Wolfgang Weitzl

Measuring Electronic
Word-of-Mouth
Effectiveness
Developing and Applying
the eWOM Trust Scale
Measuring Electronic Word-of-Mouth
Effectiveness
Wolfgang Weitzl

Measuring Electronic
Word-of-Mouth
Effectiveness
Developing and Applying
the eWOM Trust Scale
Wolfgang Weitzl
Vienna, Austria

Dissertation University of Vienna, 2014

OnlinePlus material to this book can be available on

[Link]

ISBN 978-3-658-15888-0 ISBN 978-3-658-15889-7  (eBook)

DOI 10.1007/978-3-658-15889-7

Library of Congress Control Number: 2016953682

Springer Gabler
© Springer Fachmedien Wiesbaden 2017
This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part
of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations,
recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission
or information storage and retrieval, electronic adaptation, computer software, or by similar or
dissimilar methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publication does not imply, even in the absence of a specific statement, that such names are
exempt from the relevant protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in
this book are believed to be true and accurate at the date of publication. Neither the publisher
nor the authors or the editors give a warranty, express or implied, with respect to the material
contained herein or for any errors or omissions that may have been made.

Printed on acid-free paper

This Springer Gabler imprint is published by Springer Nature

The registered company is Springer Fachmedien Wiesbaden GmbH
The registered company address is: Abraham-Lincoln-Str. 46, 65189 Wiesbaden, Germany
Acknowledgements

I wish to express sincere appreciation to some people for their indefatigable support. Without
their understanding and encouragement, as well as generosity, this doctoral thesis would not
have been possible. First, I would like to thank my supervisor Prof. Wolfgang Fritz for his trust
in the topic and his kind willingness to profoundly guide me through the research process. I
also want to thank Prof. Udo Wagner, head of the Marketing Department at the University of
Vienna, who always encouraged me to aspire to high academic standards, work independently
and do my own academic research. Both have challenged, encouraged, and motivated me to
excel in scientific study. I was really blessed to benefit from their patronage and wisdom. In
addition, special thanks to Prof. Adamantios Diamantopoulos, head of the International
Marketing Department at the University of Vienna and Joseph A. Schumpeter research fellow
at Harvard University, whose doctoral courses on structural equations modelling and
measurement theory mostly inspired my research. He and Dr. Petra Riefler provided
inestimable assistance in the preparation of this dissertation. My gratitude also belongs to my
colleagues and friends – most notably Robert Zniva, who always gave me valuable feedback,
making me a better writer, researcher, and person. Similarly, it belongs to the various scale
translators and study contributors for their generous assistance. Also to my parents and my
grandmother, who were supportive and encouraging throughout the writing process. I owe my
thankfulness to many, but to none more than Eva. Without her loving support I would not have
completed this work.

Wolfgang Weitzl

V
Abstract

Companies have learned that consumers not only collect purchase-critical insights from online
advertising, but are also increasingly turning to social media and particularly to consumer
reviews and recommendations (eWOM). However, this loss of control and “share of voice”
potentially directs consumers towards unintended business outcomes, as well as diminishing
return on marketing expenditures. Literature suggests that segmenting the communication
audience can be helpful to companies in identifying and targeting the consumer groupings most
likely to respond to their own marketing efforts and anticipating the impact of other forms of
market communication across different kinds of consumers. As trust has been found to affect
human thoughts, feelings and behaviours in various conditions, consumer trust in
eWOM/Online ads is suggested here as a valuable basis for segmentation. While valid, reliable
instruments are already available to measure generalised trust in online advertising, the same is
not true for eWOM trust.

To advance a novel segmentation approach in the era of online social shopping, this thesis (i)
clarifies the conceptual domain of eWOM trust by investigating the construct’s meaning, scope
and boundaries; (ii) identifies its key components – specifically, eWOM integrity/honesty,
ability, benevolence, and consumers’ willingness to rely and take risks; (iii) proposes a new
multi-item quantitative scale; and (iv) advances an online trust/shopping framework that relates
the construct to its causes, correlates and consequences. The conceptualisation derived from
interdisciplinary literature was the genesis of a multi-stage research process which targeted the
development of a context-specific measurement instrument – the eWOM-TRUST scale (eWT-
S) – that complies with psychometrical standards of social science. This process included a
series of complementary studies with multiple samples from various countries.

Finally, empirically based insights into the characteristics of the online communication
audience are provided by (a) describing eWOM trusters in terms of consumption-relevant
characteristics; (b) exploring the impact of eWOM trust on consumers’ usage of reviewers’
opinions; and (c) providing a new typology of eWOM/Ad trusters by means of cluster analysis.
Results show that online consumers are best segmented into four discrete clusters: (i) pure
review trusters (19%), (ii) non-trusters (10%), (iii) moderately trusting consumers (41%), and
(iv) high eWOM/Ad trusters (30%). From the managerial perspective, this thesis provides
insights into the opportunities and perils of targeting consumers with online advertising. Most
importantly, however, it enables marketers to anticipate consumers’ reliance on, as well as the
impact of, positive (negative) C2C communication across consumer groupings.

VII
Table of Contents

Acknowledgements ................................................................................................................................ V
Abstract ................................................................................................................................................ VII
Table of Contents .................................................................................................................................. IX
List of Tables ...................................................................................................................................... XIII
List of Figures ......................................................................................................................................XV
Acronyms and Abbreviations ............................................................................................................ XVII
Chapter 1 – Introduction to the Research ................................................................................................ 1
General Introduction............................................................................................................................ 1
Objectives of the Research ................................................................................................................ 11
Theoretical Importance...................................................................................................................... 12
Practical Importance .......................................................................................................................... 14
Organization of the Thesis ................................................................................................................ 16
Chapter 2 – Review of Literature and Theoretical Background ............................................................ 19
The Role of Trust .............................................................................................................................. 19
The Concept of Trust......................................................................................................................... 23
Overview ....................................................................................................................................... 23
The Disciplines’ Conceptualizations and Measurement of Trust .................................................. 32
The Philosophers’ Perspective .................................................................................................. 32
The Personality Psychologists’ Perspective .............................................................................. 33
The Sociologists’ Perspective ................................................................................................... 40
The Social Psychologists’ and Organizational Researchers’ Perspective ................................. 48
Economic, Business and Marketing Perspectives ..................................................................... 56
Towards an Integrated View of Trust Components....................................................................... 68
Components of Trust ................................................................................................................. 68
Trusting Beliefs ......................................................................................................................... 77
Trusting Attitudes ...................................................................................................................... 86
Trusting Intentions .................................................................................................................... 87
Risk-taking Behaviour ............................................................................................................... 89
Conditions of Trust ........................................................................................................................ 90
Conclusions ....................................................................................................................................... 92
Chapter 3 – Conceptualization of the Construct, Research Questions and Hypotheses ........................ 95
Evidence for Generalized eWOM Trust ............................................................................................ 95
Conceptualizing Trust in eWOM ...................................................................................................... 99
Construct Definition and Domain ................................................................................................. 99
Construct Composition ................................................................................................................ 103
Types of eWOM Trust ............................................................................................................ 104
Dimensions of eWOM Trust ................................................................................................... 107
IX
 Elements of eWOM Trust ....................................................................................................... 111
eWOM Trust and Similar Concepts ............................................................................................ 112
eWOM Distrust ....................................................................................................................... 112
eWOM Scepticism .................................................................................................................. 114
Credibility................................................................................................................................ 115
Attitude towards Reviews in General ...................................................................................... 121
Other Constructs ...................................................................................................................... 122
Definition summarized ................................................................................................................ 123
Research Questions and Hypotheses ............................................................................................... 124
Construct Identification and Dimensionality............................................................................... 125
Validity ........................................................................................................................................ 125
Criterion-Related Validity ....................................................................................................... 126
Nomological Validity: A Social Shopping – Trust Framework .............................................. 129
eWOM Trust Antecedents ....................................................................................................... 134
eWOM Trust Correlates .......................................................................................................... 151
eWOM Trust Consequences .................................................................................................... 152
Reliability and Generalizability................................................................................................... 153
Measure Application ................................................................................................................... 160
Chapter 4 – Research Methods ............................................................................................................ 163
Identification Stage.......................................................................................................................... 165
Specification of Content Domain, Construct Definition and Dimensionality ............................. 165
Generating and Judging Items ..................................................................................................... 175
Expert Interviews (Study 1) .................................................................................................... 177
Consumer Interviews (Study 2) ............................................................................................... 178
Reliability Stage .............................................................................................................................. 181
Research Instrument .................................................................................................................... 183
Sample Description and Data Collection Procedure ................................................................... 184
Validity Stage .................................................................................................................................. 186
Main Validation Study (Study 4) ................................................................................................ 187
Research Instrument ................................................................................................................ 187
Questionnaire Pretest ............................................................................................................... 201
Sampling Frame and Data Collection...................................................................................... 201
Return Rate .............................................................................................................................. 202
Sample Description ................................................................................................................. 202
Supplemental Reliability, Validity and Generalizability Studies (Study 5) ................................ 205
Application Stage ............................................................................................................................ 205
Chapter 5 – Research Results .............................................................................................................. 209
Assessment of Construct Dimensionality and Consistency ............................................................ 209
Results Reliability Stage ............................................................................................................. 210
Exploratory Factor Analysis .................................................................................................... 211

X
 Confirmatory Factor Analysis ................................................................................................. 218
Results Validity Stage ................................................................................................................. 234
Exploratory Factor Analysis .................................................................................................... 234
Confirmatory Factor Analysis ................................................................................................. 236
Results Validity Stage (Modified Model) ................................................................................... 245
Model Goodness-of-fit ............................................................................................................ 245
Reliability ................................................................................................................................ 246
Convergent Validity ................................................................................................................ 247
Discriminant Validity .............................................................................................................. 249
Implications for Construct’s Structure and Content .................................................................... 251
Supplemental Validity Assessments ............................................................................................... 256
Concurrent Validity ..................................................................................................................... 256
Convergent Validity and Discriminant Validity.......................................................................... 256
Known-group Validity ................................................................................................................ 266
Predictive and Postdictive Validity ............................................................................................. 267
Nomological Validity .................................................................................................................. 270
Social Desirability Bias ............................................................................................................... 273
Supplemental Reliability and Generalizability Assessments .......................................................... 274
Test-Retest Reliability ................................................................................................................. 274
Generalizability to different Online Sources ............................................................................... 274
Generalizability to individual Customer Reviews ....................................................................... 279
Generalizability to different Languages and Cultures ................................................................. 281
Scale Finalization and Development of Norms ............................................................................... 287
Scale Application: Segmenting Online Consumers ........................................................................ 289
Segmenting Consumers on eWOM/Online Ad Trust .................................................................. 289
Profiling the Segments ................................................................................................................ 290
Chapter 6 – Discussion, Implications and Future Research ................................................................ 299
Summary of Research ..................................................................................................................... 299
Conclusions and Discussion of Major Findings .............................................................................. 307
The Five Sub-Dimensions of eWOM Trust ................................................................................ 308
The Three Elements of eWOM Trust .......................................................................................... 310
The Characteristics of eWOM Trusters ....................................................................................... 312
Profiling Recipients of Online Market Communications ............................................................ 313
Contribution and Implications ......................................................................................................... 316
Limitations and Directions of Future Research ............................................................................... 320
References ........................................................................................................................................... 325
Appendices .......................................................................................................................................... 375

XI
List of Tables

Table 1: Dictionary Definitions of Trust .................................................................................. 29

Table 2: Matrix of Trust Dimensions ....................................................................................... 77
Table 3: Classification of Trusting Beliefs .............................................................................. 79
Table 4: Theorized Key Components of eWOM Trust .......................................................... 104
Table 5: Initial Set of eWOM Trust Items ............................................................................. 182
Table 6: Measures of Generalized eWOM Trust ................................................................... 189
Table 7: Measures of Attitude towards Reviews in General .................................................. 190
Table 8: Measures of Review Credibility .............................................................................. 192
Table 9: Measures of Reviewer Credibility ........................................................................... 194
Table 10: Measures of Distrust in Online Reviewers‘ Motives ............................................. 195
Table 11: Measures of Online Review Usage ........................................................................ 196
Table 12: Measures of Online Review Avoidance................................................................. 197
Table 13: Measures of Review Involvement.......................................................................... 198
Table 14: Measures of Disposition to Trust ........................................................................... 199
Table 15: Measures of Review Experience ............................................................................ 200
Table 16: Measures of Internet Experience ............................................................................ 200
Table 17: Sample Composition of Study 4 ............................................................................ 203
Table 18: Respondent Characteristics of Study 4 .................................................................. 203
Table 19: Results of the Initial EFA (Sample 3a and 3b) ...................................................... 214
Table 20: Reliability and Item-based Statistics ...................................................................... 216
Table 21: Psychometric Properties of the eWOM Trust Scale (Samples 3a and 3b) ............. 222
Table 22: Correlations among the Sub-dimensions (Samples 3a and 3b).............................. 225
Table 23: Competing Models (Sample 3a) ............................................................................ 227
Table 24: Competing Models (Sample 3b) ............................................................................ 228
Table 25: Pairings of Sub-dimensions (Sample 3a) ............................................................... 230
Table 26: Pairings of Sub-dimensions (Sample 3b) ............................................................... 232
Table 27: Results of the EFA (Sample 4) .............................................................................. 235
Table 28: Psychometric Properties of the eWOM Trust Scale (Sample 4) ............................ 238
Table 29: Correlations among the Sub-dimensions (Sample 4) ............................................. 239
Table 30: Competing Models (Sample 4) .............................................................................. 241
Table 31: Pairings of Sub-dimensions (Sample 4) ................................................................. 243
Table 32: Psychometric Properties of the eWT-S (Modified Scale, Samples 4 and 5) ......... 248
Table 33: Correlations among the Sub-dimensions (Modified Scale, Sample 4) .................. 249
Table 34: Competing Models (Modified Scale, Sample 4) .................................................... 252

XIII
Table 35: Pairings of Sub-dimensions (Modified Scale, Sample 4) ...................................... 254
Table 36: Convergent Validity ............................................................................................... 257
Table 37: Multi-Trait Multi-Method Matrix .......................................................................... 259
Table 38: Model Comparison (eWOM Trust – Disposition to Trust).................................... 263
Table 39: Results of the EFA (Sample 4) .............................................................................. 264
Table 40: Predictive Validity ................................................................................................. 269
Table 41: Postdictive Validity ................................................................................................ 270
Table 42: Nomological Validity (Sample 4) .......................................................................... 271
Table 43: Nomological Validity (Sample 5) .......................................................................... 273
Table 44: Test-Retest Reliability............................................................................................ 275
Table 45: Psychometric Properties of the eWT-S (Alternative Online Sources) ................... 278
Table 46: Influence of Generalized eWOM Trust on Perceptions of Individual Reviews .... 280
Table 47: Results of the EFA (Sample 6) .............................................................................. 282
Table 48: Psychometric Properties of the eWOM Scale (Sample 6) ..................................... 285
Table 49: Measurement Invariance (Sample 4 and 6)............................................................ 286
Table 50: Construct and Item Descriptives (Sample 4 and 6)................................................ 288
Table 51: Cluster Demographics ............................................................................................ 291
Table 52: Cluster Description ................................................................................................ 294
Table 53: eWOM Trust Scale (eWT-S) ................................................................................. 302

XIV
List of Figures

Figure 1: Typical Online Customer Review .............................................................................. 3

Figure 2: Preliminary Conceptualization of Generalized eWOM Trust ................................ 124
Figure 3: The Nomological Framework ................................................................................. 133
Figure 4: Overview of the Research Process ......................................................................... 166
Figure 5: Adopted Measurement Model ................................................................................ 219

XV
Acronyms and Abbreviations

AGFI Adjusted Goodness-of-fit Index

AIC Akaike’s Information Criterion
ATSCI Attention towards Social Comparison Information
AVE Average Variance Extracted
B2C Business-to-Consumer
C2C Consumer-to-Consumer
CDS Consumer-developed Review Sites
CFA Confirmatory Factor Analysis
CFI Comparative Fit Index
CI Confidence Interval
CSA Covariance Structure Analysis
CSII Consumer Susceptibility to Interpersonal Influence
EFA Exploratory Factor Analysis
ELM Elaboration Likelihood Model
eWOM Electronic Word-of-Mouth
eWT-S Electronic Word-of-Mouth Trust Scale
GFI Goodness-of-fit Index
KMO Kaiser-Meyer-Olkin
MDS Marketer-developed Review Sites
ML Maximum Likelihood
MSA Measure of Sampling Adequacy
MTMM Multi-trait Multi-method
NFI Normed Fit Index
NNFI Non-normed Fit Index
OCR Online Customer Reviews and Recommendations
PCA Principal Components Analysis
PFI Parsimonious Fit Index
PLS Partial Least Squares
RMSEA Root-mean-square-error-of-approximation
RMR Root-mean-square-residual
SCT Social Cognitive Theory
SEM Structural Equation Modelling
SET Social Exchange Theory
SRMR Standardized Root-mean-square-residual
TLI Tucker-Lewis Index
TPB Theory of Planned Behaviour

XVII
TRA Theory of Reasoned Action
UGC User-Generated Content
WOM Word-of-Mouth

XVIII
Chapter 1

Introduction to the Research

General Introduction
It is not assumptive to claim that trust is the key ingredient of social relationships,
communications and economic transactions. For Luhmann (1979), one of the most influential
researchers in the domain, trust is nothing less than a general prerequisite for human behaviour,
as it is “a basic fact of social life” (p. 4). For the sociologist, the absence of trust would impede
us in undertaking normal everyday activities, such as talking with others and making new
friends. Trust, its causes, nature and consequences, is recognized as a critical concept and
acknowledged not only by sociologists; academics from a wide spectrum of intellectual
disciplines, such as psychology, economics, marketing and communications, have contributed
to our current body of knowledge. For scholars in these fields, investigation into the conceptual
nature and the construct’s measurement often lie at the very heart of their research. A profound
operationalization of the construct is here recognized as a necessity for subsequent scientific
research. While the view of the concept often differs from discipline to discipline, trust has
commonly been found to have a vital role in human interactions and effective interpersonal
communication – whether it takes place offline or online.

Since the early 2000s, consumers are increasingly turning to the Internet to source information,
as this medium furnishes consumers with radically new opportunities to interact online and
conduct pre-purchase information searches (Brown, Broderick & Lee, 2007; Hu, Bose, Gao &
Liu, 2011; Jepsen, 2007). These new opportunities are often associated with the concept of Web
2.0, which basically describes the development, dissemination, and sharing of user-generated
content (UGC) and the decline in importance of other sources. Consumers’ usage of Web 2.0
applications today is mainly motivated by changes in their interests, needs, and behaviours.
These new tools have changed the way people search, find, read, gather, share or consume
information, as well as the way people communicate with each other and collaborate to create

1
© Springer Fachmedien Wiesbaden 2017
W. Weitzl, Measuring Electronic Word-of-Mouth Effectiveness,
DOI 10.1007/978-3-658-15889-7_1
new knowledge (Casalo, Flavian & Guinaliu, 2011). We are witnessing a radical change in
consumer habits and profound knowledge about this trend has become an imperative for all
marketers worldwide.

The essential difference between Web 2.0 and the traditional Web is that content is generated
and steered by consumers; moreover, there is considerably more collaboration amongst Internet
users. During the old days of the Internet it was mostly up to institutions such as companies to
provide a static and controlled content. Now, however, the web has become more dynamic,
interconnected, and uncontrolled, shaped by social media and networks. It’s the consumers who
produce content for others and who dominate this communication channel. Nowadays,
information issued by marketers in the form of online advertising is often said to considerably
lose impact to the advantage of these sources. In such an era, Web 2.0 social media applications
like bulletin boards, blogs, newsgroups, chat rooms, review sites (e.g., [Link],
[Link]), and product-, brand- or anti-brand-centric discussion forums on social networks
such as MySpace and Facebook, and video channels (e.g., [Link]) all furnish consumers
with valuable feedback and evaluation tools (Dellarocas, 2003; Goldsmith, 2006).
Contemporary consumers have found that these platforms represent an excellent opportunity to
gather and share consumers’ personal ideas, preferences, opinions and experiences with
manufacturers, retailers, products and services with a vast, geographically dispersed group of
people, who have experience and knowledge on relevant consumption issues (Duan, Gu &
Whinston, 2008; Forman, Ghose & Wiesenfeld, 2008; Godes & Mayzlin, 2004; Ratchford,
Talukdar & Lee, 2001). Today, hundreds of millions of people raise their voices and contribute
to social media or passively consume the statements of others (Kamplan & Heanlein, 2010).
And these voices are not always positive for the marketer. In times of considerable economic
downturn, individuals are regularly forced to look after their expenses and to become smarter
consumers. When consumers lack own experiences, the experiences of others are often seen as
a good way to compensate for information shortages and to anticipate the consequences of
buying actions. As consumption takes place more and more in the public arena, others’ opinions
and experiences also matter more.

Hence, consumers actively or passively engage in electronic word-of-mouth (eWOM)

communication (Bagozzi & Dholakia, 2002; Casalo et al., 2011; Dwyer, 2007; Hennig-Thurau,
Gwinner, Walsh & Gremler, 2004) today. In reference to earlier contributions, this thesis
defines eWOM (or online customer reviews (OCR), which is here used as a synonym) as peer-
generated, text-based product evaluations and recommendations made by potential, actual, and
former customers about a product or company, which are made available to a multitude of
consumers via postings on the Internet (Hennig-Thurau et al., 2004; Mudambi & Schuff, 2010).
This definition is also consistent with the one advanced by Rafaeli and Raban (2005), who view
eWOM reviews or recommendations as positive, neutral or negative information about a
product released on the Internet by a consumer. Therefore, customer reviews (see Figure 1 for

2
an example) are regarded as an often-used form of market communication or, more precisely,
a variety of consumer-to-consumer (C2C) online communication. While slight variations are
possible from website to website, OCR are typically characterized by an unstructured
discussion of the product by a consumer which can vary considerably in length, a “star”-rating
for the overall product performance or satisfaction, the user name of the consumer and, at best,
accompanying information on the consumer (e.g., origin, date of purchase), date of review, and
the possibility of responding to the review.

Others’ advice or recommendations about purchase-critical issues have a critical role for
today’s purchasing decisions as consumers often refrain from using marketer-driven
communication sources, like online advertising such as retailers’ websites or professional
advice (Lee, Park & Han, 2008). This is caused by a multitude of characteristics of this
information source. For instance, eWOM provides consumers with more choices and relevant
consumption insights, and allows them to make quick but comprehensive comparisons.
Moreover, it enables shoppers to directly communicate and exchange thoughts and ideas, as
well as compare consumption experiences with other online shoppers in the electronic
community who share the same interests. In contrast, online advertising is commonly seen as a
less objective source of information that is biased by the interests of the company. Shopping
online is often an ambiguous endeavour, frequently associated with a set of different risks;
evaluating the suitability of a product that suits a person’s needs is here regularly far from clear.
In such situations, individuals regularly look to others for cues (Darley & Latanße, 1968).
Online reviews are typically regarded as especially valuable cues, as they offer key information
about market offerings and thus typically facilitate purchase decision processes as a whole (Ku,
Wei, & Hsiao, 2012).

Figure 1: Typical Online Customer Review

The phenomenon that more and more consumers consult online reviews is supported by several
market studies. For instance, a Deloitte study (Deloitte, 2007) shows that consumers have been
used to turning to eWOM in the form of online customer reviews and recommendations for
many years. Through this action, consumers build up their arsenals of knowledge and satisfy
their need to research product information online before making purchases and sharing

3
experiences regarding the product after its using it. According to this (quite old) study, almost
two thirds (62%) of consumers read product reviews created by fellow shoppers and made
available on the Internet. Another (more recent) study provides similar figures, as 70% of
consumers consult reviews or ratings before purchasing (Businessweek, 2009). In addition,
Deloitte’s (2007) study provides evidence that these reviews exert strong societal and personal
influence, as 82% of these consumers state that there is a direct linkage between their final
shopping choice and the consumer reviews consulted, either influencing them to buy a different
product from the one they had originally been thinking of, or confirming the original purchase
intention. Another study has found that 77% of online shoppers consult reviews and ratings
when doing online shopping (McGuigan, 2008). These facts correspond roughly with a more
recent study demonstrating the increasing importance of reviews. Here, 91% of respondents
mentioned that they consult online reviews, blogs, and other user-generated content before
making a purchase; of these, 45% say that their choices are affected quite considerably by
reviews placed on social networking sites by individuals whom they follow (Harris Interactive,
2010). In a large-scale survey of online consumers, Kee (2008) found that 22% of the
respondents say that they always read online reviews before making the final purchase decision
and 43% have indicated that they visit ratings and reviews most of the time. The same study
concludes that more than 90% of shoppers deem online reviews to be extremely or highly
useful. Forrester predicts that over 50% of worldwide retail sales will be influenced by Internet
content (i.e., online reviews) by 2014 (VanBoskirk, 2009). When consumers nowadays go
online, they can hardly avoid online customer reviews and recommendations. 75% of all
Internet users say that they have created or consumed some kind of social media content,
including product reviews (Kampla & Haenlein, 2010). Only 12% say that they take no notice
of them at all (Anderson, 2013b). While demonstrating the importance of customer online
reviews today, these findings also indicate that perceptions and usage patterns concerning this
information source also vary among segments of customers.

The increasing popularity of communication between consumers on product search and

shopping choice behaviour is also mirrored by the recent success of Internet sites such as
[Link], [Link], [Link], [Link], and [Link], which
are regarded as having the fastest-growing Internet clientele for years (ComScore, 2007). These
third-party product-review websites have introduced opinion-sharing communities in order to
enable the exchange of customer reviews and consumption experiences about a tremendous
amount of products, as well as companies (Ku et al., 2012). But OCR success is also bound to
the success of social platforms such as Facebook and Twitter, which are used more and more
as recommendation channels (Anderson, 2013b). Due to the possibility of providing
personalized recommendations from known or highly reputable people, it is very likely that
social platforms will play a key role in the recommendation business. And it is quite reasonable
to argue that they will radically change online purchasing with new powerful review tools in
the near future.
4
An increasing number of companies are becoming aware of the relevance of eWOM
communication for their own market success (e.g., Thomson, 2003). Consequently, several
companies have put forward initiatives to introduce Web 2.0 applications to streamline their
business models to the new communication demands and offer added value to their customers
by including public voice in their advertising strategy. These companies regularly want to
benefit from the two worlds by integrating them: marketer-driven online advertising, as well as
consumer-based advertising. Meanwhile, it has become a common practice for retailers (e.g.,
[Link], [Link], [Link]) and manufacturers (e.g., Nike, Levi’s) to establish
their own opinion-sharing communities and offer shoppers the chance to participate actively by
posting their thoughts on and experiences with products and brands they have recently bought
or are interested in. (Casalo et al., 2011). Leading online shops such as [Link] have been
inviting shoppers to submit product reviews for many years, while others have started to offer
this communication instrument more recently or are about to launch new community platforms.
There are also examples where retailers abstain from providing product information supplied
by themselves in favour of online reviews. Firms have acknowledged that customer reviews are
a valuable form of third-party evaluation, able to increase company reputation and trust in
online shopping: 60% of retailers already use customer reviews on their websites ([Link],
2009) and 97% of the top 250 e-retailers are on Facebook (Scarpello, 2012). Other firms have
decided to purchase customer reviews from [Link] or other retail sites and to publish
them in their own online shop (Mudambi & Schuff, 2010). Consumers are already used to
accessing conversations with fellow shoppers on retailing websites: 58% read online reviews
on shopping sites and 47% visit them on company websites (Lightspeed Research, 2011). One
can say that, due to current consumer demand, the inclusion of social web features is already
an imperative for every online shop owner.

It seems that retailers have good reason to consider reviews in their communication strategy, as
research has shown that online reviews can steer consumer behaviour. In recent years, a
significant number of studies have been published which all strive to investigate the
effectiveness or persuasiveness of eWOM messages (see Cheung & Dhadani, 2012 for a
review). They all more or less agree that customer reviews can be a very powerful marketing
force (under certain circumstances). For instance, prior studies indicate that customer reviews
can have a positive influence on sales (e.g., Archak, Ghose & Ipeirotis, 2011; Chevalier &
Mayzlin, 2003; Clemons, Gao & Hitt, 2006; Formann et al., 2008). More specifically, by
investigating the impact of online book reviews at [Link] and [Link],
Cheavalier and Mayzlin (2006) find that online reviews have a positive effect on book sales,
Clemons et al. (2006) identify a similar impact on the growth of craft beer sales, and Reinstein
and Snyder (2005) conclude that reviews are able to increase customer demand for movies.
Such a relationship has been found in the same context by Liu (2006) and Basuroy et al. (2003).
These findings suggest that reviews have the most influential power if the reviewed products
are intangible in nature and can only be evaluated upon consumption (e.g., belletristic books) –
5
which holds true for experience products (Korfiatis, Garcia-Bariocanal & Sanchez-Alonso,
2012). Other studies indicate that (positive) consumer reviews and ratings can lead to increased
buyer trust in the exchange partner (Ba & Pavlou, 2002; Pavlou & Gefen, 2004), as well as
price premiums for the retailer. In addition, the presence of online reviews can impact the
overall evaluation of a website’s usefulness and perceptions of social presence (Kumar &
Benbasat, 2006). Additionally, consumer feedback systems can enhance website traffic, as well
as the time a consumer spends on the website (i.e., “site stickiness”) (Mudambi & Schuff, 2010).

Researchers have investigated the causes that make online customer reviews regularly so
influential compared to other forms of market communication. They agree that its impact is
merely the result of common features that eWOM and traditional (offline) word-of-mouth
(WOM) share, but it is also the effect of distinctive characteristics which this type of
communication has in common with its marketer-generated counterpart (online advertising), as
well as unique features of its own. WOM communication has long been recognized by
researchers and practitioners as an important element of marketing. The rich body of traditional
communication theory considers WOM to play a key role in purchasing decisions by directly
affecting shopping choices (Arndt, 1967; Katz & Lazarsfeld, 1955; Olshavsky & Granbois,
1979; Richins & Root-Shaffer, 1988). Here, WOM conveys valuable information concerning
product characteristics and performance, as well as the social and psychological consequences
(risks) that typically arise out of the purchase decision. Therefore, it is a potential consumption
risk reducer. Additionally, past research describes WOM information as “the most powerful
form of marketing communication” (Ekici, 2004) and hence it is seen to be more influential
when compared to traditional marketing, including personal selling and advertising (Katz &
Lazarsfeld, 1955). This is mainly attributable to the fact that WOM is assumed to be more
reliable, credible and believable, as it is communicated by people that are not marketers of the
product. As eWOM is also typically generated by other consumers, the lack of commercial self-
interest in consumer online reviews and recommendations makes the conveyed information
generally more believable to the reader. So trust is essential here. However, other characteristics
apply that make its impact on consumer behaviour distinctive. For instance, consumers usually
collect WOM about a market offer from a very restricted number of individuals (e.g., one or
two) (Zeithaml & Bitner, 2000). In contrast, on the Internet, products and services may be
discussed by a dozen or typically even more mostly anonymous customers who possess a
diversified background and different motivations (Litvin, Goldsmith & Pan, 2008). Hence, in
the online environment, there is typically a greater amount of additional information available
to the shopper than from traditional WOM. As a consequence, consumers are often confronted
with a multitude of positive and negative (i.e., ambivalent) reviews (Xie, Miao, Kuo & Lee,
2011), endangering trust. However, eWOM is likewise valued for its non-commercial character
and consumers rely on other consumers’ product information, as it is deemed credible and
trustworthy (Godes & Mayzlin, 2004; Mayzlin, 2006). While consumer trust in online reviews
may be limited compared to traditional WOM, shoppers are typically more susceptible to
6
eWOM than to market communication issued by companies (Bickart & Schindler, 2001;
Trusov, Bucklin & Paulwels, 2009), which suffers from a lack of credibility.

Second, WOM typically means information and opinions gathered from acquaintances (e.g.,
friends and relatives). Here, the persuasiveness of information is attributable to the close links
or “ties” that exist between the person giving the information (i.e., the reviewer) and the
receiving consumer. In this context, Duhan et al. (1997) discuss the concept of “tie strength”.
A tie is regarded as being weak where the recommender is not known by the consumer or has
only weak relationships with the consumer. In contrast, a strong tie exists in cases when the
shopper knows the reviewer or recommender personally or “feels” that he does. While most
WOM originates from people with whom the shopper shares strong ties, eWOM messages are
typically comments from invisible and anonymous strangers (weak ties). As a result, consumers
have some concern over the source credibility (Kelman & Hovland, 1953) of eWOM, as it is
often not as readily established as WOM (Park, Lee & Han, 2007). Here, the reader is generally
not familiar with the credentials of the reviewers (e.g., knowledge, motivations), which means
that his/her reactions to the reviews are largely shaped by the consumer’s general attitudes
towards this information and the population of reviewers, besides some degree of situational
inference of credibility information. This is done through cues that are present within the review
or exist in its environment. For example, the positive source credibility effect, which aids
traditional WOM coming from an impartial source, can be diminished in the case of eWOM
when it is featured on a site that sells the product at the same time. Or in other words, a context
effect is said to exist. Accordingly, reviews on independent third-party websites are more likely
to be preferred by consumers, compared to the platforms which are clearly operated by a
business entity which has a vested interest (Senecal & Nantel, 2004). To cope with the latent
uncertainty of review source credibility, some shopping portals and online retailers require
reviewers to provide personal information which enables the consumer to “identify” the review
creator or at least make some inferences (e.g., name, state of residence, gender). This should
foster the perceived “ties” between participants.

Third, prior research provides evidence that the strong impact of WOM communication on
product judgements and purchasing decisions can be partially explained by the fact that the
information is received in a face-to-face manner and hence is more accessible than information
presented in a less vivid way (Herr, Kardes & Kim, 1991). Consumers can easily interact
verbally with the individual who creates WOM. This enables him/her to obtain detailed
contextual insights about the experience with a product or service provider. This furnishes
consumers with the opportunity to make more confident attributions for and judgments about
products and services (Weiner, 2000). In contrast, eWOM comes from a virtual source and is
an asynchronous activity, and contextual information is often much less accessible. Interaction
possibilities are also restricted. In everyday conversation, people typically communicate largely
through non-verbal language and visual cues, including hand gestures, eye contact, and smiling.

7
The absence of such contextual information and the reviewer’s anonymity make eWOM
messages often appear more nebulous to consumers. This leads to the perception that eWOM
is less vivid.

Criticism of the power of online reviews to influence consumers in their buying decisions is
regularly based on the potential to fake seller postings as customer or independent reviews.
Clearly, retailers have a strong incentive to disguise promotional reviews as legitimate eWOM
communication in order to influence consumers’ evaluation of their products (Mayzlin, 2006).
This manipulation can be easily carried out by interested parties – and the educated Internet
user knows that. There is no mechanism which ensures eWOM quality. This is one reason
which makes the assessment of eWOM reviews and recommendations generally tough for the
consumer (Smith, Coyle, Lightfoot & Scott, 2007). Everybody on the Internet – including
marketers – is qualified to anonymously post considerate product reviews on products. But it is
also possible for companies to enlist or hire an “impartial” reviewer who has been paid or has
received another form of incentive in order to provide a positive review. At the same time, a
“warning” review may be a competitor’s posting on the brand striving to hurt the opposition.
And it is true that a wave of negative eWOM can enduringly harm a brand’s image. Such
unethical business practices are, however, quite common today as the trade press regularly
uncovers unethical review practices (e.g., Moe, 2012; Tuttle, 2012) or “astroturfing”
incidences. For instance, the media reported the dishonest marketing tactics of a popular coffee
machine producer that encourages an employee to post very positive reviews about a number
of the company’s products on [Link] (Pilon, 2009). Similarly, a considerable number of
“glowing” reviews on a well-known hotel booking website turned out to be the product of
contributors paid by hoteliers (Fernandez, 2011). In such cases, online reviews have essentially
become nothing more than another kind of marketer-to-consumer advertisement. This
effectively defeats the original purpose of enabling eWOM to influence consumers in their
decision-making process as the content added is no longer independent, objective or credible
(Bray & Schetzina, 2006). However, consumers are becoming more and more aware of this
kind of manipulative practice and many consumers are sceptical about the information
conveyed in eWOM messages (Baumbauer-Sachse & Mangold, 2013; Senecal & Nantel,
2004). In the eyes of these consumers, eWOM is nothing more than market communication
intended to steer and manipulate individuals in the interest of the consumption system. A market
research study by Jupiter Research suggests that only 21% of consumers actually rely on
information provided in product reviews on social networking sites such as Facebook or
MySpace (Wassermann, 2006). In 2010, 57% of shoppers trust customer reviews as a research
source, but 35% question whether they are biased (Mediapost, 2010). A more recent study
demonstrates that, despite the differentiating characteristics of eWOM and WOM, 79% of
Internet users “trust” online reviews as much as personal recommendations (Anderson, 2013a).
These data would suggest that consumers on average rely more and more on information given
in reviews and for the most part disregard threats of message integrity. Assuming further that
8
the method of measuring trust in reviews is correct, the figures indicate that people value
eWOM and WOM recommendations almost equally. But are these implications really correct?

The primary communication function that society has assigned to eWOM is its role of informing
consumers about products, services, and service providers in order to help shoppers weigh one
consumption choice against another. However, online consumer reviews can have a dual task:
(i) they provide useful and purchase-critical information, and (ii) in addition, they also act as a
recommender (Burgess, Sellitto, Cox & Buultjens, 2011). Crucial for the fulfilment of both
functions and for the consumers’ acceptance as an effective decision-making aid is that
consumers trust the information conveyed in online reviews. Hence, the most fundamental
question in evaluating the effectiveness and persuasiveness of eWOM messages is whether
consumers believe what they read – or in other words – whether they really trust the information
conveyed in online customer reviews or not (Urban, Amyx & Lorenzon, 2009). Or is the fear
of deception greater? Previous discussion on eWOM characteristics has outlined that a major
factor which determines whether eWOM functions most effectively as an information and
recommendation source, as well as which attitudes consumers form toward online reviews, is
finally the consumers’ general trust in this communication vehicle. Once consumers trust the
fellow shoppers’ reviews and recommendations, their willingness to choose the socially-
recommended product is enhanced and thus their purchasing behaviour may be meaningfully
influenced. This rationale complies with the insights of interdisciplinary research – led by social
psychology – that regards the trust concept as a key moderator of interpersonal communication.
For instance, Obermiller and Spangenberg (1998) find that trust in advertising is a primary
determinant of its influence. The relevance of trust is not only revealed by academic research,
but it is also mirrored explicitly and implicitly by non-academic research: in industry research
reports, the effectiveness of online consumer reviews and recommendations is often attributed
to the consumers’ trust in this type of market communication. For instance, Nielsen (2012)
reports that when making purchasing decisions, online shoppers rely on reviews posted by
anonymous others (70%) more than they trust traditional media (47%). These figures are very
insightful since they imply that what makes one information source more influential than
another is the trust of the receiving party. And it goes almost without saying that a person’s
reaction to any form of information is ultimately affected by his/her level of trust towards it.
For instance, if consumers do not believe that eWOM claims are true, the value of online
reviews as an information as well as a recommendation source completely disappears. In other
words, shopping ambiguity can only be reduced by OCR when individuals can truly depend on
it. Hence, trust is key to eWOM effectiveness – as it is in any form of human communication.

In contrast to the great interest in diverse disciplines, scholars put little effort into systematic
research of the trust concept in the eWOM context. When research surfaces, the main research
interest seems to be in the investigation of the antecedents and consequences of eWOM
communication. Especially the latter area of research has produced a variety of mixed results,

9
which may stem from the defiance of the key role of consumer trust in former research
frameworks. The neglect of the role of trust as a key characteristic of computer-mediated
communication is quite astonishing, as the concept itself is generally regarded as an important
antecedent of behaviours which demonstrate dependence on others (such as the acceptance of
others’ advice) by other disciplines (McKnight, Choudhury & Kacmar, 2002a; Smith, Menon
& Sivakumar, 2005). Only a few marketing scholars have done pioneering work in the
incorporation of the trust concept. Here, some academic studies have attempted to measure trust
in eWOM several times (e.g., Briggs, Burford, Angeli & Lynch, 2002; Burgess et al., 2011;
Hsiao, Lin, Wang, Lu & Yu, 2010; Liu & Zhang, 2010; Pan & Chiou, 2011; Racherla,
Mandviwalla & Connolly, 2012; Sen, 2008; Sen & Lerman, 2007; Smith et al., 2005). And also
industry reports are typically interested in adequately assessing the amount of consumer trust
in eWOM (e.g., eMarketer, 2008; Nielsen, 2012; PowerReviews, 2010).

However, eWOM trust research – as a young field of inquiry – is still characterized by

conceptual confusion. Given the prominence of credibility (a trust-related construct) in research
on communications and information systems (IS), researchers have incorporated the concept
quite early in the context of peer recommendations and online reviews (e.g., Brown et al., 2007;
Lis, 2013; Mackiewicz, 2007; Wathen & Burkell, 2002). Nan (2009) even claims that
credibility is one of the most frequently studied triggers of source convincibility. Research
contributions agree that the concept is key to online as well as offline persuasion (e.g., Cheung,
Luo, Sia & Chen, 2009; Tormala & Petty, 2007). These insights document the important role
of trust-related concepts for information gathering. However, literature suggests that credibility
and trust are distinct constructs with several divergent conceptual characteristics (Doney &
Cannon, 1997; Hovland, Janis & Kelley, 1953; Lewis & Weigert, 1985; Ohanian, 1990). The
dictionary defines trust as an assured reliance, confident dependence, and confident anticipation
(Merriam-Webster’s New Encyclopedic Dictionary, 1996). The trust literature, which is
reviewed in Chapter 2, characterizes trust as a complex construct, often defined differently as
“a willingness to behave based on expectation about the behaviour of others” (Luhmann, 1988),
“an expectancy that the word, promise, verbal or written statement of another can be relied
upon” (Rotter, 1967), or “a willingness to rely on an exchange partner in whom one has
confidence” (Moorman, Deshpande & Zaltman, 1993). The same literature also suggests that
trust is composed of several elements that are not typically included in the credibility construct,
such as confidence in, or a positive expectation of, the trusted party’s benevolence, mutual
emotional bonds between the involved parties, as well as a person’s willingness to rely on the
trusted party (e.g., Doney & Cannon, 1997; Johnson-George & Swap, 1982). Hence, it would
be far from correct to regard credibility and truth as synonyms.

The lack of adequate research on eWOM trust in the face of its latent theoretical, as well as
practical, importance calls for a more systematic investigation into the conceptual nature and
measurement issue of the construct. These issues as well as the implications of consumer trust

10
in online customer reviews and other forms of market communications are the focus of this
thesis.

Objectives of the Research

The research described in this thesis contributes by developing and validating an instrument to
quantitatively measure generalized trust in eWOM-conveyed information (i.e., the eWOM trust
scale or eWT-S). Additionally, this dissertation applies the scale by investigating the concept’s
role for segmenting online shoppers. More specifically, the objectives of the research at hand
are:

1. To clarify and advance the theoretical conceptualization of eWOM trust by investigating

the construct’s conceptual meaning as well as scope and to provide evidence on the
unique characteristics of the construct which distinguishes it from other eWOM
concepts.
2. To introduce a reliable as well as valid multi-item scale that quantitatively captures trust
in eWOM by applying a rigid measurement development process.
3. To investigate the role of eWOM trust in an e-commerce/trust framework by evaluating
the concept’s antecedents, correlates, and consequences and to characterize eWOM
trusters.
4. To apply the new scale to segment online consumers in accordance with their
generalized trust in C2C (eWOMTrust) as well as B2C communication (oADTrust) and
describe the segments in terms of attitudinal, perceptual and habitual characteristics.

In order to satisfy the aforementioned objectives, a multi-stage research design was followed
which meets both the demands of classical test theory and traditional scale development
standards – advanced, for example, by Churchill (1979) as well as Netemeyer et al. (2003).
Such an approach is justified due to the nature, and this thesis’ conceptualization, of the
discussed construct. Based on synthesizing previous trust research of various intellectual
disciplines, this work advanced a preliminary definition of eWOM trust by describing it as a
consumer’s general confidence that the information conveyed in online customer reviews is
reliable. More specifically, eWOM trust was formally defined as a five-dimensional, second-
order construct capturing the extent to which a consumer has the belief that information given
in online customer reviews is (a) honest, (b) useful, and (c) benevolent; (d) has a favourable
attitude towards this kind of market information; and (e) is willing to rely on online customer
reviews in general. This definition built the reference of this research’s identification stage
(stage one). As described in Chapter 3, the definition covers the nature of eWOM trust as it
includes the three underlying elements that are critical for the formation of consumer trust (i.e.,

11
cognitive, affective, and behavioural elements), as well as the distinct dimensions of trust. The
preliminary trust definition and an initial pool of items were evaluated for conceptual
correctness and translation validity by means of various consumer and expert interviews.

During the reliability stage (stage two), a student survey in two universities was conducted in
order to verify the construct’s dimensionality. Here, the survey data was split into two data sets,
of which one half of the sample was analysed by applying exploratory factor analysis (EFA)
and internal consistency analysis in order to select the items most representative of its intended
domain. Then, a confirmatory factor analysis (CFA) was applied. The same procedure was also
applied to the second half of the sample (the holdout sample) in order to verify and validate the
structure of the items resulting from the first data set. Following this, new representative data
sets were collected by surveying Internet users from Austria and Germany and analysed to
examine diverse issues of construct validity in the validity stage (stage three). This process also
led to a refinement of the original definition.

By using additional fresh samples, additional reliability (e.g., generalizability of the scale to
different cultural/language contexts) as well as validity checks (e.g., known-group validity,
predictive validity, nomological validity) were achieved. The third and fourth stage (i.e.,
application stage) also offered empirically-based insights into the characteristics of eWOM
trusters by (i) describing eWOM trusters in terms of consumption-relevant characteristics (e.g.,
susceptibility to informational and normative influence) ; (ii) exploring the impact of eWOM
trust on consumers’ usage of reviewers’ opinions for their own purchase decisions and on active
engagement in offline/online WOM; and (iii) providing a new typology of eWOM/Ad trusters
by means of cluster analysis and profiling the identified segments. The details on the research
methodology are presented in Chapters 4 and 5.

Theoretical Importance
In recent years, marketing scholars as well as practitioners in the business sector have both
recognized the extensive exchange of information among shoppers online. A growing body of
literature mirrors the increasing research interest in the field of online advertising and eWOM
effectiveness. In particular the topic of online reviews and recommendations as well as social
media impact have become a vital field of scientific inquiry (e.g., Chung, 2011; Dhar & Chang,
2009). Much is already known. However, despite this heightened interest, research dedicated
to providing adequate eWOM measures attempting to capture the characteristics of eWOM
review and recommendations that make them persuasive rarely surface. Only a handful of the
paper-and-pencil scales typically applied in the greater context of online shopping are validated.
Until now, the focus of fruitful eWOM research was on topics such as the motivational
background of contributing to eWOM (Hennig-Thurau et al., 2004; Walsh & Mitchell, 2010),
or the optimal approach to spreading viral communication (Berger & Milkman, 2012). In
contrast, contributions on the appealing and critical concept of eWOM trust remain scarce.
12
In online research, the concepts of trust in shops or online shopping (e.g., Beldad, De Jong &
Steehouder, 2010; Bente, Baptist & Leuschner, 2012), commercial websites (e.g., Ahrholdt,
2011; Aiken & Bousch, 2006; Chang & Chen, 2008), vendors (e.g., Becerra, 2006;
Benedicktus, 2011; Chau, Hu, Lee & Au, 2007), products (e.g., Becerra, 2006), brands (e.g.,
Alam & Yasin, 2010), and the Internet as a shopping institution (e.g., Grabner-Kraeuter &
Faullant, 2008; Lee & Turban, 2001) are generally recognized as critical variables for online
consumer behaviour. More recently, researchers have also discovered the influence of social
presence or trust in fellow, typically unknown, shoppers and reviewers on shopping outcomes
(e.g., Casalo et al., 2011; Fang & Chiu, 2010; Liu & Zhang, 2010; Pan & Chiou, 2011). They
all agree that trust between strangers can foster or impede shopping outcomes. At the same time
they provide insight that purchasing online best takes place in an “atmosphere of trust”, where
consumers produce attitudes towards the vendor, the medium, the product, and other shoppers.
Various research calls have been put forward to capture additional insights into social shopping
and eWOM effectiveness – including demands for a precise measure of consumers’ trust in
online reviews and recommendations (e.g., Gefen, Benbasat & Pavlou, 2008; Urban et al.,
2009). This research makes a pioneering step towards answering these calls. However, it also
accords with other impulses for the advancement of research in the online shopping context
(Chen & Hung, 2010; Taylor & Lee, 2008).

In accordance to Straub’s (1989) claim on scale development, the development of a reliable and
valid scale will (1) bring more rigor to research and allow future scientific inquiry to be
executed in a more systematic manner, as it brings standardization to the study of eWOM trust;
(2) promote cooperative research efforts in the research community; (3) help to assure that
research results are valid and “can be trusted”, and (4) make earlier and future eWOM research
results more interpretable and clear (p. 148). For example, while one can notice several research
attempts which aim to evaluate eWOM impact, several of these contributions are contradicting,
unclear on certain aspects or need further explanation. This thesis argues that eWOM trust is a
key determinant of consumers’ responses to customer online reviews, which has been regularly
ignored in earlier surveys. Therefore, a profound investigation into individual trust differences
may bring more clarity in respect to so-far inexplicable phenomena. In addition, this research
helps to better understand eWOM’s role in a modern communication environment.

Up to the present, several trust-related constructs have been measured in the greater context of
consumers’ attitudes toward eWOM messages (e.g., eWOM usefulness, credibility,
helpfulness, quality, and attitude). However, research suffers because of a missing domain-
specific instrument which assesses the level of consumer trust in eWOM. This gap in research
is caused by several critical problems:

First, there is no agreement among scholars concerning how eWOM trust should be
conceptualized. For instance, some authors confuse eWOM trust with eWOM trustworthiness

13
(e.g., Burgess et al., 2011). The distinction between these two constructs is, however, an
important one. While trust is often described as a general willingness of the consumer to depend
on the information conveyed in the message in situations of risk, trustworthiness, in contrast, is
the set of beliefs about the information that often precedes that willingness (e.g., Mayer, Davis
& Schooman, 1995) or it is regularly regarded as only one of the aspects of trust (e.g., McKnight
& Chevany, 1996). Related to this problem is the lack of agreement on the number of types or
dimensions that trust in eWOM comprises. Existing measures of eWOM trust often do not cover
the entire scope of the trust construct (e.g., Chen & Hung, 2010). Even though various items
have been applied to measure trust in eWOM, they have to a large extent focused on measuring
the integrity (e.g., deception, manipulation, falsity of the message) or the reliability (e.g.,
keeping a promise) of eWOM. Hence, these research approaches advocate a single-dimensional
understanding of the concept. Trust literature, in contrast, provides meaningful insights that
trust is best understood as a higher-order construct and is generally more than an individual’s
belief that the trusted party is not deceptive or manipulative. Rather, trust regularly includes
beliefs in the trusted party’s abilities to fulfil promises or tasks, benevolence, and predictability,
but also the consumer’s willingness to rely on the trusted party under conditions of risk. While
the attempts to measure eWOM trust are rare, the majority of the few attempts to date do not
capture the true meaning of the construct.

Another critical issue in current eWOM trust research is the scales’ lack of established validity
(and reliability). Additionally, eWOM trust has been regularly quantified with single-item
measures (e.g., Burgess et al., 2011) or ad-hoc multiple item scales (e.g., Briggs et al., 2002;
Casalo, Flavian & Guinaliu, 2008) that have not been validated. To my knowledge, there are
no validated multi-item measurements of trust in eWOM which have been tested for internal
reliability and validity either. Most of the typically applied scales were adapted from research
in other contexts and/or were developed conceptually without the necessary empirical testing.
Specific characteristics of the eWOM context were often neglected. The fact that these
measures are derived from other domains makes it likely that eWOM-specific trust elements
remain uncovered. In contrast to most trust theorists who conceptualize trust as a construct
composed out of several dimensions, these single-item measures as well as ad-hoc scales
commonly treat trust as a single dimension. The further use of invalidated measures which
neglect the true nature of trust in online reviews and recommendations not only limits the
comparability across studies but also limits future advances in the field of eWOM research.

Practical Importance
The development of a valid and reliable scale to measure eWOM trust is also beneficial to
marketers and communication experts. In the current shopping environment, the role of the
Internet is continuing to evolve. The rise of Web 2.0 applications is only one indicator of the
fundamental changes that take place in this environment and these changes will accelerate

14
within the next years. Today’s Internet is characterized by user control and ownership of
information (e.g., consumer empowerment), collaborative networks, and social networks which
have evolved out of communities. Revolutionary phenomena have taken place such as MySpace
and Facebook, blogging, online brand and anti-brand communities, customer review and
complaint sites. As the era of social shopping (i.e., an e-commerce method that combines social
networking and shopping) characterizes the mode of purchasing, it is imperative to understand
(1) which consumers are most likely to be influenced by online customer reviews; (2) how
consumers use and evaluate eWOM information about products, brands, and firms; as well as
(3) how peer information consequently affects consumers’ attitudes and purchasing decisions
– both online and offline.

Firms have recognized the potential of WOM and make major efforts to introduce the concept
to online marketing strategy by offering instruments of active consumer participation (e.g.,
customer feedback mechanisms) on their own websites as well as the most popular social
networking sites such as Facebook. These communication efforts often parallel traditional
marketing instruments such as advertising – but increasingly, eWOM instruments take a
dominant role in firms’ communication strategy as they try to organize online communities.
However, these efforts are often based on relatively small, or no, knowledge about consumer
psychological backgrounds and the effects of such factors. In most instances, companies
introduce active consumer participation mechanisms simply because doing so seems to be the
industry’s doctrine. In contrast, firms should make such decisions based on exact knowledge
about the effectiveness of eWOM communication. As trust plays a central role in consumers’
adoption of information conveyed in fellow shoppers’ opinions and comments, its measurement
can be considered a prerequisite to effective “buzz management” (Godes & Mayzlin, 2004) and
to anticipating its effects. Here, it seems reasonable to paraphrase W. Edwards Deming: “you
can’t manage what you can’t measure”. A better knowledge concerning consumers’ trust in
eWOM will enable firms to better target the right audience with adequate integrated marketing
strategies and to better interact with their clients, as most influential eWOM communication
channels (e.g., on which platforms shall the firm foster consumer interaction?), messages and
reviewers (e.g., which people shall be furnished with additional product information or
experiences?) can be identified.

Being able to adequately measure the level of consumer trust in eWOM will also enable better
integration of a firm’s communication instruments. In the era of information overload, the
topics of “share of voice”, advertising effectiveness and accountability are matters of
heightened interest for both marketing researchers and practitioners (e.g., Breuer & Brettel,
2012; Clark, 1999; McDonald, 2010; Rust, Lemon & Zeithaml, 2004). Various studies have
strived to investigate the variables that contribute to advertising effectiveness in the
online/offline environment. Meanwhile, much is already known about the conditions that make
ads work (e.g., Tellis, Chandy, McInnis & Thaivanich, 2005). For instance, Park, Chung, and

15
Ho (2009) suggest that the matter of a person’s trust in advertising information should be
considered, as it is a key determinant of ad persuasiveness. Further support comes from various
researchers (e.g., Goldsmith & Lafferty, 2002). Another key to online advertising effectiveness
is the segmentation of the target audience (Iyer, Sobereman & Villas-Boas, 2005; Reutterer,
Mild, Natter & Taudes, 2006): Advertisements – be they from marketer-to-consumer or
consumer to consumer – are most influential when they are offered to the most open-minded
target groups. Segmenting individuals according to their trust in online advertising as well as
eWOM – the two major forms of today’s market communication – enables firms to effectively
adapt their online communication strategies and target responsive market segments with the
right communication vehicles. The research at hand makes a first contribution towards a better
understanding of online information recipients, as it provides answers to the question: which
online consumers are most likely to be susceptible to consumer- vs. marketer-released
information? How responsive are they? Or, simply speaking: does online advertising make
sense anymore in the era of consumer enforcement and collaboration? The answers provided
should make strategy enhancements possible and scope out the opportunities as well as limits
of company-issued online market information. Moreover, in-depth knowledge of consumers’
general trust in and perceptions of online inter-shopper communications allows companies to
identify the advantages and perils of eWOM for their current and prospective customers. Here,
the anticipation of positive (negative) eWOM consequences in diverse eWOM channels as well
as consumer segments seems to be especially valuable.

In addition, the new eWOM scale provides a source of reference for the measurement of other,
but conceptually related, constructs (e.g., eWOM trustworthiness), which can be used to
evaluate and better design online recommendation systems. The developed measures might also
provide the ground for sub-scales that would be of practical use in situations where use of the
full scale would be inappropriate or impractical. While the scale at hand has been developed in
the context of text-based online customer reviews, it might be also applicable in slightly
different contexts or “forms of eWOM” – like communication in online brand communities –
which would make it possible to anticipate their impact but also the participation of its
members.

Despite the latent need of practitioners to understand what causes consumers to trust (or not
trust) eWOM information and what makes them apply the conveyed information, research to-
date has not fully explored the concept nor the measurement of trust in eWOM. To overcome
this shortcoming is the target of this thesis.

Organization of the Thesis

The aim of the current chapter is to clarify the role of trust in eWOM and the current problems
research faces in measuring the construct. Chapter 2 provides a review of relevant literature on
trust, emphasizing the often divergent conceptualizations and measurement approaches across
16
various academic disciplines such as psychology, sociology, economy and business. This
chapter also discusses two frameworks that help to integrate interdisciplinary research on trust
for a common conceptual basis. In the subsequent chapter (Chapter 3), the construct of trust in
eWOM is conceptualized and the thesis’ research questions and hypotheses which guide the
assessment of the measure’s psychometric qualities are presented. Chapter 4 gives a description
of the procedure of constructing a quantitative scale for the measurement of eWOM trust and
additionally presents details of the research methodology and the executed studies. Chapter 5
is dedicated to the results of the empirical data analyses. Chapter 6 concludes this thesis by
summarizing and discussing its key findings, offering theoretical as well as practical
implications, and suggesting future avenues of research.

17
Chapter 2

Review of Literature and Theoretical Background

This chapter reviews prior research on the concept of trust across various disciplines. First, the
importance of trust for human affairs and communications is emphasized. This is followed by
a discussion of the disciplines’ different approaches to conceptualizing, defining and
operationalizing trust. Subsequently, this chapter provides a classification framework of trust
components which integrates interdisciplinary knowledge and on which further investigation
of the trust concept in the domain-specific context of eWOM is based. A discussion of the
conditions of trust that arise as well as trust-related concepts follows. In the next chapter, a
conceptualization of trust in eWOM is presented based on this chapter’s conclusions.

The Role of Trust

The notion of trust has existed as long as the history of mankind and human social interactions.
While the term itself dates back to the 13th century (e.g., Möllering, Bachmann & Lee, 2004) –
having its roots in expressions symbolizing faithfulness and loyalty – the concept of trust is
most likely to be “as old as the earliest forms of human association” (Watson, 2005). It is now
widely agreed that almost every aspect of human life seems to be based on some form of trust.
Certainly, trust is positive and essential to human beings, since it is part of the essential
ingredients of life such as love and friendship, and different kinds of meaningful relationships
seem to depend on this notion (Wang & Emurian, 2005). Trust is regarded as a central necessity
for the effective functioning of diverse levels of human relations: it determines relationships
between nations, organizations, groups, and dyads but can also be found to be a critical variable
for relationships between individuals (Butler, 1986). Trust is an important element of human
life (e.g., Gambetta, 1988; Luhmann, 1979), as it can be best described as a governance
mechanism in social, business and communication exchange relationships (e.g., Golembiewski
& McConkie, 1975; Morgan & Hunt, 1994), which are characterized by risk, uncertainty,
vulnerability, and dependence (e.g., Bradach & Eccles, 1989; Mayer et al., 1995; Mishra, 1996).

19
© Springer Fachmedien Wiesbaden 2017
W. Weitzl, Measuring Electronic Word-of-Mouth Effectiveness,
DOI 10.1007/978-3-658-15889-7_2
The critical role of trust can be underlined by saying that when there is social uncertainty
concerning how others will behave – which certainly is the case in most social interactions –
trust is a pivotal factor that determines what people expect from the situation or context. It
makes one outcome more likely than another. Trust as a mental mechanism determines what
individuals ultimately decide to do, both in social interactions (Blau 1964) as well as in business
relationships (Fukuyama 1995; Gefen et al. 2003a) and diverse aspects of life.

Over several decades, the concept has been extensively investigated by scholars from multiple
disciplines, such as psychologists, anthropologists, sociologists, economists, organizational
researchers and marketing researchers (e.g., Doney & Cannon, 1997; Larzerele & Huston,
1989; Mayer & Davis, 1999; Mayer et al., 1995). Practitioners as well as academics from almost
every discipline have commonly acknowledged the role and value of trust. The key message
that trust is critical remains the same, but the concept’s importance is explained differently
across those disciplines. For instance, Rotter (1967), an often-cited representative of the
psychological school, emphasizes the importance of trust by making the claim that the truster’s
(i.e., the trusting party) expectancy that the trustee (i.e., the trusted party) is reliable is a key
element in the formation of social relationships. Without trust, communication and interaction
among individuals become pointless. Likewise, the theorists Lewis and Weigert (1985a, p. 968)
also make an adequate recognition of the concept in the same school of thought by concluding
that trust is simply “indispensable in social relationships”. A critical consideration for
individuals in typical exchange relationships is choosing the other parties with whom an
individual is willing to interact (or communicate). Trust plays a key role in such decisions
(Fisher & Chu, 2009; McKnight, Cummings & Chervany, 1998; Stewart, 2003). Truly, the
obvious frequency with which decisions are typically made makes us assume that trust is
regularly produced and can be regarded as being fundamental to human exchanges in social
systems. In reference to this, Zucker (1986, p. 56) notes that trust is “vital for the maintenance
of cooperation in society and necessary as grounds for even the most routine, everyday
interactions”. Gundlach and Murphy (1993, p. 41) conclude that “the variable most universally
accepted as a basis of any human interaction or exchange is trust”. Hence, the concept is critical
from complex, one-of-a-kind to very simplistic, everyday decisions.

The key role of trust for business and social exchange is evidenced in a number of ways and
the positive effects are omnipresent. Trust enables people to live in risky and uncertain
situations (e.g., Deutsch 1958; Deutsch 1962; Mayer et al. 1995); it provides the means to
decrease social complexity in an overwhelming, complex world by reducing the set of options
a person has to consider in a given situation (e.g., Barber 1983; Lewis and Weigert 1985a;
Luhmann 1979); it allows us freedom from rational prediction (Aikgen & Bousch, 2006) and
provides us with confidence in social interactions (Lewis and Weigert 1985a). Further, trust is
a necessity to positive relational exchange in various settings (e.g., Lewis and Weigert 1985a)
as the degree of trust is critical to how one interacts with others in a given environment (e.g.,

20
Golembiewski and McConkie 1975); trust reduces harmful conflict; and promotes effective
responses to crises. Trust can also be described as a kind of social capital that enables
cooperative behaviour and coordination between people (Deutsch, 1973; Gambetta, 1988;
Misztal, 1996). In this vein, trust is important in various commercial activities and relationships,
such as those between a supervisor and his/her subordinate or in buyer-seller relationships.
Accordingly, trust has been shown to have a substantial effect on business relationships in
general (Dasgupta, 1988; Fukuyama, 1995; Gambetta, 1988; Gulati, 1995); trust is an
alternative solution to control, besides price and authority (Creed & Miles, 1996); it reduces the
need for extensive negotiations, enforced regulation, as well as tight and costly organizational
control (Fukuyama, 1995); trust is a cricital ingredient of effective working relationships
(Gabarro, 1978). In addition, trust cultivates long-term orientation (Fukuyama, 1995; Ganesan,
1994; Morgan & Hunt, 1994); is key to successful transactions and long-term relationships
Koehn, 1996); trust creates commitment (Moorman et al., 1992; Morgan & Hunt, 1994); and
increases a person’s acceptance of interdependence (Schurr & Ozanne, 1985; Zand, 1972).
Trust becomes a critical success variable especially during periods of uncertainty because of an
organizational crisis (Mishra, 1996). The presence of trust is also able to reduce perceived risk
(Fukuyama, 1995) as well as transaction costs (Fukuyama, 1995; Williamson, 1985), and is to
some degree important to almost any contractual agreement due to the possible opportunistic
behaviour of the other party (Williamson, 1985). Trust can reduce uncertainty created by other
people or artefacts and is therefore generally essential in commerce (Blau, 1964). Analogously,
trust also heightens the willingness of the truster to interact with the trusted party in situations
where the trusting party may be vulnerable to the other’s actions (Blau, 1964; Luhmann, 1979).
It is a significant element leading to actual and anticipated purchases (Doney & Cannon, 1997).
Research has shown that trust affects consumers’ fears of unreliability and risks of being
cheated (Jones & Leonard, 2008). Without such a mental mechanism, any interaction would be
paralyzed as the risks of failure would be perceived as overwhelming. In essence, all these
academic contributions demonstrate that trust is shown to play a major and indispensable role
in determining individuals’ diverse forms of intentions and behaviours. Practitioners’
acknowledgement of the value of trust is as pronounced as it is among scholars (Bartlett &
Ghoshal, 1995; Peters, 1992). For example, in their book on partnering in organization
Rackman et al. (1996, p. 75) have quoted a business executive on the importance of trust, citing:
“[…] there are a lot of issues in partnering, […] but trust is truly the key. Everything else has
to be based on it. Without trust, there is no basis for partnering. It’s the bottom line […]”.

One idea to explain the prominence of trust in exchange relationships is social exchange theory
(SET) (e.g., Homans, 1961; Kelley & Thibaut, 1978). This theory is based on the fundamental
assumption that people typically form their exchange relationships on the basis of trust. In its
essence, SET views social interactions in a similar manner to economic exchange theory. That
is, they are composed of costs paid and rewards received. Similarly to economic theory, people
are assumed to only take part in interactions in situations where their outcomes are expected to
21
be satisfactory. This essentially means that the subjectively perceived rewards have to exceed
the subjective costs of the interaction (Blau, 1964; Homans, 1950) – or at least satisfy their
expectations and exceed their alternative investments (Thibaut & Kelley, 1959). Here, however,
SET contrasts social exchange with its economic counterpart by arguing that the former
typically deals with situations that are characterized by the absence of explicit or detailed
contracts which bind the involved parties. Hence, because rewards cannot be guaranteed in
social exchange, trust is essential and determines people’s expectations of the personal reward
from the relationship (Blau, 1964; Lewis & Weigert, 1985a; Luhmann, 1979). However, the
same has been found to be true for economic relationships, as various researchers emphasize
that trust, in general, enables the parties to form appropriate favourable expectations about the
party they are doing business with (Gefen, 2002a; Luhmann, 1979; Williamson, 1985). Various
scholars argue that even in the presence of a contractual framework, the parties have to “trust
the trust” of the other in a proper business order. Contracts typically do not completely
safeguard interaction outcomes and parties sometimes have to make a serious effort to enforce
the contractually ensured outcomes with penalties, if they are successful at all. So even when
there are contracts, some kind of interaction risk is still unavoidable. Hence, in both economic
and social relationship situations, trust increases the perceived certainty concerning other
people’s expected behaviour (Luhmann, 1979; Zand, 1972) and reduces the fear of being
exploited (Zand, 1972).

Luhmann (1979) advances a similar theory and strives to enlighten our understanding of the
importance and functionality of trust in human affairs. In essence, he refers to everyone’s need
to control and anticipate – or at least to better understand – the social environment with which
individual interacts as a central necessity of human behaviour. Specifically, all people have a
general desire to “know” before a social interaction takes place (i) how their behaviour will
influence the behaviour of others, and (ii) how the behaviour of others will consequently affect
themselves (Gefen & Straub, 2004). The sociologist states that in social relations, interaction
partners are best described as free agents whose behaviour can neither be entirely controlled
nor predicted. This leads to an inherent uncertainty of relationships. Consequently, the social
environment is so overwhelmingly complex that without a mechanism striving to minimize this
social complexity, only short-term interactions would be possible at the most. Hence, people
are forced to develop certain mechanisms or heuristics to effectively reduce this complexity.
According to Luhmann (1979), one of these mechanisms is trust and people regularly use this
mental shortcut in order to predict the behaviour of other individuals and organizations. People
may also apply other mechanisms, such as pre-determined rules and regulations resulting from
laws, contracts or societal standards, which likewise strive to enable complex and long-term
transactions. However, when, as is usually the case, laws, rules and regulations are not enough
(because they do not automatically safeguard interaction partners from harm), individuals often
reduce (unwanted) social complexity by “assuming away undesirable, yet possible, behaviours
of others” (Gefen, 2002b). The most critical factor that prohibits social interaction is the
22
participant’s fear of potential negative outcomes. With trust, the probability of those negative
outcomes is reduced to a level the person can handle. Trust offers people multiple beneficial
consequences as – in additional to social complexity – it also reduces general feelings of
uncertainty as well as risk. According to Luhmann (1979), the favourable presumption about
the acceptable future actions of others is the essence of trust (Gefen, 2002b).

Luhmann’s theory had been stressed by various scholars, amongst them online consumer
researchers, as they regularly recognize that the sociologist’s work provides a valuable
framework for explaining why people often refrain from online shopping. According to these
authors, the online shopping environment is a less verifiable and less controllable business
environment compared to traditional retailing (Gefen, 2000; Gefen & Straub, 2004; Reichheld
& Schefter, 2000). This generally leads to an increased transaction complexity and a heightened
uncertainty of conditions (Lee & Turban, 2001; McKnight & Chervany, 2001). Furthermore,
online interactions are not only characterized by uncertainty, but in addition by anonymity of
the interaction partner, and potential opportunism. This makes risk and trust crucial elements
in online shopping and interactions. Trust – be it in the Internet as a shopping institution or in
the seller itself – is regarded as being one of the most efficient mechanisms to reduce inherent
online risks and elicit positive expectations of the outcome of the online transaction (Gefen,
Karahanna & Straub, 2003; Jarvenpaa & Tractinsky, 1999; Kaplan & Nieschwietz, 2003;
McKnight & Chervany, 2002; Mishra, 1996; Pavlou, 2003). Trust is a facilitator of online
interactions.

In a nutshell, trust is a critical concept that determines the type and nature of the social and
business order (Blau, 1964; Fukuyama, 1995; Lewis & Weigert, 1985a). People have to produce
trust in order to participate in many activities with others and would refrain from social
interactions (e.g., communication) with other people whom they do not trust (Blau, 1964;
Luhmann, 1979). Zand (1972, p. 229) adds to the concept’s role by stating that trust “is the
salient factor in determining the effectiveness of many relations”. Konovosky and Pugh (1994)
as well as Schurr and Ozanne (1985) more forcefully note that trust is nothing more than the
prime motivator of behaviour in general. These opinions support the claim that individuals
should in general develop more trust. However, this should only take place under conditions
which give them adequate reasons to do so.

The Concept of Trust

Overview

Because trust is considered so vital and important for human life, the construct has been the
object of scientific investigation over several decades. In fact, this interest dates back to the
1950s (Corritore, Kracher & Wiedenbeck, 2001; Corritore et al., 2003) and has meanwhile
resulted in an extensive literature string which can be attributed to the trust concept itself, but

23
also deals with the nature of trust relationships and trust-related concepts (e.g., endorser
credibility). During this time, many intellectual disciplines have researched the concept and
each discipline has investigated trust from its own perspective and in varying contexts,
advocating their own definitions, concepts as well as findings. Calder (1977) explains the vast
scientific interest by suggesting that the study of topics such as trust – which can be said to
“belong to the world of everyday explanation” (p.182) – inevitably leads to a proliferation of
approaches to explaining them. The contributions’ richness is essentially caused by the fact that
such terms (or mental constructs) typically involve numerous connotations. Likewise, trust has
received a great amout of interest from scholars in multiple academic disciplines, such as
sociology (e.g., Lewis & Weigert, 1985), psychology and especially social psychology (e.g.,
Deutsch, 1960; Lewicki & Bunker, 1995), economics (e.g., Dasgupta, 1988; Williamson,
1993), as well as management (e.g., Garbarino & Johnson, 1999; Mayer et al., 1995; McAllister,
1995), organizational behaviour (e.g., Cummings & Bromiley, 1996; Zaheer, McEvily &
Perrone, 1998; Zucker, 1986), and marketing (e.g., Anderson & Weitz, 1989; Dwyer & Oh,
1987; Ganesan, 1994; Moorman et al., 1993). For these scholars, trust represents an appealing
mental construct that is capable of explaining otherwise irrational outcomes, behaviours or
processes.

When looking at the diversity of these disciplines, one could assume that the concept of trust is
a well-understood and established concept, free of ambiguity. And it’s true that current
scientific work is able to build on a quite robust fundament of trust-knowledge which helps
researchers to understand the essence or nature of the concept. Many social phenomena can be
explained as a result of this knowledge. However, this knowledge – when examined in detail –
is also quite fragmented. It appears that one is not able to speak of a single trust construct
because of heterogeneous meanings, but a set of concepts. As various disciplines agree on the
key role of trust in the conduct of human affairs – while the concept has especially critical
importance in social relationships (Hosmer, 1995), each discipline was always striving to
contribute to trust research from their own perspective. This led to unique insights into the
nature of trust, its definitions, and the process by which it regularly develops. Here, trust has
often been defined in significantly different ways naturally resulting in a great number of
divergent conceptualizations and operationalizations. The outcome is a multi-dimensional
family of unique concepts that share only some communality and are therefore often not directly
applicable to other disciplines due to their narrow and contextual understanding of trust (Doney
& Cannon, 1997). In fact, sometimes concepts are not even applicable to different research
areas within the very same scientific discipline. Research hence suffers from a lack of general
agreement and focus of effort (Lewicki & Bunker, 1995).

The fragmented knowledge on trust becomes more evident when one compares the disciplines’
divergent views of trust. For instance, psychologists generally frame their assessment of trust
in terms of the attributes of the trustee (or trusted object) perceived by the truster (the trusting

24
party). Additionally, research in the field of psychology focuses on a number of internal
cognitions that personal attributes yield. Traditionally, personality psychologists have regarded
the concept of trust as an individual’s personal trait, often called a “propensity to trust”, that is
deeply rooted in the personality and has its origins in the individual’s early psychological
development (e.g., Rotter, 1967; Worchel, 1979). Social psychologists and management
researchers, on the other hand, often define trust as an expectation about the behaviour of other
individuals in transactions, which is affected by contextual factors (e.g., Lewicki & Bunker,
1996; Mayer et al., 1995). But interactions between generalized trust (personality) and
situational trust (environment) can also take place (Bandura, 1977a; Bowers, 1973).
Sociologists hold a different view, in that they have considered trust as a social good which is
embedded in social relationships among people (e.g., Granovetter, 1985) or institutions (e.g.,
Zucker, 1986). Among economists, it is common to regard trust as either calculative
(Williamson, 1993) or as an institutional phenomenon that is able to reduce the anxiety and
uncertainty associated with transactions (North, 1980).

Within business schools there are various approaches to the study of trust across different
domains. For instance, management researchers’ interest in the topic began in the mid-1980s.
From this perspective, trust has been examined as a factor that facilitates cooperative behaviour
among individuals, groups, and organizations (e.g., Gambetta, 1988; Mayer et al., 1995;
McAllister, 1995). Trust as a concept has also attracted major interest in the marketing
community and has found its way to the core of modern marketing understanding. The
dominant paradigm of relationship marketing would not be possible without the concept of
trust. Here, marketing scholars have especially focused on trust in a particular interaction
partner (i.e., interpersonal trust) and on its consequences in buyer-seller relationships – widely
drawing on trust constructs and theories developed in other disciplines. Marketing
understanding is therefore not stand-alone research but is deeply rooted in interdisciplinary
research. The role of trust in marketing has been studied in the context of distribution channels
(e.g., Anderson & Narus, 1990), between industrial buyers and sellers, as well as within the
context of retailing (e.g., Garbarino & Johnson, 1999) and on the consumer-level, trust in the
salesperson (e.g., Doney & Cannon, 1997), trust in the brand (e.g., Chaudhuri & Holbrook,
2001), trust in the vendor (e.g., Sirdeshmukh, Singh & Sabol, 2002), as well as trust in e-
vendors (to name but a few). Marketing knowledge about consumer behaviour essentially
mirrors the understanding and assumptions of psychology in respect of a person’s environment,
mental structures, and their consequences. Contributions in this field mainly stem from two
literature strings, namely research on relationship marketing (e.g., Ganesan & Hess, 1997;
Garbarino & Johnson, 1999) and research on e-commerce/online shopping (e.g., harris &
Goode, 2004; Jarvenpaa et al., 2000; McKnight, Choudhury & Kacmar, 2002b). The latter topic
surely led to a revival of the concept in the last decade. Within the marketing field, trust is often
regarded as having a pivotal role in holding relationships between buyers and sellers
successfully together (Berry, 1995; Dwyer, Schurr & Oh, 1987; Morgan & Hunt, 1994) and
25
enabling positive business outcomes for both parties. This stream of research is valuable, as it
demonstrates the need for domain-specific investigation of the nature and role of trust.

Simultaneously, trust has often been defined and conceptualized according to the features of a
particular context. For example, the construct of trust has been investigated in the context of
personal and romantic relationships (Rempel, Holmes & Zanna, 1985), organizations (Mayer
et al., 1995), public administration (McLain & Hackman, 1999), politics (Barber, 1983),
bargaining (Schurr & Ozanne, 1985), and buyer-seller relationships (Doney & Cannon, 1997;
Ganesan & Hess, 1997).

Although these multi-disciplinary and multi-contextual perspectives have both created a

recognizable breadth that strengthens the trust literature and our domain-specific knowledge
(Bigley & Pearce, 1998; Rosseau, Sitkin, Burt & Camerer, 1998), it generally also contributes
to a latent confusion among researchers, leaving essential questions unanswered: what is trust
really? and: can we identify something that we can generally label “trust” or are there diverse
forms meaning that we have to construct the meaning within each domain? If trust is considered
to be a socially-defined construct (Searle, 1995), a person’s understanding of the concept may
be different from others. This is true for our everyday understanding but naturally also in
sciences. Therefore, some researchers regularly define trust as a belief or expectation (e.g.,
Anderson & Weitz, 1989; Crosby, Evans & Crowles, 1990; Morgan & Hunt, 1994), while
others define trust as an attitude (e.g., Hallen & Sandstrom, 1991; Jian, Bisantz & Drury, 2000).
Others, still, view trust as a purely behavioural intention (e.g., Mayer et al., 1995; McKnight et
al., 1998; Moorman et al., 1993; Rosseau et al., 1998) or an internal process, similar to choosing,
judging, or preferring (Lewis & Weigert, 1985). A fraction of trust scholars regard trust as a
natural feeling or faith. Others view trust as being synonymous to trustworthiness, emphasizing
trust in the context of personal characteristics that inspires or causes positive expectations on
the part of other individuals (Butler & Cantrell, 1984; McKnight et al., 1998). Among those
characteristics of the trusted party that imply trust, scholars often refer to a set of attributes
including concepts of fairness, goodness, strength, ability, benevolence, honesty, and
predictability (Gefen, Rao & Tractinsky, 2003; Mayer et al., 1995; McKnight & Chervany,
2001). Still others view trust as a disposition or a constant mental state. Accordingly, trust is a
facet of personality that develops in the early stages of life and remains stable (or relatively
stable) through the lifespan (e.g., Ganesan, 1994; Moorman et al., 1993; rotter, 1966; Webb &
Worchel, 1986). For others, trust is a choice (Alpen, 1997). Finally, some researchers regard
trust as a synonym for human cooperation or risk-taking (Kee & Knox, 1970; Zand, 1972) and
often attempted to measure the construct by using cooperative behaviours in dilemma games
(Deutsch, 1958, 1960). In a similar manner, trust has been discussed to be a mental structure
having cognitive (i.e., trust is a matter of opinion or a prediction that things will turn out in a
certain way or that people will behave in a certain way – trust as a matter of belief or

26
knowledge), affective (i.e., trust is a matter of feeling or mood) (e.g., Lewis & Weigert, 1985;
McAllister, 1995), and/or conative (i.e., trust is a matter of choice, motivation or will) aspects.

According to Kuhn (1962), scientific endeavours should, over time, lead to some level of
consensus concerning the topic of interest. But while trust literature suggests common
agreement on the importance and the positive effects of trust, the same literature stream has so
far failed to find a needed consensus on what trust means (Kee & Knox, 1970; Wang &
Emurian, 2005; Yamagishi & Yamagishi, 1994).

Literally dozens of definitions of trust coexist today. Some of these definitions overlap, but
more often each single definition offers an assertion of a different aspect of trust (Grabner-
Kraeuter & Kaluscha, 2003). As a result, investigations into the nature of trust are often seen
as quite uncomfortable and complex. Giffin (1967, p. 104) says that trust “has been viewed as
a somewhat mystical and intangible factor, probably defying careful definition”. Being aware
of the confusing conceptual difficulties, various trust researchers refrain from specifically
defining the concept at all (e.g., Granovetter, 1985; Jarvenpaa et al., 2000), while other scholars
come to the conclusion that the concept is almost impossible or elusive to define (e.g.,
Gambetta, 1988; Yamagishi & Yamagishi, 1994). The difficulty in defining trust is also
demonstrated by numerous statements of researchers. For example, some scholars find existing
definitions contradictory and confusing (e.g., Gulati & Sytch, 2008). In 1985 Lewis and Weigert
(Lewis & Weigert, 1985, p. 464) state that “trust is one of those primary forms of experience
that elude simple definition”. The authors further remark that trust definitions are characterized
by a “conceptual confusion” (Lewis & Weigert, 1985, p. 975). Similarly, Barber (1983, p. 1)
speaks of a “conceptual morass” and Golembiewski and McConkie (1975, p. 131) express their
belief that the study of trust is essentially “a paradox”. Zucker (1986, p. 56) concludes that
“recognition of the importance of trust has led to concern with defining the concept, but the
definitions proposed unfortunately have little in common”. Shapiro (1987, p. 624) agrees with
the general tenor and adds that the considerable attention to trust research results in “a confusing
potpourri of definitions applied to a host of units and levels of analysis”. More recently, Hosmer
(1995) consistently states that the different disciplines agree on the quality of trust in the
conduct of human affairs, but on the other hand, there also appears to be a similar lack of
agreement on a suitable scholarly definition of the concept. In a similar manner, Luhmann
(1980, p. 8) comments on the unintegrated and incomplete state of trust research that “diverse
conceptualizations of interpersonal trust coexist […] with intense convictions that the various
somethings described are central in all of human life”. Over several decades, trust researchers
have called for conceptual clarity (Golembiewski & McConkie, 1975), however, with little
success. Today, the problem with trust as a concept remains the same, limiting future research
because of the absence of a universally accepted definition (Barber, 1983; Beldad et al., 2010;
Bhattachary & Devinney, 1998; Butler, 1991; Das & Teng, 1998; Husted, 1998; Kee & Knox,
1970; Rosseau et al., 1998). This problem particularly applies also to the online shopping and

27
the communication domain in marketing research (Bramall et al., 2004; McKnight et al.,
2002a). This is mirrored by the statement made by Keen and colleagues (Keen, Balance, Chan
& Schrump, 1999, p. 4-5), who noted that “[…] the basic conclusion in all these fields [is that]
trust is becoming more and more important, but we still can’t really say what it exactly is”.
Naturally, past research on eWOM is also affected by the problems described above (Briggs et
al., 2002; Burgess et al., 2011), having the consequence that various scholars simply refrain
from precisely defining the concept (e.g., Pan & Chiou, 2011) or understanding only a
fragmented element of trust as being the same as overall eWOM trust (e.g., Smith et al., 2005;
Smith, 2002). This often leads to a critical confusion with other concepts, and consequently to
a poor and imprecise measurement of trust.

Therefore, an interesting question that needs answering is why the term “trust” elicits either
confusion or reluctance to define the concept under investigation. The existence of the multitude
of trust definitions is likely to be attributable to several reasons. First, “trust” as a mental
concept is quite a vague term and accordingly an abstract construct. Or, as Williamson (1993,
p. 493) puts it, “[…] trust is a term with many meanings”. Trust shares this characteristic with
many other everyday language terms. A look at the definitions from three different dictionaries
shows the massive breadth of everyday usage of the word trust and provides evidence that the
term is really conceptually “massive” in terms of the different meanings it conveys (see Table
1). From earlier dictionary reviews it can be learned that trust has more (and also longer)
definitions than the terms Mayer, Davis and Schoorman (1995), for instance, use to
discriminate trust from related concepts such as “cooperation”, “confidence”, and
“predictability”. However, trust has, on the other hand, fewer and shorter definitions than other
also abstract and nebulous-appearing concepts such as “like” and “love”, which are both
seemingly very broad and difficult to define (Brehm, 1992). Earlier, Taylor (1989) conducted
a review and states that if someone had to compare the length of trust definitions with those of
other concepts, “[they] would most likely conclude – and quite correctly – that trust is not a
single or simple concept”. Additionally, from the dictionary definitions it is evident that similar
words appear repeatedly, including words such as “reliance”, “confidence”, “integrity”, “faith”,
“risk”, and “hope”. By referring to these words, one can conclude that “trust” is often used
interchangeably with related but conceptually different concepts in everyday usage (e.g.,
confidence, credibility). Hence, our usage of the word is often imprecise. However, this
problem can also be recognized in various scientific research approaches where a variety of
authors were inclined to use the trust concept interchangeably with related but different
constructs. The boundaries of the concept hence remain unclear.

28
Table 1: Dictionary Definitions of Trust

Merriam-Webster The American Heritage Oxford Dictionary (1995, p.

Dictionary (1996, p. 1116) Dictionary (1992, p. 1498)
1920)
Noun: 1. Assured reliance on 1. Firm reliance on 1. A firm belief in the
the character, the integrity, reliability or truth or
ability, strength, or ability, or strength of a person or a
truth of someone or character of a thing.
something. person or a thing. 2. The state of being relied
2. Someone or 2. The condition and on.
something in which resulting 3. A confident expectation.
confidence is obligation of
4. Reliance on truth of a
placed. having confidence
statement etc. without
3. Dependence on placed in one.
examination.
something in the 3. Reliance on
5. Confidence placed in a
future. something in the
person
future
4. Reliance on the
intention and
ability.
Verb: 1. To place confidence. 1. To have or place 1. Believe in the reliability,
2. To rely on the reliance; depend. truth, or ability of someone
truthfulness or 2. To be confident; or something; believe.
accuracy of hope. 2. To rely on the character or
someone or 3. To have or place behaviour of someone or
something. confidence in; something.
3. To believe in depend on. 3. Have faith or confidence or
someone or 4. To expect with hope that a thing will take
something. assurance; place
4. To commit or assume. 4. To place reliance in
consign with trust or 5. To believe 5. To accept something (an
confidence. assertion, claim, etc.)
5. To permit to remain without evidence or
or go somewhere or investigation
to do something
without fear of
consequences.
6. To risk; to venture
confidently
Synonyms: Confidence; credence; Faith, confidence; n.a.
faith; belief; assurance reliance; dependence
Antonyms: Distrust; mistrust n.a. n.a.

The second reason for the miserable conceptual confusion comes from the fact that trust exists
at multiple levels of human affairs. Nissenbach (2001, p. 104) recognizes that the object of trust
can vary dramatically as he states, “Trust is an extraordinarily rich concept, covering a variety
of objects. One can trust (or distrust) persons, institutions, movements, information, deities,
psychological things, systems, and more”. Psychologists typically concentrate their efforts on
the individual (intrapersonal, interpersonal) and occasionally on the group level. In contrast,

29
sociologists are used to referring to groups or social entities which may be families, neighbours
or societies. Organizational researchers investigate trust as it exists within an individual, as a
subordinate’s trust or distrust in his/her manager (Gabarro, 1978). Further, the researchers in
the same field also recognize that trust also exists within groups or organizations (Ouchi, 1981)
as well as between organizations (i.e., on an institutional level), such as in the establishment
and maintenance of strategic alliances (Barber, 1983). Finally, economists advance trust
research in the context of individuals or larger firms. Recently, the concept has been broadened
as a result of investigation into people-to-computing systems trust by researchers in computer
and communication sciences (Kantsperger & Kunz, 2010).

Third, disciplines disagree over whether trust incorporates cognitive, emotional, and
behavioural dimensions or not. It seems true that the kind of relationship is likely to determine
the ingredients of trust and their relative importance. Fourth, contemporary trust research is
primarily driven by empirical studies in which researchers tend to develop trust
conceptualizations that are quite specific and narrow in order to fit this type of research
(McKnight & Chervany, 2001). Such approaches typically target the identification of the
attributes of trust rather than defining trust itself (Tan & Sutherland, 2004).

Various scholars have recognized this confusing proliferation of trust types, as well as
perspectives, and have attempted to synthesize the findings in a composite trust definition or to
categorize interdisciplinary research on the topic (e.g., Doney & Cannon, 1997; Kee & Knox,
1970; Mayer et al., 1995; Rosseau et al., 1998). For example, in their broad view on trust, Keen
and Knox (1970) defined the concept as a set of five trust-related constructs; namely,
dispositional factors, contextual factors, perceptions of the interaction partner, subjective trust,
and behavioural trust. Mayer et al. (1995) introduce an integrated model that defines trust as
the willingness of a truster to be vulnerable to the actions of a trustee (i.e., a motivation) based
on the expectation (i.e., cognitive beliefs) that the trustee will carry out a particular action. They
further recognized the truster’s general disposition to trust as a major determinant of trust. Their
framework therefore recognizes personality as well as environmental aspects. In a similar
tradition, McKnight and his colleagues (2002a) propose a trust model that integrates trusting
intentions, trusting beliefs, disposition to trust as well as institutional trust, which can be mostly
attributed to the social psychologists’ and sociologists’ trust perspective. More recently, Gefen,
Karahanna and Straub (2003) summarize the conceptualizations of trust by reviewing prior
definitions along with measures used and suggest that trust can be operationalized as (1) a set
of distinct beliefs about the trustee dealing primarily with the integrity, benevolence, and ability
of another party (e.g., Doney & Cannon, 1997; Ganesan, 1994); (2) a general belief (e.g., Gefen,
2000; Hosmer, 1995) or “trusting intention” that another party can be trusted (e.g., McKnight
et al., 1998), or “the willingness of a party to be vulnerable to the actions of another” (Mayer et
al., 1995, p. 712); (3) “feelings of confidence and security in the caring response of the other
party” (Rempel et al., 1985, p. 96); or (4) a combination of these factors.

30
Various valuable parallel research attempts are striving to profoundly categorize trust in
general. For instance, Hosmer (1995) evaluates trust from the perspectives of (1) individual
expectations; (2) interpersonal relations; (3) economic transactions; (4) social structures; (5)
organizational theory; and (6) ethical principles. By borrowing from Worchel’s (1979) earlier
contribution, Lewicki and Bunker (1996) categorize academic works on trust into three major
perspectives: (1) the view of personality theorists who affirm that trust can be regarded as an
individual difference; (2) the perspective of social psychologists who see trust as an individual’s
expectation derived from interpretation of the social environment; and (3) the views of
sociologists and economists. For the latter, trust is an institutional phenomenon. More recently,
Beldad, Jong and Steehoulder’s (2010) review classifies trust contributions in (a) works that
emphasize trust as an individual feature; (b) articles that view trust as an expectation; (c)
scientific inputs which regard trust as the acceptance of and exposure to vulnerability; and (d)
works of sociologists and economists who typically understand trust as an institutional
mechanism.

In their classical works on measurement and validity, Cronbach and Mehl (1955), as well as
Loevinger (1957), likewise advocate the importance of theory in measurement. Analogous
requirements have been also proposed by other scholars (e.g., Netemeyer et al., 2003). Kaplan
(1964) states that effective conceptualization is critical to the cusses of construct validation and
substantive research. Therefore, an essential step towards an effective measurement of eWOM
trust is to profoundly understand the meaning of the concept and to derive an appropriate
conceptualization. Such an approach helps to exactly delineate what is included and what is
excluded from the domain of interest. A comprehensive literature review should build an
adequate basis for a precise conceptualization of the construct, its boundaries and content.
Further, this approach is likely to uncover potential antecedents and consequences which enable
the concept to be embedded into a nomological framework and to profile eWOM trusters. The
author proposes that – to be most valuable – a conceptualization of eWOM trust should build
on the multi-dimensional and diversified family of trust concepts and, hence, should span across
disciplines. The roots of the scientific investigation of the trust concept can be found long before
the rise of the Internet or eWOM communication. Scholars made valuable attempts to clarify
the conceptual basis of the parent construct. By following the proposition of Corritore et al.
(2003), this thesis grounds its contribution on coexisting trust research in the offline as well as
online world. In this respect, many offline trust findings appear to be applicable to an online
environment, since offline and online situations typically have much in common. One obvious
communality is, for example, the existence of social exchange. Here, the same social rules of
interaction seem to function in both the offline and online environment. Humans tend to solve
social problems wherever they arise in a similar way. Further, since trust is able to mitigate
perceptions of risk, fear and complexity in conversations in the offline environment, it is likely
that it can perform the same function in the online environment (Corritore et al., 2003). The
mechanism remains the same. Therefore, offline trust research with its rich understanding of
31
the human mind and nature is regarded to be essential for the aim of defining the eWOM trust
concept as clearly as possible.

Since the concept of trust has its roots in different disciplines, the literature on the topic has
become – as demonstrated – very manifold over the years. Nevertheless, each of the proposed
conceptualizations seems to add some insight and understanding. Subsequently, the construct
of trust will be reviewed within the context of (1) philosophical thinking, (2) personality
theories, (3) societal perspectives, (4) interpersonal relationships, and (5) economic/business
transactions. This categorization originates from the earlier classifications advanced by Hosmer
(1995), Lewicki and Bunker (1995), and Soh (2007). This thesis follows their classification
because this integrated view appears to reflect the characteristics of different types of trust and
specific perspectives of each discipline as well. Here, the functional perspective of trust is
emphasized. Among ancient and contemporary philosophers, trust is viewed as an ethical
principle. Personality theorists commonly agree that trust is a personality trait or an individual
difference. Sociologists regard trust as an institutional phenomenon. On the other hand, trust
has been conceptualized by social psychologists and organizational science researchers as an
individual’s expectations of an interpersonal relationship. Finally, economists and marketing
scholars are particularly interested in the role of trust in economic transactions. Grounding our
understanding in the originating literature helps to define the eWOM-specific trust concept as
clearly as possible and to evaluate its conceptual framework for logical validity and consistency.

The Disciplines’ Conceptualizations and Measurement of Trust

The Philosophers’ Perspective

Philosophical thoughts on the nature and conditions of trust can be traced back to the ancient
Greek scholars. Early philosophers strove for a better understanding of trust as a characteristic
of human nature (Bailey, 2002). Among these scholars – Plato (424-348 BC) being one of the
most prominent representatives – it was widely accepted that individuals rely on others only if
they are confident that the other's fear of detection and punishment prevents them from breaking
the law and doing evil for their own self-interest. The picture drawn seems to be reasonable.
This perspective also reveals an important characteristic of trust: in the situation when people
trust others, they are confidently relying on them to take care of something which is valuable
to them, but can be stolen or harmed by others if they wish. This means that if we trust, we
make ourselves vulnerable at the same moment. However, we do so in the confidence that the
other party will not exploit this vulnerability, but instead show a cooperative behaviour and
actively take care of what makes us vulnerable (Bailey, 2002). This general understanding of
trust was not forgotten during history. It now represents an important benchmark for a
recognizable number of contemporary trust researchers. The ancient philosophers additionally
also made the first investigations into the antecedents of trust, and identified love and sympathy

32
as elements that elicit more trust, but they also stress the determinant role of cooperation and
peace among human beings (Wang & Emurian, 2005).

Modern scholars – such as the moral philosopher and Hume-fellow Annette Baier - advanced
the discipline’s knowledge on trust by focusing on interpersonal trust (i.e., trust between
individuals) and the morality of trust relationships. One central philosophical definition of trust
was pioneered by Baier (1994), as she views trust as an individual’s accepted vulnerability to
another’s opportunity to harm the trusting person. Political philosophy can be regarded as an
area of fruitful research, as members of this school have explored trust in terms of social values
and benefits. For example, one prominent member of this school is Francis Fukuyama, who has
not only recognized the role of interpersonal trust, but expanded his perspective to include
issues of culture and society. In his politically oriented book entitled Trust, the social virtues
and the creation of prosperity, Fukuyama (1995, p. 26) defines trust as “the expectation that
arises within a community of regular, honest, and cooperative behaviour, based on commonly
shared norms, on the part of other members of that community”. He continues with the assertion
that these norms can be “(…) about deep 'value' questions like the nature of God or justice, but
they also encompass secular norms like professional standards and codes of behaviour”.
According to Fukuyama, trust is regarded as a mechanism which engenders spontaneous
sociability leading to shared values (or social norms) that are confirmed by the members of a
social group (Wang & Emurian, 2005). These insights make it reasonable to argue that trust is
regularly produced or is, in fact, a necessity for online interaction as it takes place in online
review communities. The absence of trust would automatically lead to a loss of shared values,
norms, and ethical principles, making information exchange meaningless.

The Personality Psychologists’ Perspective

Personality theorists are used to studying trust as a belief, an expectancy, or feeling that is
deeply rooted in an individual’s personality (Cheung & Lee, 2006). The personality
psychologists’ research interest is dominated by the fundamental question of why some people
are just more likely to trust than others. According to several authors, the answer lies in
substantial variations in the nature of individuals’ personalities – or more precisely in the
peoples’ general willingness to trust others. In general, personality refers an individual’s
general style of dealing with the world and in particular with other people – that is the social
environment. More precisely, the concept considers individual differences in characteristic
patterns of thinking, feeling, and behaving (Kazdin, 2000). Trust is here regarded as a constant
personality trait or characteristic (such as sociability or open-mindedness) and hence element
of an individual’s personality characteristic that influences a person’s patterns of interaction
with the world at large in a variety of situations and across time. Within this context, trust is
often defined as a generalized tendency of human beings to believe in the trustworthiness of
others (e.g., Das & Teng, 2004), or in other words, trust is a person’s generalized expectancy

33
that behaviours or words of other people can be relied on. Here, the object of trust is not a
specific other but other people in general (i.e., the public). Hence, dispositional trust means
primarily the general propensity to have the intention to depend on others (Mayer et al., 1995)
and this propensity will influence how much trust one has in a trustee prior to data on that
particular party being available. Bigley and Pearce (1998, p. 410) comment that “factors exist
within individuals that predispose them to trust or distrust others, whom they do not know”.
Because trust is regarded as a trait, it is assumed to be independent across situations and persons,
being a stable within-party factor which encompasses individual characteristics of the truster
that colours the interpretation of situations and actors in situations. It is assumed that people
develop trust in varying degrees depending mainly on their personality types but also on
personal experiences, prior socializations, and cultural backgrounds (Hofstede, 1980). An
example of an extreme case of this kind of trust is what is commonly called “blind trust”. Some
people can be observed to repeatedly trust in situations which the majority of people would
agree do not warrant trust. Here, trust seems to be a synonym for pure faith. On the other hand,
others are unwilling to trust in most situations, regardless of the circumstances that would
support doing so (Mayer et al., 1995). These people hold a very sceptical world view.

Rotter (1967) was one of the first scholars to discuss trust as a form of personality. More
specifically, he defined what he termed interpersonal trust as a generalized expectancy (see
also Rosenberg, 1956). By following the trust typology proposed later by McKnight and
Chevany (1996), this type of trust is commonly referred to as dispositional trust (see also
Kramer, 1999). Several other authors have referred to this habitual pattern or personality-based
form of trust by using different terms for more or less the same concept. In this tradition, trust
has been termed generalized trust (Stack, 1978) or trust propensity (e.g., Bigley & Pearce,
1998; Cheung & Lee, 2001; Lee & Turban, 2001; Mayer et al., 1995; McKnight & Chervany,
2001).

Rotter’s (1967, 1971, 1980) work can be regarded as being one of the most representative and
influential contributions in this area. Still, numerous researchers elaborate on his conceptual
definition and his view widely influenced their conceptualizations of trust in various disciplines
(e.g., Zaltman & Moorman, 1988). The scholar (Rotter, 1967, p. 651) himself defines trust as
“an expectancy held by an individual or a group that the word, promise, verbal or written
statement of another individual or group can be relied upon”. It has to be noted that, although
this definition seems to suggest that the scholar is referring to trust for a specific referent,
Rotter’s definition is aimed at and focuses on a generalized trust of others that constitutes a
relatively stable personality characteristic that a person would presumably carry from one
situation (or person) to another (Mayer et al., 1995). Rotter’s interpersonal trust definition
therefore describes a general expectancy about the behaviour of generalized others and not a
specific person. In his later work, Rotter describes trust in the same tradition as an expectation

34
that the commitments undertaken by another person (or organization) will be fulfilled (Gefen,
2002b; Rotter, 1971).

In a similar vein, Farris and his colleagues (Farris, Senner & Butterfield, 1973, p. 145) define
trust as “a personality trait of people interacting with peripheral environment of an
organization”. Here, trust is again conceptualized as a trait that leads to a generalized
expectation about the trustworthiness of others. In their work, this trait is referred to as the
propensity to trust. The same notion has been used by Mayer, Davis, and Schoorman (1995, p.
715), as they define propensity to trust as “a stable within party factor that will affect the
likelihood the party will trust”. In line with other definitions of propensity to trust, Bigley and
Pearche (1998, p. 410) state that “factors exist within individuals that predispose them to trust
or distrust others, whom they do not know”. This leads to the notion that people can be
positioned on a spectrum ranging from high to low. Mayer, Davis, and Schoorman (1995) note
that at one extreme people grant total blind trust, while individuals at the other end of the
continuum will not trust anybody, regardless of the circumstances that make it relevant to trust.

A slightly different perspective has been taken by McKnight and Chervany (2001). They define
disposition to trust as “the extent to which one displays a consistent tendency to be willing to
depend on others in general across a broad spectrum of situations and persons” (McKnight &
Chervany, 2001, p. 45). Within this definition, trust does not literally refer to a personality trait;
rather it means that a person has a general disposition to be willing to depend on others. A
person’s disposition to trust does not automatically mean that one generally thinks that other
individuals are trustworthy or believable, but for whatever the reason, one tends to be willing
to depend on others. Dispositional trust hence has nothing to do with having good reasons to
trust, but merely the concept represents a general approach towards interacting with the social
environment. People scoring high on dispositional trust would trust another person under any
circumstances.

Personality psychologists recognize that this individual characteristic is a product of peoples’

lives, where they develop “generalized expectations about the trustworthiness of other people”
(Rotter, 1967). Here, the tendency to trust is not based upon experience with or knowledge of
a specific trusted party (McKnight et al., 1998), but rather the result of an ongoing lifelong
experience (Rotter, 1971), socialization (Fukuyama, 1995) but also learning. To some degree,
the concept seems to be linked to naiveté. Trust as a personality trait is for the most part
developed during childhood as an infant asks for and receives help from benevolent
caregivers/parents (Bowlby, 1982; Erikson, 1968; Rotter, 1971; 1980). Erikson (1968)
advances that people grow up with a disposition to trust or may develop it later in life. However,
as this trait is developed, it is acted out as a generalized reaction to life’s experiences with other
people and a mental strategy which causes habitual patterns of behaviour, thought, and emotion.
According to the model of human nature advanced by the psychoanalyst Erik Erikson, trust

35
represents the first stage of human development. In developmental child psychology, trust is
regarded as the fundamental relationship between a mother and her infant which is based on
consistency and regularities. Erikson (1963) suggests that a necessary precondition of a
“healthy personality” is what he labels “basic trust”. Due to these early-life experiences, the
young child develops a constant tendency to produce trust in a broad range of situations (Rotter,
1967). However, the readiness to trust is subsequently formed and potentially altered by the
socialization process. By drawing from the social learning theory, Rotter (1967) argues that
individuals typically have different experiences with other people’s fulfilment of their promises
and varying reinforcements of previous social interactions (i.e., reinforcement history).
Therefore, people develop different expectancies that specific (negative/positive)
reinforcements will occur when something is promised by other people. The scholar
furthermore states that an individual’s expectations that the promise of a social agent will be
kept will be generalized from one social agent (trust object) to another. As a consequence,
people differ in generalized expectations that the oral or written statements of other people can
be relied upon. This is also certainly true online.

The importance of social interaction for the development of trust is also advocated by Worchel
(1979), who views trust as a function of an individual’s environment. He emphasizes that trust
is a personality trait or response pattern which is the result of early socialization. Hence, “a
person exhibits a disposition to trust to the extent that she/he demonstrates a consistent tendency
to be willing to depend on others across a broad spectrum of situations and [persons]”
(McKnight et al., 1998). Analogously, Tyler and Kramer (1996) note that an individual’s
willingness to trust is based on the person’s estimation of the probability that those trusted will
reciprocate the trust.

The above argumentation outlines the potential of personality characteristics to impact a

person’s trust response and to influence peoples’ amount and level of trust in their interaction
partners. However, results on the influence of generalized trust are controversial, as some
empirical studies suggest that in a particular situation the influence of trust as a context-free
generalized trait is limited compared to the attributes of the trustee (e.g., Schlenker et al. 1973).
In general, research has shown that the impact is particularly relevant in cases when the former
has limited knowledge about the latter or in the absence of available or experiential information
on which to base a judgment (Hofstede, 1980). Thus, as an determinant of (situational) trust,
disposition to trust is most influential in the early phases of a relational exchange when the
trusting parties are still widely unfamiliar with each other, and before extensive ongoing
relationships provide the necessary background information for the formation of other trust-
building beliefs (Gefen, 2000). Rotter (1971, p. 445) comments that “the more novel the
situation, the greater weight generalized expectancies have”. Further, he notes, “the situation
partially determines the response, and the theory predicts that situations of considerable
familiarity are less predictable from a generalized tendency than those involving more novelty”.

36
The importance of dispositional trust is likely to diminish when the person interacts more with
the trusted party (McKnight et al., 1998; Rotter, 1971). Or, as McKnight and Chervany (2001,
p. 45) put it, “disposition to trust will affect trust in a specific other (interpersonal trust), but
only when novel situations arise in which the other and the situation are unfamiliar”. Therefore,
the kind of relationship and its development seem to determine the impact of generalized trust.
Accordingly, it can be assumed that, for online review communication, dispositional trust is a
key variable. However, it is also reasonable to argue that people develop generalized trust that
is specific to the context (e.g., using online reviews) and which coexists with the more abstract
context-free personal trust disposition. Research in the field of personality trait also tells us that
people regularly develop trust that is independent from a specific situation and predisposes them
to think, feel, and act in a certain stable way.

For a long time, personality psychologists have refrained from developing a psychometric
measure of trust as a personality characteristic. A notable exception has been made by Rotter
(1967), who applied his definition of trust to the development of the Interpersonal trust scale
(ITS) – one of the most prominent psychometric scaling approaches for trust measurement still.
By conducting a series of experiments, the researcher was able to compose and test a trust scale
which enables researchers to measure individuals’ general beliefs about the expected reliability
of others, or as Rotter (1967; p. 653) puts it, the instrument can be applied to measure the
individual’s “general optimism regarding the society”. Hence, solely the aspect of constant
positive expectation is emphasized. The ITS is a self-report inventory that consists of 25 Likert-
format items that essentially deal with the perceived credibility of social agents. Rotter was
striving to deduce a person’s tendency to trust society by asking people to characterize certain
agents critical for society. More specifically, the ITS measures a person’s general tendency to
trust a wide variety of social entities, including but not limited to teachers, politicians,
physicians, classmates, journalists, salespersons and strangers. In order to develop an
instrument that is able to capture a generalized expectation towards society, Rotter particularly
focused on selecting questions on social objects that do not represent close social relationships,
such as parents, friends, and life partners (i.e., primary group relationships), but rather constitute
“generalized” others. The scale includes 12 positive and 13 negative formulated expectations.
Many of the items deal – as noted – with the credibility of social agents and include items such
as “most elected officials are really sincere in their campaign promises” or “in dealing with
strangers one is better off to be cautious until they have provided evidence that they are
trustworthy”. Other phrases reflect an individual’s general expectancy on others’ motives and
reliability (e.g., “it is safe to believe that in spite of what people say, most people are primarily
interested in their own welfare” or “parents usually can be relied upon to keep their promises”),
and general scepticism about the future of society (e.g., “hypocrisy is on the increase in our
society”). Meanwhile, the scale is frequently applied to differentiate “high trusters” (those who
are mostly trusting others) from “low trusters”. In his subsequent research, Rotter (1980) found
significant differences between these two groups, as he provides some evidence that people
37
who trust more are less likely to lie and are additionally less likely to cheat or steal. Moreover,
these people are less likely to be unhappy, conflicted or maladjusted. In general, it appears that
high trusters have many social and psychological advantages – an insight that has been very
recently explored by evolutionary psychologists.

Besides Rotter’s (1967) approach, several other researchers have discussed trust in similar ways
and have attempted to measure a person’s general tendency to trust others. For instance,
Dasgupta (1988) proposes a treatment of trust that includes generalized expectations of others;
for example, “can I trust people to come to my rescue if I am about to drown?” (p. 53). However,
the most commonly used measures of trust include Rosenberg’s (1956) Faith-in-people scale,
the trust facet of the NEO PI-R Agreeableness scale (Costa & McCrae, 1992a), Mayer and
Davis’s (1999) Trust propensity scale, and several other self-report personality inventories
incorporating constructs similar to trust. For example, the trust-defensiveness subscale of the
Comprey personality scale (Comrey, 1970), the trust in humanity subscale of personal
orientation dimensions (Shorstrom, 1975), and the trustworthiness subscale of philosophies of
human nature (Wrightsman, 1974). While these personality researchers do not offer an explicit
definition of the meaning of trust, their contributions to the measurement of trust commonly
imply that they regard trust as a faith in humanity. These measures contribute to an
understanding of the nature of trust in that they broaden the construct by recognizing the
necessity to introduce general optimism and confidence for the concept of trustworthiness and
honesty of people in general.

While Rotter’s (1967) ITS represents one of the classical works in the field of dispositional
trust, the scale itself has also been criticized. For example, Heretick (1981) researched on the
constructs of suspicion and trust by relating them to gender differences. She states that the ITS
scale generally suffers from a lack of discriminant validity and being confounded with another
construct referred to an individual’s expectation about the social environment, namely locus of
control. In particular, Heretick’s criticism refers to the circumstance that the ITS, while being
supposed to measure interpersonal trust, shows relatively high correlations with Rotter’s (1966)
earlier developed Internal-External locus of control (I-E) scale. Motivated by the need for a
scale that exhibits psychometrically independent measurement characteristics, Heretick’s
(1981) 6-item Trust-Suspiciousness scale (T-SS) was developed to measure “expectancies
concerning the motives of other individuals” (p. 269). The T-SS comprises mostly items from
the Minnesota Multiphasic Personality Inventory (MMPI), including items striving to measure
others’ honesty and selfishness, as well as the individual’s general caution towards others’
trustworthiness (e.g., “it is safer to trust nobody”). The author herself states that the scale
exhibits reasonable reliability and only weak correlations with I-E. However, additional
evidence concerning its validity has not been reported.

38
Rosenberg (1957) introduced a 5-item scale that measures a person’s general level of
confidence in the trustworthiness, honesty, goodness, generosity, and brotherliness of people in
general. The scale includes items such as “if you don’t watch yourself, people will take
advantage of you”. Kanter and Mirvis (1989) developed for their survey on cynicism a 7-item
measure of mistrust, which contains items such as “most people will tell a lie if they can get
away with it”. Subsequent research reports (Gurtman, 1992) that the measure is reliable and
highly correlated with scores on Rotter’s (1967) ITS. Another stream of literature implies that
trust and mistrust are two distinct concepts rather than simply a bipolar construct. For example,
Wrightsman (1974) constructed the 84-item Philosophies of human nature measure, which
incorporates a trustworthiness subscale (Omodei & McLennan, 2000).

More recently, Omodei and McLennan (2000) introduced the scale of Global interpersonal
mistrust-trust (Global IMTM), which strives to measure an individual’s global interpersonal
mistrust. The construct is conceptualized by the authors as a general mistrust of the motives of
others in situations related to one’s well-being: “A general tendency to view others as mean,
selfish, malevolent, or unreliable people who are, thus, not to be depended on to treat one well”
(p. 283).

The above-mentioned scales have in common that they all treat trust as a uni-dimensional
construct without underlying factors. However, the individual items typically used seem to
reflect several distinct elements of trust. For example, even though Rotter (1967) does not argue
for multi-dimensionality of the ITS, it is apparent that the scale consists of three underlying
dimensions, or factors: (i) credibility of social institutions; (ii) belief in others’ sincerity; and
(iii) caution (Kaplan, 1973). This parallels the often-applied method by researchers to
operationalize trust in the context of the individual scales. The concept of trust has often been
broadened to include items that are usually associated with the concepts of fearfulness,
cynicism, caution, optimism, or confidence in social institutions (e.g., Omodei & McLennan,
2000). Only recently, the multi-dimensional character of trust as a personality characteristic has
been recognized. For example, McKnight et al. (1998) identify two sub-constructs of
disposition to trust: First, faith in humanity, which refers to one’s belief that others are usually
upright, well-meaning and reliable. Second, trusting stance, which means that “people believe
that they will obtain better interpersonal outcomes by dealing with others as though they are
well-meaning and reliable, regardless of whether those others are reliable or not” (Beldad et al.,
2010). While faith in humanity deals with the attributes of general others, a trusting stance is a
more personal approach to dealing with others (McKnight & Cherany, 2002; McKnight et al.,
2002b). Besides telling us that trust can be regarded as a stable trait (be it towards the
generalized others or online reviewers in general), these approaches also give us the insight that
this trait is likely to be multifaceted and more complex than at first glance.

39
The Sociologists’ Perspective

The scholars Lewis and Weigert (1985a) stress the notion that humans would have no occasion
or requirement to trust apart from their relationships with others. They say: “[trust is] applicable
to the relations among people rather than to their psychological states taken individually” (p.
968). This statement emphasizes the sociological function of trust which can be clearly
separated from the concept’s psychological aspect (Paliszkiewicz & Klepacki, 2013). In
contrast with personality psychologists, who view trust as a psychological event within an
individual, sociologists conceptualize trust as an interpersonal process, deeply rooted in a
fundamental human need for cooperation and interdependence with the social world. Hence,
according to this viewpoint, trust is examined as it exists in social properties of relationships
between people (Granovetter, 1973). Therefore, trust is regarded as a property of collective
units (e.g., dyads, groups, teams, and collectives) – inseparably linked to interrelationships
between people. Such a perspective points to trust as a functional condition for the very
possibility of society (Luhmann 1979). By referring to Barber (1983), Lewis and Weigert
(1985a) state that trust is said to exist in a social system when its members behave accordingly
to it and are secure in the expected futures constituted by the presence of each other of their
symbolic representations (Barber, 1983). Simmel (1964) states that it is the mutual
“faithfulness” that represents nothing less than the foundation of all social relationships.
Consequently, according to the sociological view, trust can be thought to be an important
mechanism for the functioning of society in that the only alternatives to appropriate trust are
“chaos and paralyzing fear” (Luhmann, 1979, p. 142); or, in other words, the breakdown of
human relationships. The social philosopher Sissela Bok (1978, p. 26) likewise emphasizes the
role of trust as a fundamental social good for the existence of society: “when [trust] is destroyed
societies falter and collapse”. Therefore, trust is generally a critical factor for many social and
transactional exchanges, which ultimately determines the kind of the interactions and the
individual’s expectations of it (Fukuyama, 1995; Hosmer, 1995; Lewis & Weigert, 1985;
Luhmann, 1979). Trust is at its heart a necessity for all levels of social exchange relationships.
For instance, Blau (1964), who views trust as a set of three distinct beliefs (i.e., integrity,
benevolence, and ability) that other people would fulfil their expected favourable commitments,
describes trust as an essential component for the early phase but also for the maintenance of
solid social relations, just as exchange obligations promote trust. Similarly, Pruitt (1981) views
trust as a precondition for relationship exchanges that result from coordination and
collaboration. In essence, among sociologists, trust is commonly understood as a social
structure which is regularly situationally constructed.

While striving for an explanation of why trust is essential to social life, sociologists’ salient
argument is that trust can be regarded as a mechanism which “reduces the complexity” of social
interactions. Niklas Luhmann (1979) advanced this notion and today he is regarded as one of
the pivotal representatives of this school of thought. In his short but nevertheless influential

40
book Trust and Power (1979) the scholar placed trust at the heart of sociological theorizing for
understanding modern society. Like other sociologists such as Barber (1983) later, Luhmann
proposes that trust is an irreducible and multi-dimensional social reality. More specifically,
Luhmann’s work borrows heavily from the phenomenological and symbolic interaction
traditionalists (Parson, 1970; Simmel, 1964) in order to develop a global theory that strives to
blend the macro- and micro-sociological aspects of trust. The scholar states that “Trust occurs
within a framework of interaction which is influenced by both personality and social system,
and cannot be exclusively associated with either” (Luhmann, 1979, p. 6). The idea that the
concept of trust cannot be completely understood by exclusively studying the psychological or
the institutional level is perhaps one of the most important theoretical insights of Luhmann’s
work (Lewis & Weigert, 1985). A critical claim of this statement is the assumption that trust
serves the same basic function at the personal level as well as at the more general system level,
as it enables humans to deal with uncertainty, social complexity and often ominous images of
the future as it simplifies life by relieving social agents from risk. Luhmann advances his
thought that modern (industry) society is generally organized by complex and well-integrated
temporal structures. Within such a system, social interactions are controlled by “socially
expected durations” which determine when activities start and end (Lewis & Weigert, 1985).
Social interaction is characterized by the fact that the interaction with other individuals – who
are inevitably independent and their behaviour is not fully predictable, combined with an inborn
need to understand the actions of the other members of society – faces individuals with an
overwhelming social complexity (Gefen, 2000). This complexity can so severe that it
completely paralyzes any social action. Generally, interaction partners are not able to control
the actions of the other party – in fact it is often impossible to understand the real motivation
that leads to such a behaviour. For instance, people are not able to take an introspective view
of their interaction partners and are therefore never able to know the true motivations which
finally drive their behaviour. Are the reviewers attempting to help me or are all or some of them
trying to manipulate me in a certain way? In such an environment, one would have to consider
all possible future outcomes which maybe arise with equal probability. Hence, a person would
have to find and anticipate the outcomes of all possible futures of an interaction yet to come –
an almost impossible endeavour, as everything/anything can be true/false. What here is truly
required is a strategy that reduces this intricacy to manageable and calculative proportions.
Otherwise, this complexity would be so dramatic that it would actually inhibit most intentions
to perform cooperative behaviours. Since people have to interact with others under
unpredictable circumstances on a regular basis or in similar situations, they apply a variety of
approaches that aim to reduce this crushing complexity. Rational prediction is one of these
strategies. By collecting and processing information about our interaction counterpart and
known causal relationships in the social environment, we are able to make predictions that
certain futures are more probable than others, while other futures can be regarded as being too
remote or too uncertain to require serious consideration in the present planning. However,

41
because of human nature, rational prediction alone is not sufficient. People simply do not have
the required resources such as cognitive abilities, but also time, to rationally predict and control
the effects of all possible futures. This seems to be also especially true in online environments,
where information on others’ reputation is simply missing or can be only indirectly derived.
Such predictions are also often hindered because humans tend to be more irrational than rational
entities. Instead, Luhmann (1979) proposes that trust is a (simple) functional alternative to
rational prediction. To trust is convenient to humans as the concept reduces complexity far more
quickly, economically, and thoroughly than rational predication does. Therefore, the author
proposes – in line with many sociological scholars that followed – that trust is one of the most
effective methods to reduce social complexity and therefore represents a focal aspect in most
interactions with other people. He assumes that trust succeeds where prediction based on
rational thinking alone would fail – because to trust means to live as if certain at least rationally
possible futures will not occur, while others surely will (Lewis & Weigert, 1985). Trust makes
an otherwise unjustifiable belief in the future subjectively justifiable and facilitates social
interactions to continue on a simple and confident basis. When trust vanishes, these interactions
would again be impaired by immense social complexity posed by contingent futures which
would finally paralyze action again. Overall, Luhmann proposes that people who trust reduce
their perceived social complexity through a (generalized) belief. Sometimes, such a belief can
be irrational and may appear to be unwise. Nevertheless, it still eliminates the risk of
disadvantageous future behaviours of the trusted party to an acceptable level. Trust is a shortcut
to intense mental efforts to make predictions about something that is unpredictable.

Luhmann’s work on the role of trust in social interactions had an important impact on
subsequent scholars that followed his view by recognizing the sociological function of trust.
For instance, Gefen (2000) defined trust “[…] in the broad sense, [trust] is the confidence of a
person […] in his or her favourable expectations of what other people will do, based, in many
cases, on previous interactions (experience)” (p. 726). Trust is especially important in situations
that are not governed by other complexity reduction methods, such as rules and regulations
(Fukuyama, 1995). Even if rules exist, trust is critical, because there is still no guarantee that
other individuals will completely abide by them. Luhmann recognized that trust, of course, does
not really enable individuals to control or even anticipate without error the behaviour of others;
however, the construct facilitates for people the creation of a comprehensible organization of
their social exchanges. In his subsequent work (Luhmann 1979; 1988), the sociologist sharpens
his position that trust is a prerequisite of human behaviour, as it represents not less than a “basic
fact of social life” (Luhmann, 1979, p. 4).

Luhmann (1979) also advances the view that different types of trust do exist. Specifically, he
states that society passes through different developmental stages. While there exist small and
relatively undifferentiated societies (e.g., tribes) where social order depends widely on personal
or interpersonal trust (trust between people), complex industrial societies in contrast are based

42
more on system trust. Here, system trust can be defined as trust in the functioning of
bureaucratic sanctions, norms and safeguards. A typical example is the legal system. Luhmann
(1979) advocates that interpersonal trust and system trust differ in respect of their foundation.
In general, the basis of personal trust is built by reciprocated altruism or emotional bonds
between individuals. For system trust, however, intensive emotional involvement is untypical.
Instead, system trust is established when one perceives that everything appears to be “normal”
and the system behaves as expected. Further, system trust is not trust in a particular person, but
rather a generalized trust that all other members of the society will continue to trust in the
system Lewis & Weigert, 1985). Luhmann (1979) calls this assertion “trust in trust”: This
means that we all trust in the system because we trust that others trust in it as well (Lewis &
Weigert, 1985). This makes their actions more calculable. Luhmann characterizes this process
as “overdrawing” of one’s information base. According to the scholar, there exists a collective
cognitive reality that transcends the realm of individual psychology and individual cognitive
bases (Lewis & Weigert, 1985). However, he does not pass the line to assert that people do not
have to individually assess information and make personal judgments of where, when and how
much to trust – rather he recognizes the need for an investigation of the psychological
perspective of trust.

By focusing on the issue of how institutions and incentives can be introduced to reduce anxiety
as well as uncertainty typically associated with relations among relative strangers, sociologists
have further delved into the issue of trust as an institutional phenomenon. Sociologists typically
make the claim that trust is a socially rational expectation which is based on formal
requirements and informal obligations. In this tradition, Zucker’s (1986) conceptualization
refers to trust as a “set of shared social expectations” of the institutional environment. More
specifically, trust constitutes various expectations that are shared by everyone involved in a
social relationship or an economic exchange. Trust includes broad social rules, such as how
does a “fair” rate of interest for a given situation look, as well as legitimate social processes,
such as who has the “right” to dictate this rate in a given situation (Zucker, 1986, p. 54).
Therefore, trust includes the beliefs that all participants follow the same social rules and that in
an established society all social processes function in a proper way. Such beliefs represent the
background expectations of society, as they are taken for granted. According to Zucker (1986),
trust in social exchange cannot only be developed on knowledge based on prior experience
between the exchange partners, but furthermore can be based on the belief in the institutional
environment that has been established to support the trust (e.g., bureaucratic organizations,
professional associations, and laws). In the essence of her work, the scholar emphasizes the
importance of the institutional arrangements, social structures, processes and norms in the
context of trust – an insight that has been advanced also by other authors. For instance, Shapiro
(1987, p. 685) terms these arrangements “guardians of trust”. A communality shared by such
instruments is their capability to reduce both system-dependent as well as transaction-specific
uncertainty. This is accomplished by the elicitation of trusting beliefs which are based on the
43
institutional arrangements functioning as contextual cues that make a social environment or the
circumstances of an economic transaction appear normal. Shapiro (1987) terms these
expectations as “structural assurance beliefs” signifying structural protections or governance
mechanisms. On the societal level, such supporting structures, guarantees, and safety nets
surface as elements of the legal systems that protect the individual’s personal rights and his/her
property. The role of institutional mechanisms for the establishment of trust has been also
recognized by various other authors (e.g., Lane & Bachmann, 1996). However, the fact that
Zucker’s work has been recognized as a milestone in sociological as well as organizational
research can be attributed to the scholar’s proposed typology of trust. Institutional-based trust,
as discussed before, is only one of the three sources on which trust could be based. She argues
that institution-based trust historically has been supplemented by process-based and
characteristic- or person-based trust. Process-based trust is generally tied to past expected or
past experiences. For example, trust is only developed in exchange relationships where
exchange histories were known and respected or the other party is perceived to be reputable.
On the other hand, characteristic-based trust is related to similarities between individuals.
Exchanges are limited to those individuals who have the same ethnicity, culture, and other
backgrounds, such as shared expectations.

Shapiro (1987) also advances the view of trust as an impersonal rather than a personal state.
Shapiro (1987, p. 626) defines trust as “a social relationship in which principals […] invest
resources, authority or responsibility in another to act on their behalf for some uncertain future
return”. By drawing from the principal-agent theory, the scholar argues that in a regular setting
the principal is not able to perfectly monitor the actions of the agent, who performs a dedicated
task for the principal and on whom they depend. A social control framework ensures that trust
can be developed between the parties in these situations. The instruments of such a framework,
including institutional arrangements and generally accepted structural constraints, are aimed at
ensuring the agent’s honesty. However, Shapiro also challenged the role of institutional-based
trust as she proposes the question: “who guards the guards?” She concluded that “in complex
societies in which agency relationships are indispensable, opportunities for agent abuse
sometimes irresistible, and the ability to specify and enforce substantive norms governing the
outcomes of agency action nearly impossible, a spiralling evolution of procedural norms,
structural constraints, and insurance-like arrangements seems inevitable” (Shapiro, 1987, p.
649).

It is an omnipresent reality that most sociological researchers regard trust as the expectation
that social interaction partners have of one another. For instance, when defining trust in social
relationships and social systems, Koller (1988) refers to the expectation of one interaction
partner that his/her counterpart has the willingness to behave advantageously toward him/her,
although the person who is trusted has the freedom to select among alternative behaviours that
may also have harmful consequences for the individual that trusts.

44
Barber (1983), who also provides an excellent discussion of the earliest works on the concept
of trust, agrees by advocating that trust is basically a set of optimistic speculations regarding
the behaviour of the interaction partner. Nevertheless, in doing so, Barber takes a predominantly
cognitive and functional stance and develops the idea that one has to shift the focus of those
speculations from the ultimate outcome of a single event, which is uncertain, to the three basic
conditions that determine that outcome. By proposing a trust concept that is anchored by three
expectations which determine the fundamental meaning of trust, Barber (1983) simultaneously
argues for a multi-dimensional construct. The author includes (1) expectations of the
persistence and fulfilment of the natural social order in which the individual found himself or
herself. Within this context, Barber cited Niklas Luhmann (1980, p. 4), who characterizes the
social world as being unmanageably complex and states that trust is capable of reducing this
complexity with “cognitive, emotional, and moral expectations that some things will remain as
they are or ought to be”. In essence, according to Barber, trust rises with a person’s speculation
that the world will continue to exist without any fundamental change. (2) The expectation of
technically proficient role performances from those who are involved with the interaction
partner. Many of the earlier writers (e.g., Gabarro, 1978) were concerned about a conceptual
linkage between trust and competence. Barber (1983), on the other hand, avoids this corundrum
with the expectation of technically competent role performance. The author distinguishes
between daily routine actions, technical possibilities, as well as expert knowledge. (3) The
expectation of morally correct role performance from those associated with the individual.
Respectively, the expectation that interaction partners will perform their fiduciary obligations
and responsibilities. Some authors before Barber tend to synthesize the trust concept with
fiduciary duties and responsibilities, but they refrained from defining those terms. However,
Barber avoided this problem in his work by claiming that fiduciary duty of professionals, in
certain situations, means to place the interests of the individual who is trusting before the
interests of the professional who is trusted. In essence, Barber emphasizes the central role of
the expectation that a truster, at a specific point in time, will allow someone else’s (the trustee’s)
interests to prevail above his/her own. This expectation depends to a lesser extent on the earlier
shown behaviour, but mainly on the felt moral obligations and intentions. In the case where the
referent’s behaviour can no longer be anticipated, the person falls back on this. This fact
highlights the strategic aspect of the trust concept. Barber recognizes that trust dimensions can
also be related, as he suggests that greater trust resides in a referent when trust along multiple
dimensions (e.g., competence and concern) is present. Barber (1983) makes a valuable
contribution to the “personal expectations” literature on trust. Here, the author developed his
thoughts on trust towards an interpersonal definition in that for his conceptualization, two
conditions have to be met: (i) the existence of a person who is trusting; and (ii) a second
individual who is worthy of that trust. However, within Barber’s conceptualization, trust
remains basically the optimistic expectations of a person relative to the eventual consequence
of a future, uncertain event.

45
Barber’s view of the dimensionality of trust integrates earlier sociological insights such as the
work of Garfinkel (1967), whose micro-sociological conceptualization of trust is based on the
individual’s expectation of order and stability in a world of everyday interaction. Like Barber
(1983), Liebermann (1981) also states that in fiduciary relationships the dependent party’s trust
is based on a belief in the “competence and integrity of the professional or official in accordance
with the highest ethical standards” (Mayer et al., 1995).

Many sociological scholars, with Barber (1983) amongst them, overly restrict their trust
conceptualization to expectations, which leads to a purely functionalistic analysis of the
construct. Lewis and Weigert (1985) likewise agree with earlier sociologists that trust is a set
of expectations with particular contextual conditions, parameters and constraints (see also
Cheung & Lee, 2006). However, they criticize the previous conceptualizations of trust as being
too rational and cognitive centred. Here, the scholars (p. 970) emphasize the critical point that
“trust is based on a cognitive process which discriminates among persons and institutions that
are trustworthy, distrusted, and unknown. In this sense, we cognitively choose whom we will
trust and we base the choice on what we take to be ‘good reasons’, constituting evidence of
trustworthiness. However, knowledge alone can never cause us to trust. The manifestation of
trust on the cognitive level of experience is reached when social actors no longer need or want
any further evidence or rational reasons for their confidence in the objects of trust”. Based on
this criticism, Lewis and Weigert (1985a) conceptualized trust as consisting of distinct
cognitive (cognitive-based trust), emotional (affect-based trust), and behavioural aspects, which
are combined into a unitary social experience or overall attitude towards the trust object. This
viewpoint appears to be deduced from the “age-old trilogy of cognition, affect, and conation”
(Ajzen & Fishbein, 1980; Katz & Stotland, 1959) or the ”three-component view” of attitude
(Himmelfarb & Eagly, 1974). Lewis and Weigert state that the existence of the three dimensions
is nothing new as they are implicitly included by a variety of sociology’s key trust contributors
(especially in Luhmann, 1979). The cognitive aspect of trust includes the notion that we
cognitively decide whom we should trust or distrust, in which respects and under which
circumstances this takes place. In this process, we base our choice on what we take to be “good
reasons” which constitute corresponding evidence for the interaction partner’s trustworthiness.
On the cognitive dimension, trust is based upon knowledge (or perception) of the trustee’s
personal and institutional attributes. In this context, Luhmann (1979, p. 19) states, “familiarity
is the precondition for trust as well as distrust, i.e., for every sort of commitment to a particular
attitude toward the future”. Similarly, Simmel (1964) observes that trust involves a certain
degree of cognitive familiarness with the trusted object which is somewhere on the continuum
between “total knowledge” (also known as completely objective evidence) and on the other end
“total ignorance” (or completely subjective evidence) (Hinnant, 2007).

At the beginning of a social relationship, people trust only when they have access to the
evidence that the other party is trustworthy, but as trust develops gradually, interaction partners

46
no longer need rational reasons for trusting each other (Lewis & Weigert, 1985). Here, Lewis
and Weigert further state that in such a condition it is of no importance how much more
knowledge about an object we may collect, such knowledge alone can never give us reason to
trust. They further assert that the rational component of trust is characterized by a cognitive
“leap” beyond the expectations that reasons and experience alone would warrant. They simply
represent a kind of “platform” from which the leap is created. This process can be termed, by
referring to Luhmann’s (1979) terminology, as “overdrawing” on the informational base.
Typically, each individual makes the leap not only because of his/her own psychological
constitution, but in parallel on the assumption that others join the leap. Therefore, cognitive
trust can be characterized as being a collective cognitive reality that goes beyond the realm of
the person’s psychology. According to Lewis and Weigert (1985), trust is not ultimately based
on rational evidence alone, instead, individuals trust by virtue of the norm of trust.

The researchers (Lewis and Weigert 1985a, b) further identify an emotional or affective base
of trust that is complementary to its cognitive counterpart. In this context, Lewis and Weigert
argue that affective trust is motivated by strong positive feelings among those who participate
in the relationship. Trust is viewed as a force that allows for intense emotional investments.
More specifically, expectation-violating behaviours elicit emotional reactions. They state that
“The affective content of the trust relationship is evident in the strong emotional response
elicited by its betrayal. Although feelings of anger and guilt are usually strongest when an
interpersonal trust is betrayed, we also feel a sense of outrage when a public trust is violated,
as occurs, for example, when a political official violates the oath of office. Because each
participant in the trust relationship knows that its betrayal would bring emotional pain not only
to the person betrayed but also to the betrayer, affect complements the cognitive dimension of
trust”. Therefore, emotional-based trust is characterized as being just as reciprocal and
intersubjective as its cognitive base. For Lewis and Weigert (1985a, b), trust represents a
construct composed of feelings as well as rational thinking. This accords with the psychological
insight that emotions (feelings) have to parallel cognition in order to cause human action. The
exclusion of either one of these two dimensions may mis-conceptualize trust as blind faith
(without any cognitive base) or a rationally calculated prediction (without an emotional base).

The third component of trust is its behavioural enactment. Here, to trust means that the
individual acts as if the future behaviours of the other party were indeed certain, while the
violation of these assumptions leads to negative consequences for those persons involved. In
other words, by borrowing from Barber’s (1983) perspective, Lewis and Weigert (1985a)
define the behavioural ingredient of trust as “the undertaking of a risky course of action on the
confident expectation that all persons involved in the action will act competently and dutifully”
(p. 971). The authors emphasize that, although analytically distinct, behavioural trust is closely
intertwined with the cognitive as well as emotional elements of trust. Moreover, each dimension
implies the others. Nevertheless, Lewis and Weigert (1985a, b) argue that the comparative

47
strength and relative importance of the individual dimensions may vary due to the type of
relationship, situation, and social system. Specifically, they assume that cognitive trust is more
characteristic of the macro level such as large social settings and societies, whereas affective
trust is more typical in primary, close-knit groups (e.g., kin) or close relational situations.
Overall, Lewis and Weigert once again reinforce the sociological assumption that trust is a
“collective attribute” based upon the people’s relationships that co-exist in a social system.
They conclude that trust is essentially social and normative rather than individual and
calculative and is hence a necessity for social relationships. Without trust, such interactions
would be impossible.

In a nutshell, sociology provides meaningful insights to gain a profound understanding of how

humans develop trust, and which fundamental role trust captures in social interactions. While
theoretical contributions on the societal macro level are manifold, only few empirical studies
aim to investigate trust in social settings or on the individual level. Only recently, due to the
rise of research in the context of the Internet, some scholars have attempted to operationalize
and measure types of trust that has its roots in the sociological school of thought. Here, the
concept of institution-based trust is applied to draw on the idea that people are able to form trust
in the Internet as a social system or (more specifically) the Internet as a shopping or research
channel. For instance, McKnight and his colleagues (McKnight et al., 2002b) define Internet-
trust as the belief that structural conditions are present. These beliefs make the probability of
achieving a successful outcome in an endeavour like commerce more likely. The authors draw
their research on the conceptualization of trust being two-dimensional: structural assurance and
situational normality. Here, by mainly drawing on Shapiro (1987) and Zucker (1986), structural
assurance is conceptualized as the consumer’s belief that structures (e.g., guarantees,
regulations) are in place in order to facilitate success. On the other hand, situational normality
is the notion that the world is in proper order and success is likely, as the situation is normal or
favourable (Baier, 1986; Lewis &Weigert, 1985a, b). Other e-commerce and marketing
researchers followed a similar operationalization, as demonstrated later.

This literature stream also tells us that the universe of online reviews can be regarded as a
certain kind of social institution. Society has dedicated certain functions to this system and it is
up to the online consumer himself/herself to judge whether this institution can be trusted.

The Social Psychologists’ and Organizational Researchers’ Perspective

In the psychological literature, a dedicated research stream can be identified that emphasizes
the conceptualization of trust in the context of interpersonal relationships with a specific other
person. In contrast to personality psychologists, who view trust as a characteristic of the
individual that is a stable psychological trait and independent of contextual and time influence,
social psychologists define trust as an expectation about the behaviour of others in social
interactions and transactions respectively (Kautonen, 2008). Here, interpersonal trust is
48
regularly conceptualized as an individual’s state of mind which is mainly influenced by
situational/contextual factors (the environment). These variables serve to either enhance or
inhibit the formation and maintenance of interpersonal trust (Lewicki & Bunker, 1995). They
include the characteristics of the trustee, such as his/her personality, competencies and
perceived reliability, as well as the truster’s expectations about the outcomes of their
relationship. Consequently, the research focus of social psychologists traditionally has been on
the identification of these characteristics that make the trusted party trustworthy, as well as on
the search for situational or contextual elements that affect the development of trust in
interpersonal relationships. More recent approaches investigate the relationship between
personality and mental representations of the environment. Such research attempts to
conceptualize interpersonal trust regularly surface in two different contexts. Social
psychologists have recognized that individuals develop trust (i) in intimate (close) personal
relationships as well as (ii) in business or working relationships (Lewicki & Bunker, 1995). In
both kinds, trust plays an important but different role.

Trust in Intimate Personal Relationships

Larzerle and Huston (1980) contend that trust is a behavioural intention (or motivation) that
reflects the dependence on a partner. By building on previous literature, Larzelere and Huston
(1980) advance the Dyadic trust scale (DTS), which strives for a better understanding and
measurement of the degree of interpersonal trust in close relationships. The authors make a key
contribution of the recognization of trust as the key to the functioning and continuance of
relationships, including marriages and families but also an element in romantic relationships
(Regan & Berscheid, 1999). In their work, trust is conceptualized as a strong belief of the truster
in certain personal attributes of the partner. The scholars conclude that trust can be established
on the basis of two different characteristics of the trustee, namely (i) the partner’s benevolence
and (ii) his/her perceived honesty. Benevolence refers to the truster’s belief that the trustee has
a genuine interest in the other’s welfare. It further targets the evaluation of the extent to which
the other believes that the partner is motivated because of self-interest or cooperation.
Benevolence is deeply related to reciprocal altruism and therefore mostly determined by a social
emotion or feelings towards the interaction partner. The emotion-in-relationships model –
stating that close relationships typically elicit strong emotions (Hogg & Vaughan, 2011) –
basically explains the proper inclusion of this affective trust element. Honesty, in contrast, is
described as the extent to which the truster can believe that the interaction partner’s statements
reflect his/her real future intentions or the reality in general. In reference to the importance of
these two dimensions, Larzelere and Huston (1980, p. 596) define interpersonal trust as dyadic
by proposing that “trust exists to the extent that a person believes another person (or persons)
to be benevolent and honest”. The scholars recognize that benevolence and honesty are
conceptually distinct. However, these two elements of trust are so intervened in personal
relationships, especially when these relationships are intimate in nature, that they are

49
operationally inseparable (e.g., if a partner’s honesty is questioned this may lead to questioning
his/her benevolence, and vice versa). So, it is very unlikely that you perceive a person to be
very altruistic towards you, but you deem the trustee dishonest at the same time. They contend
that trust is a behavioural intention, which reflects the dependence on a partner. This view
parallels other approaches typically surfacing in other disciplines. Larzerle and Huston (1980)
make a distinction between dyadic trust and generalized trust. While the former refers to the
judgment of an individual about the benevolence and honesty of a significant other, in contrast
the latter refers to a person’s belief about the character of people in the aggregate (society).
Therefore, generalized trust represents the understanding of trust in the eyes of personality
psychologists as a stable trait (e.g., Rotter, 1967). This approach acknowledges that different
kinds of trust exist which vary in their context and resistance to change. Larzerle and Huston’s
(1980) DTS measure is composed of eight Likert-format items to measure the two dimensions
of trust in close relationships. Among the final scale, items such as “my partner is primarily
interested in his (her) own welfare” and also the item “my partner is perfectly honest and
truthful with me” can be found.

Larzerle and Huston’s (1980) as well as Johnson-George and Swap (1982) recognize the
importance of a multitude of situational factors and contend that trust is both target and context
(domain) specific. For example, we may trust auto mechanics with our cars, but we probably
would not trust them with regard to investment advice. When somebody asks you “do you trust
this person?” it is reasonable to respond with the question “for doing what?”. Additionally,
these authors agree that their conceptualizations of interpersonal trust reflect assumptions about
the trustee and these inferences seem to be manifold. Therefore, they further advance the view
of trust as a multi-dimensional or complex construct, where the truster’s judgments on certain
characteristics give him/her reason to trust/distrust the interaction partner. The judgments which
have to be made vary according to different authors. Johnson-George and Swap (1982), for
example, describe that interpersonal trust is established when the trustee is expected to be
reliable. According to them, in addition, the truster has to perceive a favourable attitude (i.e.,
emotional trust) towards him/her/it. The authors define reliability as the extent to which one
believes a trustee will keep his/her promises and commitments. On the other hand, what is often
called emotional trust corresponds to one’s confidence in the relationship. If a person has high
emotional trust, he/she is confident that he/she is free from criticism and embarrassment from
the trustee. The person feels emotionally attached. Johnson-George and Swap (1982) introduce
two trust scales that aim to measure trust in meaningful interpersonal relationships, which are
labelled as the Specific interpersonal trust scales (SITS). The authors propose two different
trust scales because their research showed that men and women tend to evaluate trusting
relationships differently.

Rempel, Holmes and Zanna (1985) make a distinction of three main dimensions of trust:
predictability, dependability, and faith. Predictability refers to the consistency and desirability

50
of previously shown behaviours. More specifically, the predictability of a partner’s behaviour
is influenced by the amount of past experience in the social interaction, as well as the degree to
which consistent and stable behaviours were typically shown. A totally predictable trustee
would give the truster the opportunity to completely anticipate his/her future behaviours. The
element of dependability has some connection with the concept of stable motives that influence
the referent’s behaviour. In fact, dependability refers to the sum of a trustee’s performed actions
that provide some insight into the extent to which a referent can be trusted in times to come. As
such, dependability refers to dispositional attributions of the partner’s reliability. Faith is
understood as an emotional security of the partner, which makes individuals feel that their
partner will be responsive towards them in the future. This enables them to go beyond the
available evidence. It reflects the referent’s underlying motives and intentions that represent the
foundation of trust in a social interaction which typically arises after a longer period of time.
Such an evaluation needs an introspection of the trustee’s personality and cannot be sometimes
explained out of good reason. Rempel, Holmes and Zanna (1985) recognize that humans require
varying amounts of time and emotional investment in order to develop perceptions of
predictability, dependability, and faith in relationships. Therefore, in contrast to other scholars
such as Barber (1983), the authors emphasize the dynamic aspect of trust as their model,
implying that trust is developmental with hierarchical stages. At the beginning of a relationship,
predictability constitutes the basis of trust. In the course of the interaction, trust develops
gradually via dependability, and finally resulting in pure faith. For the researchers, the elements
of trust are not mutually exclusive. Rather they conclude that the relative importance of one
element affects the quality of the relationship in a specific stage. For instance, in an early form
of a relationship, predictability and dependability would explain interpersonal trust better,
whereas in mature relationships, faith may be the dominant element of trust.

While the relationships in online review communities cannot be best described as closely-knit
relationships, the ideas presented above foster the insight that interpersonal trust has to be
assessed in terms of multiple characteristics of the interaction partners and that, in this specific
kind of relationship, specific kinds of trustee characteristics are more important than others.
However, as it is a social interaction, the spectrum can be very broad.

Trust in Professional Relationships in Organizations

There are a variety of definitions of interpersonal trust within organizations. However, in the
organizational trust literature, most conceptualizations of trust refer to the concept as
representing a belief or expectation about the other (trusted) party (similarly to the concept
described above), or as a behavioural intention or willingness to depend or rely on another
party, coupled with a sense of vulnerability or risk if the trust is violated (Grabner-Kraeuter &
Kaluscha, 2003).

51
First of all, the contributions of behavioural psychologists on the notion of trust are notable.
These scholars concentrate their efforts on studying trust in laboratory experiments where they
typically investigate trust in prisoner’s dilemma games. In respect of this research tradition,
trust is generally conceptualized by following a strictly behavioural interpretation of the
construct, as they equate trust with cooperation with others in such games. Here, to trust simply
means that an individual chooses to show a cooperative behaviour. Parallel to most of the
earlier-described research approaches to interpersonal trust, behavioural scholars typically
focus on the evaluation of situational variables that increase or decrease a person’s level of trust
(i.e., cooperation) or distrust (i.e., competition) towards other people in the game. However,
behavioural researchers typically neglect the events in the human mind and the psychological
nature of trust. Using behaviours as a proxy of trust also gives rise to some concerns. A
subsequent section will discuss this issue more in detail.

Morton Deutsch (1958) is one of the best-known members of the experimental research group.
Equivalently to other behavioural scientists, he termed trust as being the conceptual equivalent
of cooperative behaviour. However, he made this claim in specific reference to the context of
game theory. Hence, for this scholar, trust is a risk taking behaviour which is evaluated in a
situation-specific manner by the individual. Deutsch’s main assertion is that trust is a non-
rational choice of a person who faces the uncertainty of a future event where the expected gain
is smaller than the expected loss. He concludes that trust is an expectation set that triggers
behavioural intentions which involve potential harm, because of the absence of control over
those persons upon whom the truster ultimately depends (Gefen & Straub, 2004). Hence,
Deutsch (1958) stresses the vulnerability aspect of trust, as the trusting person “will be worse
off, if he (she) trusts and his (her) trust is not fulfilled than if he (she) does not trust” (p. 266).
In line with this argumentation, trust is the non-rational expectation of the outcome of an
uncertain event, given conditions of personal vulnerability. To trust means being at risk of
potential loss if the trusted referent takes the advantage of truster’s vulnerability. This
expectation seems to be best described by the term “confidence”; however, Deutsch never made
this explicit claim. Later, Deutsch (1960) made a more concrete reference to the concept, as he
investigated the background of why an individual would trust another to produce some
beneficial events (Mayer et al., 1995). He concludes that the “individual must have confidence
that the other individual has the ability and intention to produce it” (p. 125). The concept of
confidence also surfaces in subsequent research of social psychologists, with Cook and Wall
(1980) among them. The scholars define trust as “the extent to which one is willing to ascribe
good intentions to and have confidence in the words and actions of other people” (p. 39).

The vulnerability aspect of trust (or the central condition of present risk) is emphasized also by
Zand (1972), as the scholar notes that to trust means that a person increases his/her vulnerability
to others whose behaviour he/she cannot control nor completely predict. Trust is here
conceptualized as having two dimensions, namely personal behaviour and individual

52
expectations. The behaviour, which is also termed “decision to trust”, can be best described as
the choice of giving up control. Such a decision is certainly influenced by the kind of problem
or the extent of personal vulnerability, but further by the expectations of the consequences.
Here, confidence can once again be regarded as being synonymous with trust, but Zand’s central
claim is still the individual’s decision to give up control over the outcome. In sum, interpersonal
trust is a decision of the individual person which is based on an optimistic expectation or the
confidence about the result of a yet uncertain event in the condition of personal vulnerability
and a lack of personal control over the behaviours of the interaction partner. Golembiewski and
McConkie (1975) further equate the notion of trust and confidence by stating that trust generally
“implies reliance on, or confidence in, some event, process or person”. However, this reliance
is acknowledged to be non-rational and, being rooted in personal perceptions and experiences,
subjective. In opposition to Zand (1972), who views trust as a duality, the authors instead
conceptualize trust as a continuum. Trust equals the amount of a person’s hope for a beneficial
outcome. They conclude that trust is “strongly linked to confidence in, and overall optimism
about, desirable events taking place” (p. 134). In their conceptualization of trust, Ring and Van
de Ven (1992) go even further and emphasize the predominant role of confidence for
interpersonal trust. To them, trust rises due to a mixture of two aspects, namely (i) the
confidence or predictability in one’s expectations and (ii) the confidence in the other’s goodwill.
In line with the earlier definition of trust proposed by Barber (1983), moral values are the key
to trust. Specifically, benevolence, which can be described as the duty to take care for the
protection of others, and good will, which represents a person’s intent to respect the interests
of others, both take a prominent role in the behavioural definitions of trust.

Discroll (1978) views organizational trust as “the belief that the decision makers will produce
outcomes favourable to the person’s interests without any influence by the person” (p. 44).
Sitkin and Roth (1993) define trust as an individual’s belief and expectation about the likelihood
of having a desirable action performed by the trustee. Lewicki and Bunker (1995) understand
trust as an expectation about the behaviour of others in transactions. Hosmer (1995), who
provides one of the most complete reviews of the meanings and conceptualizations of prior trust
literature in the past, emphasizes the importance of the concept on the societal level in the
formation of mutually pleasing business relationships. In reference to early psychology and
sociology research, he defines trust as an optimistic expectation that other individuals or
companies will behave ethically. More specifically, Hosmer (1995, p. 393) proposes the
following definition: “Trust is the reliance by one person, group or firm upon a voluntarily
accepted duty on the part of another person, group or firm to recognize and protect the rights
and interests of all others engaged in a joint endeavor or economic exchange”.

Today’s multi-dimensional understanding of trust is mainly based on a key conceptual model

proposed by Mayer, Davis and Shoorman (1995). By striving for an integration of perspectives
from multiple disciplines, Mayer et al. (1995) advance a definition of trust that has been cited

53
over 1,100 times in a wide range of disciplines (Schoorman, Mayer & Davis, 2007). Today one
can confidently claim that their conceptualization is of the most frequently cited definitions in
the contemporary trust literature. The authors mainly take a modern social psychological
standpoint. In essence, they define trust as a truster’s intention to take a risk and suggest that
the truster’s perceptions of the trustee’s characteristics (i.e., the trustworthiness of a trustee) are
the antecedents of trust. More specifically, Mayer et al. (1995) understand trust as “the
willingness of a party to be vulnerable to the actions of another party based on the expectation
that the other will perform a particular action important to the truster, irrespective of the ability
to monitor or control that other party” (p. 712). The proposed conceptualization of trust parallels
the work of Gambetta (1988); however, they make the critical addition of vulnerability.
Specifically, Mayer et al.’s (1995) definition embeds two significant components of trust: trust
is composed of (1) a truster’s set of specific beliefs about certain qualities or characteristics the
trustee owns or lacks, and (2) a set of general beliefs, which represent the truster’s willingness
to be vulnerable. The latter is represented by behavioural intentions which are based on the
former or the reasons to behave (i.e., the trustee’s trustworthiness). Trusting beliefs as well as
intentions have to be present in order that “true trust” exists. The authors are able to elicit three
(contextual) perceptions regarding the other person’s attributes that stimulate trustworthiness,
namely ability, benevolence, and integrity. Mayer et al.’s (1995) model focuses on the truster
and recognizes the interaction partner only in terms of what they are about to do to the truster
and which motives rule their behaviour. According to the authors and in contrast to others’
assumptions, confidence can also be logically differentiated from trust. Furthermore, the
definition emphasizes the role of the truster’s purpose of interaction and the truster’s outcomes
of interaction. The choice to trust is purely made by the truster and it is the product resulting
from trust that furnishes the truster with a personal benefit. In general, the distinction between
trust as a set of specific beliefs and trust as a general belief has been mainly applied to studies
dealing with interpersonal interactions (Gefen et al., 2003) as it is applicable to an exchange
relationship between at least two parties and in case this relationship is of a personal nature.
The relevance for online review interactions follows.

Another author who emphasizes a conceptual relationship between trust and a person’s
willingness to be vulnerable is Mishra (1996). By applying this view the author follows the
notion of vulnerability earlier proposed by various scholars (Deutsch, 1973; Luhmann, 1979),
but he additionally grounds his definition of trust on the relevance of expectations or beliefs
(Barber, 1983; Luhmann, 1979. More specifically, the author defines trust as a party’s
willingness to be vulnerable to another party based on the belief that the latter party is
competent, open, concerned, and reliable. Mishra states that trust is an overall concept,
nevertheless the scholar advances the view that the construct is multi-dimensional, being
composed of these four belief dimensions. Additionally, these elements of trust seem to be
combined in a multiplicative way as Mishra (1996, p. 229) states “[…] a lower level of trust in
terms of any of the dimensions offsets high levels of trust in terms of the other dimension”.
54
A great proportion of research on interpersonal trust suggests that the truster’s willingness to
act represents a critical element in the notion of trust (Albrecht & Travaglione, 2003; Mayer &
Davis, 1999; McAllister, 1995). Accordingly, Ferres et al. (2004) define trust as the employee’s
willingness to act. This behavioural intention, however, is based on the employee’s perception
of the trustworthiness of the other party. The essence of these conceptualizations is that trust is
a person’s willingness to rely on the trustee under the condition that the truster is exposed to
personal vulnerability, as well as dependence on the actions of the trustee.

In contrast to earlier research approaches that have implicitly assumed trust to be uni-
dimensional (e.g., Larzerele & Huston, 1980; Rotter, 1967), the above mentioned
conceptualizations further advance the impression that trust can be expected to be, in fact, a
higher-order construct composed out of diverse (multi-dimensional) factors which may vary
due to the context or trust object itself. For instance, Mishra (1996) concludes that approaches
that conceptualize and operationalize trust as consisting of a single dimension are not able to
discriminate trust from other related constructs such as familiarity and cooperation. Therefore,
the author insisted on a four-dimensional understanding of interpersonal trust. In an early
contribution on the “bases of trust” in working relationships, Gabarro (1978) also advances a
slightly different multi-dimensional conceptualization. In a similar vein, Butler and Cantrell
(1984) investigate the interpersonal nature of trust and the relevance of the concept for
relationships in organizations. To them, trust is a condition for cooperation. By synthesizing
earlier work, Butler and Cantrell propose five distinct components of trust or attributes of the
referent. They assumed that the relative importance of each dimension would differ according
to the occupational position of the person within the firm (superior or subordinate): (1) integrity
– means the reputation of the trustee for being honest and truthful; (2) competence – describes
both the technical expertise and interpersonal skill (soft-skills) required to perform the job; (3)
consistency – is the reliability, predictability, and good judgment in handling certain situations;
(4) loyalty – is understood as the benevolent motives or the willingness to protect, support, and
encourage others; and (5) openness – is the mental accessibility, or the willingness to share
ideas and information freely with others (Pirson, 2007). Butler and Cantrell’s (1984)
conceptualization particularly stresses the importance of ethical considerations, as four of the
five components represent moral values (i.e., integrity, consistency, loyalty, and openness). By
advancing his research, Butler (1991) later specifies the loyalty dimension, as the scholar
changes his earlier notion from a proposed attitude of general benevolence to an implicit
promise from one individual in the dyad relationships not to cause harm to the other party.
Mayer et al. (1995) and Mayer and Davis (1999) also add “benevolence” to the factors of
trustworthiness. Here, benevolence refers to the extent a trustee is believed to intend to do good
to the truster.

Another critical contribution towards a multi-dimensional trust scale has been made by
McKnight, Cummings and Chervany (1998). The scholars conceptually extend the model of

55
Mayer et al. (1995) and propose a comprehensive model on the formation of initial trust. It is
based on the assumption that trust can be developed without experience with the object of trust
(i.e., level of trust at first sight). The scholars define trusting as “that one believes in, and is
willing to depend on, another party” (p. 474). According to their definition of trust, trust can be
broken down into two major sub-constructs: (1) trusting beliefs (i.e., trustworthiness in Mayer
et al.’s model), which are cognitive concerns in respect to the trustee’s trustworthiness in a
given situation; and (2) trusting intentions (i.e., “trust” in Mayer et al.’s model), capturing the
notion that the truster is willing to depend on the trustee in a given situation. These trusting
intentions are in fact a function of the four trusting beliefs, namely benevolence, competence,
honesty and predictability. Later, this model was successfully empirically tested (McKnight et
al., 2002a) and various studies used this measurement approach (e.g., Gefen & Straub, 2004;
Schlosser et al., 2006).

The previously cited research has concentrated its efforts on the investigation of the
characteristics of the trustworthy party or trustee which can be best described as consisting of
multiple (mostly cognitive) trusting beliefs. However, another research stream has attempted to
examine the multi-dimensional nature of trust in a broader context. Here, the contributions of
McAllister (1995) are especially notable, as he differentiates trust’s cognitive and social-
emotional aspects. The role of emotions was recognized implicitly in earlier approaches by
including elements of perceived social bonds or morality. However, the scholar’s work provides
first evidence that a more clear distinction between cognitive-based trust, which represents the
rationale part of human judgment that involves a referent’s use of evidence and rational analysis
in order to form attributes of the trust components, and affect-based trust, which stems from
emotional bonds among individuals. Affect-based trust, on the other hand, is rooted in
reciprocated interpersonal care and concern. According to McAllister (1995), this type of trust
represents a high level of interpersonal trust.

Economic, Business and Marketing Perspectives

Trust in the Economic Context

Trust is considered to be a crucial element in many economic activities that potentially can
involve opportunistic behaviours disturbing the relationship between the business partners. In
general, the main focus of economic research is the creation of institutions and incentives that
are able to minimize the anxiety and uncertainty which are typically related with economic
transactions (Williamson, 1993). Economic transactions can be regarded as a specialized kind
of interpersonal behaviour. Most contributions in economic research seem to be especially
influenced by Williamson’s (1974) terminology and typically use the terms “principal” and
“agent” to refer to the interaction partners. Normally, these are individuals, but also economic
exchange between groups or with firms are in the focus of the research contributions. This
literature string represents a considerable expansion of the trust (or better distrust) concept - as
56
economic exchange theory typically investigates the concept from its negative or “dark” side.
This view is mostly attributable to one of the basic assumptions of transaction cost theory; here,
it is argued that the agent is always striving to maximize his/her own welfare. Therefore, the
risk of opportunistic behaviour is always high and the agent can actually never be trusted.
Williamson (1985, p. 47) refers to opportunism as “self-interest seeking with guile”. According
to the author, people are not opportunistic per se but they are able and often likely to show such
behaviours. Consequently, firms (i.e., principals) face an overwhelming difficulty and task to
identify trustworthy agents. As a consequence, firms have to structure themselves (and make
their decisions) as if all agents cannot be trusted. In such an environment, for firms it is a basic
necessity to negotiate and supervise detailed contracts in order to protect themselves against
opportunistic behaviour. These contracts and controls represent substitutes of trust. Williamson
states that hierarchies and markets both breed transaction costs and that the differences in the
costs of contracts versus the costs of controls finally determine the strategic options of the firm
(Hosmer, 1995). Another author, Hill (1990), also recognizes that firms are not able to identify
economic actors in the market that are cooperative or opportunistic. However, the scholar
advances the perspective that the reputation of an agent can be used as a meaningful selection
mechanism to identify those who are more likely to cooperate. This can as a consequence reduce
transaction costs. Reputation is, of course, the result of trustworthy behaviour. Hill never
explicitly defines the concept of trust. Nevertheless, in line with his work, one can derive a
conceptualization of trust as the economically rational decision to hold close to the promises
given in the contract and to do exactly that what others expect. A departure from this principle
will otherwise cause an eventual loss in one’s reputation and, as a consequence, fewer
contracting opportunities in future (Hosmer, 1995). Bromiley and Cummings (1992), on the
other hand, make the explicit claim that trust reduces transaction costs and also propose a
specific definition of trust, as they write, “trust is the expectation that another individual or
group will (1) make a good faith effort to behave in accordance with any commitments, both
explicit or implicit; (2) be honest whatever negotiations preceded those commitments; and (3)
not take excessive advantage of others even when the opportunity (to renegotiate) is available”.
Therefore, the authors implicitly describe trust as one’s assessment of others’ goodwill and
reliability, while not using the exact terms. These scholars further state that trust not only
reduces the costs of monitoring the agent’s performance but also eliminates the need for
implementing additional control mechanisms that are bonded with short-term financial results
of the firm. Bradach and Eccles (1989) argue that markets as well as hierarchies are not the only
ways to govern economic transactions, but in addition price, authority, and (most notably) trust
are alternatives and represent independent methods which can be combined in a variety of ways.
For these scholars, trust does not replace the market or the hierarchy completely, but can be
regarded as a critical complement. In accordance with Gambetta (1988, p. 217), they define
trust as the “probability that one economic actor will make decisions and take actions that will
be beneficial or at least not detrimental to another” and state that in such conditions cooperation

57
would be a more valid strategy than competition”. In contrast to Hill, Bromiley and Cummings,
and Gambetta, the authors, however, differ in that they view social norms and interpersonal
relationships as the sources of trust, rather than rational considerations of self-interest (Hosmer,
1995). While in economic research, precise definitions of trust are missing, this literature string
makes a valuable contribution to the explanation of the role of trust as a governance mechanism
in economic relationships. Additionally, there often exists an obvious overlap in the
conceptualization of the concepts between economic and especially social psychology
contributions in cases where the former approaches have decided to define trust in their own
discipline. But also the understanding of trust as a specific kind of behaviour regularly surfaces.

Trust in the B2B Marketing Context

Among marketers there is a consensus that trust plays a central role in a variety of relationships:
This includes buyer/seller relations in industrial marketing settings (e.g., Anderson & Narus,
1990; Ganesan, 1994; Morgan & Hunt, 1994) and consumer/seller relationships in relational
retail marketing (e.g., Johnson & Grayson, 2005; Kennedy, Ferrell & LeClair, 2001; Young &
Albaum, 2003) – but also in consumer/consumer relationships. The latter is a more recently
discussed topic, acknowledging that the role of interpersonal interactions between shoppers is
increasing in importance due to the evident trends towards social online shopping. In essence,
trust has been found to play a pivotal role in holding a relationship between buyers and sellers
together (Berry, 1995; Dwyer & Oh, 1987; Moorman et al., 1993). Most efforts of marketing
research have focused their attention on trust in a particular interaction partner, its meanings
and its effects in buyer-seller relationships. This literature stream has produced several
definitions of trust. Here, a large proportion directly stems from earlier trust conceptualizations
made in other disciplines, such as organizational research. Some of these definitions can be
attributed to trust research in the B2B marketing context. Various trust researchers have focused
on salespersons as a prominent object of trust (see Swan, Bowers & Richardson, 1999 for a
review). This is due to the fact that salespersons have always been regarded as the primary
source of communication and as organizational representatives. This makes their boundary-
spanning role critical in relationship success (Sharma & Stafford, 2002). However, the supplier
firm itself has also become a target of research, as Doney and Cannon (1997) state that
individuals “can trust the supplier, its salespeople, or both” (p. 36).

According to Dwyer and Oh (1987), the introduction of the trust concept to marketing literature
can be credited to Schurr and Ozanne (1985). Their research was guided by the proposition that
trust takes an important role in exchange relationships, as it results in constructive dialogue
among participants and cooperative problem-solving. This argument is well supported in other
disciplines. More specifically, Schurr and Ozanne especially enhanced our knowledge of how
trust as well as bargaining influences buyer-seller interactions. They conceptualize trust as an
expectation that the relationship partner will fulfil his/her expected commitments. So for them,

58
reliability seems to be the key component. Moorman et al. (1992) state that Anderson and Narus
(1990) share the similarity that they describe trust as a causing factor of the amount of
cooperation and the functionality of conflict between parties. Hence, trust is viewed in terms of
perceived benefits of dealing with another company. Here, trust is conceptualized as a critical
antecedent of relationship quality (Bramall et al., 2004). The authors define trust as “the firm’s
belief that another company will perform actions that will result in positive outcomes for the
[trusting] firm, as well as not take unexpected actions that would result in negative outcomes
for the firm” (p. 45). Here, it appears that the predictability of the interaction partner’s
behaviours is an essential element of trust development. The strength of this belief ultimately
determines the trusting response or action of the firm (e.g., the willingness to trade), whereby
the company commits itself to a potential loss, depending upon the further actions of the other
firm. To trust implicitly means here to make oneself vulnerable to the actions of others.

Similarly to some organizational researchers, various marketing scholars have equated trust
with confidence. For instance, Crosby, Evans and Cowles (1990) conceptualize trust in a
salesperson as a “customer’s confident belief that the salesperson can be relied upon to behave
in a manner that serves long-term customer interests” (as cited in Swan et al., 1999, p. 95).
Likewise, Morgan and Hunt’s (1994) examination of the nature of relationship marketing
characterizes trust as one party’s confidence in an exchange partner’s reliability and integrity,
which can be linked with qualities such as consistency, competency, honesty, fairness,
responsibility, helpfulness, and benevolence. This definition shows parallels to the definition
of Moorman, Deshpande and Zaltman (1993), but also draws on Rotter’s (1967) classic view
of trust as a generalized expectancy held by an individual. Nevertheless, Morgan and Hunt’s
(1994) definition excludes the trustee’s behavioural intention of “willingness”, as they state that
genuine confidence that a partner can rely on another will automatically include the behavioural
intention to rely. They further say that “if one is confident, then one would be willing; if one is
not willing, then one is not genuinely confident” (p. 23-24). Therefore, they propose that for a
measure of trust, such willingness is irrelevant because it is redundant.

In contrast, Moorman and his colleagues (1992) together with Ganesan (1994) both emphasize
the need to bring “willingness to rely on” into a trust definition and do not see this type of
redundancy. According to Morgan and his colleagues (1993; 1992), trust is the “willingness to
rely on an exchange partner in whom one has confidence” (1992, p. 315). As trust typically
involves some form of vulnerability and uncertainty on the part of the truster, the scholars
advance the view that behavioural intention is a necessity for trust to exist. Without
vulnerability, trust is meaningless because the outcomes of the interaction are inconsequential
for the trusting partner. Trust would shrink to a metaphor. Morgan and his colleagues stress that
this view builds on the insights of earlier research in the field of interpersonal trust (Deutsch
1962; Giffin 1967; Schlenker et al 1973; Zand 1972). They further suggest that both belief and
behavioural intention must be present in order to have trust, because “if one believes that a

59
partner is trustworthy and yet unwilling to rely on that partner, the trust is limited. However, if
one is willing to rely on a partner without holding a belief about that partner’s trustworthiness,
reliance may be more a function of power and control than trust” (1992, p. 315). By extending
this view, one could say that a willingness to rely without a corresponding belief of trust
indicating characteristics could also indicate pure faith in the interaction partner. Nevertheless,
Morgan and Hunt (1994) as well as Moorman and his colleagues (1992, 1993) both emphasize
the relevance of confidence and reliability for the concept of trust in a similar manner.
Moorman, Zaltman and Desphande (1992) view trust as a confident belief, sentiment, or
expectation about the exchange partner’s trustworthiness. They further theorize that this
attribution is mainly based on the partner’s extent of expertness, reliability, or intentionality. It
hence focuses only on the belief dimension.

The question of whether trust in a business relationship is a unitary or multi-dimensional

construct is not meaningful nor consistently answered. Exemplars can be identified for both
schools of research. For instance, trust has been regularly treated as a uni-dimensional construct
(e.g., Anderson & Weitz, 1989; Gefen et al., 2003; Guenzi, 2002; Kennedy et al., 2001)
typically measured by a limited number of items (e.g., Anderson & Narus, 1990). By drawing
on literature in social psychology (Larzelere & Huston, 1980) as well as prior marketing
insights, Doney and Cannon (1997) include the concept of reputation as part of the definition
of trustworthiness. They describe trust as the perceived credibility (i.e., the expectancy that the
partner’s word or written statement can be relied upon) and benevolence of a target of trust (i.e.,
the extent to which one partner is genuinely interested in the other partner’s welfare and
motivated to seek joint gain). Nevertheless, the authors conclude that, even though trust can be
conceptualized by two distinct concepts, the concepts may be so intertwined that in practice
they are operationally inseparable.

In contrast, various other authors were able to identify multiple dimensions of trust. For
instance, Ganesan (1994) and Ganesan and Hess (1997) explicitly describe two major
ingredients of trust: benevolence and credibility. More specifically, Ganesan (1994) draws on
the definition by Moorman and his colleagues (1992) and defines trust as the willingness to rely
on an exchange partner in whom one has confidence. Credibility is one of the underlying
trusting beliefs that focus on the extent to which the retailer believes that the partner has the
required expertise necessary to fulfil the job effectively and reliably. The belief in the trustee’s
benevolence, as the second dimension, is best described as “the extent to which the retailer
believes that the vendor has intentions and motives beneficial to the retailer when new
conditions arise, conditions for which a commitment was not made” (Ganesan, 1994). Later,
Ganesan and Hess (1997) propose the same trust dimensions, however, with slightly different
descriptions. According to them, credibility refers to the main “partner’s intention and ability
to keep promises and deals with partner characteristics such as task specific competencies,
reliability in the delivery of goods and services, and predictability in terms of job related

60
behaviours”. Hence, it appears to be a theoretically multifaceted construct. In contrast,
benevolence relates to the belief of genuine concern that the interaction partner strives for a
purely egocentric profit motive. Benevolent trustees strive for fairness within the interaction.
They further advance the view that trust can have multiple referents.

Kumar, Scheer and Steenkamp (1995) conceptualize trust with two essential components,
namely (1) trust in the partner’s honesty (i.e., the belief that the partner stands by his/her words,
fulfils promised role obligations, and is sincere); and (2) trust in the partner’s benevolence (i.e.,
the belief that the partner is interested in the firm’s welfare and will not take unexpected actions
that will negatively affect the firm). By developing the Organizational trust inventory to
quantitatively assess trust in business relationships, Cummings and Bromiley (1996) propose a
multi-dimensional conceptualization of trust in terms of three belief dimensions: keeping
commitments, negotiating honestly, and avoiding taking excessive advantage. The first two
beliefs seem to be attributable to the honesty dimension. However there seem to be some
differences concerning the time frame (i.e., negotiating commitments precedes keeping
commitments).

Aforementioned research provides the basic insights that trust in the context of business
relationships has been frequently treated as a cognitive process. The role of social emotions is
only implicitly acknowledged. Nevertheless, there are various researchers that also take the
affective dimension of trust explicitly into account. For instance, Swan et al. (1988), and Young
and Albaum (2003) emphasize that trust consists of a cognitive dimension together with an
affective dimension. Others have included “likeability or positive affect toward the party” as an
affective component of trust (e.g., Andaleeb & Anwar, 1996; Doney & Cannon, 1997). In
general, likeability is defined as the extent to which a trusted party is friendly, pleasant, and
likeable (Andaleeb & Anwar, 1996). By investigating trust from the buyer as well as the seller
side, Hawes et al. (1989) discovered that both parties agree that a trusted salesperson has to be
likeable, competent, and dependable. Other scholars describe affective trust as an emotional
security or confidence in the relationships (e.g., Johnson & Grayson, 2005; Swan et al., 1988).
Swan et al. (1988) have investigated how customers grant their trust to industrial sales persons.
They explicitly contend that trust consists of the emotion of a buyer feeling secure about relying
on the salesperson as well as specific beliefs about the trustworthiness of a salesperson. To be
perceived to be trustworthy, they conclude, salespersons have to exhibit a full arsenal of positive
characteristics: They have to be dependable, honest, likeable, and have to possess a dedicated
customer orientation. Young and Albaum (2003) conceptualize trust as an evolving
psychological state consisting of both emotional and cognitive components that result from the
trustee’s perceptions of competence and a well-meaning, caring motivation in the relationship
partner. To recapitulate, one has to learn that trust includes social-emotional aspects even in
highly “formalized” and “rationalized” contexts.

61
Trust in the B2C Marketing Context

Research on the consumer trust concept has been extremely fruitful since the recognition of the
concept’s importance for relationship marketing. Various marketing scholars have integrated
the insights of interdisciplinary research in their work or were at least inspired by other
disciplines. For instance, by drawing from social psychologists and organizational researchers,
Huff (2000) defines trust as “the willingness to rely on and be vulnerable to a marketer, based
on confidence that the marketer will benefit and not harm the consumer” (p. 206). An inbound
characteristic of this definition is the psychologists’ view of trust as a psychological state
defined as a confident (positive) expectation regarding an action important to the truster.
Further, critical elements such as dependence, risk and uncertainty are all mirrored in this
conceptualization of trust. In addition, Huff recognizes in his conceptual paper that to trust
means that one must have the willingness to be vulnerable to his/her interaction partner (i.e.,
trusting intentions), and that this willingness involves cognition as well as (social) emotions.
Similarly, Johnson and Grayson (2005) draw on a well-established theoretical basis (e.g., Lewis
& Weigert, 1985) when suggesting a hierarchical multi-dimensionality of trust, and
conceptualize trust as having cognitive, affective, and behavioural dimensions. Their multi-
faceted concept of trust found empirical support as cognitive and affective elements, while
being highly correlated, turned out to be empirically distinguishable.

However, most recognizable advances towards a better understanding of consumer trust have
been made more recently in the field of online trust (refer to Beldad et al., 2010 for a review).
The term online trust is not consistently idenfied as it is regularly understood as a consumer’s
trust in (i) the e-vendor, (ii), the online shop or commercial website, or (iii) online shopping in
general. This research string also borrows heavily from earlier well-established
conceptualizations developed in other disciplines. Here, the work of McKnight, Choudhury and
Kacmar (2002a), which is specifically based on earlier contributions of Mayer et al. (1995) as
well as McKnight et al. (1998), represents a dominant approach to regard and measure online
trust as a multi-dimensional concept. Meanwhile, their fundamental work has been adapted by
numerous online researchers (e.g., Aiken & Bousch, 2006; Kim & Tadisina, 2007). Originally,
McKnight et al. (2002a) focus on the formation of initial trust, which means to build trust in an
unfamiliar trustee. That is, “a relationship in which the actors do not have credible, meaningful
information about, or affective bonds with, each other” (p. 335, referring to Bigley & Pearce,
1998). In such a situation, trusters have to build trust through cognitive processes and thought.
Hence, they apply contextual cues as information sources and have to “construct” trust by
“adding up” judgments of the characteristics of the trustee. McKnight et al.’s contribution is
grounded in the cognitive-based trust literature which contend that trusting beliefs can be
produced quickly due to various reasons; for instance, social categorization, reputation,
illusions or irrational thinking, disposition, institutional roles and structures, or out of the need
to immediately cooperate on a task (McKnight et al., 1998). In contrast, scholars of the

62
knowledge-based trust approach (e.g., Blau 1964; Ring and Van de Ven 1994; Lewicki and
Bunker 1995; Shapiro et al. 1992) suggest that trust develops successively through social
exchange and hence represents some kind of continuity. Nevertheless, according to McKnight
et al. (2002a, b) the construct can be conceptualized as both trusting beliefs (i.e., “one believes
that the other party [trustee] has one or more characteristics beneficial to oneself” (McKnight
& Chervany, 2002, p. 46) and trusting intentions (i.e., “one is securely willing to depend on, or
intends to depend on, the other party even though one cannot control that party” (McKnight &
Chervany, 2002, p. 46)). Besides these two elements, their definition also includes risk, as they
state: “trust is the willingness of a party to be vulnerable to the actions of another party based
on the expectations that the other will perform a particular action important to the truster,
irrespective of the ability to monitor that other party” (here, the authors refer to Mayer et al.
(1995), p. 712). In essence, cognitive-based or initial trust is mostly grounded on rational
expectations about the behaviour of the object of trust (McKnight et al., 1998), or trustee. In
turn, those rational expectations are based on beliefs about the attributes of the trustee as being
beneficial to the consumer, here also called the truster (McKnight et al. 2002a, b). According
to the authors’ review of 32 articles, three trusting beliefs are most essential for the existence
of trust: (1) competence (i.e., the ability of the trustee to do what the truster needs); (2)
benevolence (i.e., the trustee’s motivation to act in the interest of the truster); and (3) integrity
(i.e., the trustee’s perceived honesty and promise keeping). McKnight et al. (2002a) build on
the holistic trust model proposed by McKnight et al. (1998) and establish conceptual
relationships between the concepts of institution-based trust, trusting intentions, trusting beliefs,
and disposition to trust by applying a parsimonious model of the theory of reasoned action
(TRA). According to the scholars, trusting beliefs (or the perceptions of the e-vendor) lead to
trusting intentions (i.e., the customer’s intention to engage with a specific e-vendor for
purchasing purposes – e.g., willingness to follow advice, share information, and make a
purchase), which in turn results in risk-taking behaviours. McKnight et al. (2002a) advance that
both institution-based trust and disposition to trust are determinants of trusting beliefs as well
as trusting intentions. This acknowledges that basic traits and general attitudes impact
contextual trust formation.

McKnight et al.’s (2002a, b) basic conceptualization of trust as a combination of trusting beliefs

(i.e., a set of specific beliefs) and trusting intentions (i.e., a general belief) has been advocated
by various researchers (e.g., Becerra & Korgaonkar, 2011). For instance, for Aiken and Bousch
(2006), trust is “a partner’s willingness to rely on an exchange partner in the face of risk” (p.
309). Aiken and Bousch describe trust as an attitude that has cognitive, affective, and conative
(behavioural) components. The first two dimensions are covered by beliefs about the trustee’s
credibility (i.e., reliability) and benevolence (i.e., the motivation to seek joint gain). Kim and
Tadisina (2007) build on McKnight et al.’s (2002a) model. However, they argue for a more
comprehensive trust model, as they conceptualize online trust as (1) trusting beliefs (i.e.,
perceptions of trustee’s competence and goodwill); (2) trusting attitudes (i.e., customer’s
63
positive affect and confidence in the e-vendor); and (3) trusting intention (i.e., a customer’s
willingness to take a risk in the relationship). A similar approach is also advanced in the context
of advertising trust by Soh et al. (2009).

Parallel to McKnight et al. (2002a, b), research emphasizing the nature of online trust has
conceptualized the construct in two other ways (Gefen, 2002b). First, trust has been viewed as
a set of specific beliefs about the other party, dealing with beliefs about critical combinations
of the trustee’s characteristics (e.g., integrity, benevolence, ability) (e.g., Chang & Chen, 2008;
Chen & Dibb, 2010; Chen, Griffith & Wang, 2005). This literature string is, for example,
represented by the work of Chiou (2004). While giving no explicit definition for trust, the author
measures the construct with a five-item measure adopted from Smith (1997), which strives to
mirror trust in the truster’s belief in the honesty, responsibility, consumer understanding,
profession, and company’s care. Gefen, Karahanna and Straub (2003) also make a distinction
between behaviour and beliefs and conceptualize trust as a set of specific beliefs including
evaluations of integrity, benevolence, ability, and predictability. They explicitly refrain from
integrating trust affect since, according to the authors, such elements are arguably irrelevant in
business relations. Trust is operationalized with a single dimension while implicitly mirroring
the diverse trusting beliefs dimension. The model of Gefen and his colleagues (2003a) suggests
that the antecedents of trust can be categorized into knowledge-based, calculative-based, and
institutional-based trust antecedents. In a later article, Gefen and Straub (2004) define trust as
a context-dependent multi-dimensional social concept whose vital dimensions depend on the
specific circumstances framing the interaction. In reference to an online bookstore, the scholars
propose integrity, predictability, ability, and benevolence as the key (cognitive) elements of
consumer trust in an e-vendor (nevertheless, he contends that behavioural intentions are also
critical elements of trust). According to Pavlou and Fygenson (2006), trust is “the belief that
the trustee will act cooperatively to fulfill the truster’s expectations without exploiting its
vulnerabilities” (p. 123). Wang and Benbasat (2005) conceptualize initial trust in online
recommendation agents as the truster’s set of specific beliefs in the competence, benevolence,
and integrity of the trusted party.

Second, trust has also been described as a general belief that the specific other party can be
trusted (e.g., Arnold, Landry & Reynold, 2007). For example, Gefen (2000) used a single
dimensional scale (with three items) to measure consumers’ overall trust assessments in an
Internet vendor. He defined trust as “the confidence a person has in his/her favourable
expectations of what other people will do, based, in many cases, on previous interactions” (p.
726). Gefen’s (2000) scale was later modified and applied by Gefen and Straub (2003), who
followed Luhmann (1979) in defining trust as the belief that other individuals will act in
foreseeable ways.

64
In reference to the second approach, researchers sometimes view the specific beliefs in the
trustee’s characteristics (e.g., integrity, benevolence, ability) – which are often labelled as
trustworthiness – as determinants of the general belief in trust (e.g., Büttner & Göritz, 2008).
Gefen (2002b), for instance, adopts this perspective by naming the specific beliefs as
“dimensions of trustworthiness”, and nominating the general belief as “overall trust”.
According to him, three specific trust beliefs (i.e., ability, integrity, and benevolence) together
compose trustworthiness and are viewed as antecedents of overall trust. Gefen (2002b) defines
trust (by referring to Mayer et al. 1995) as an expectation that the other party’s commitments
will be fulfilled under conditions of risk by saying that trust is the “willingness to be vulnerable
to the actions of another person or people” (p. 39). The author developed a multi-dimensional
measure of trust with six items measuring trusting beliefs and nine items to capture a
consumer’s window-shopping and purchase intentions. Jarvenpaa and Tractinsky (1999) is
another example. Trust is here described as “a consumer’s willingness to rely on the seller and
take action in circumstances where such action makes the consumer vulnerable to the seller”
(taken from an online resource). This definition is explicitly built on the earlier insights of
Mayer et al. (1995) as well as McKnight et al. (1998). It analogously concentrates on the
cognitive elements of trust (due to context) and describes the process of trust formation as a
rational choice. For Jarvenpaa and Tractinsky, a store’s reputation and perceived size are both
antecedents of trust, which is conceptualized as a combination of trustworthiness, integrity and
benevolence, and is measured by a uni-dimensional (3-item) scale. The same conceptual
approach was later advocated by Jarvenpaa, Tractinsky and Vitale (2000), who define it by
referring to Doney and Cannon (1997, p. 37) as “a truster’s expectations about the motives and
behaviours of a trustee”. In their research, the authors also applied Doney and Cannon’s uni-
dimensional measure. Later, Jarvenpaa and Tractinksy’s (1999) single-dimension trust measure
was also applied in the work by Pavlou (2003). In his definition of trust, he recognized
McKnight et al.’s (2002a) conceptualization and describes trust as a belief in the other party to
behave responsibly and to fulfil the trusting party’s expectations. Additionally, the trustee does
not aim to take advantage of the risky situation of the interaction partner. Specifically, Pavlou
(2003, p. 106) defines trust in B2C e-commerce as “a belief that allows consumers to willingly
become vulnerable to Web retailers after having taken the retailers’ characteristics into
consideration”.

Lee and Turban (2001) propose an integrative model for consumers’ trust in Internet shopping.
The authors present trust in terms of expectations and willingness, as they hold that trust is the
“willingness of a consumer to be vulnerable to the actions of an Internet merchant in an Internet
shopping transaction, based on the expectation that the Internet [vendor] will behave in certain
agreeable ways, irrespective of the ability of the consumer to monitor or control the Internet
merchant” (p. 79). According to Lee and Turban (2001), the trustworthiness of an e-vendor is
composed of ability, benevolence, and integrity. In addition, the research found that the
propensity to trust moderated the relationship between the trusting belief of perceived integrity
65
and consumer trust in Internet shopping (measured with a uni-dimensional scale with three
items: “Internet shopping is unreliable”, “Internet shopping cannot be trusted, there are just too
many uncertainties”, and “In general, I cannot rely on Internet vendors to keep the promises
that they make”). Koufaris and Hampton-Sosa (2004) also define trust in terms of a general
belief, as they define initial trust as “the willingness to rely on a third party after the first
interaction with that party” (p. 378). They recognize that uncertainty is a necessary condition
for trust to exist. Additionally, they assume that the perception of uncertainty is generally based
upon the beliefs describing the trustee’s ability, integrity, and benevolence. Hence, in their
model, they relate various perceptions of the online company (e.g., perceived willingness to
customize, perceived reputation, perceived size) and perceptions about the website (e.g.,
perceived usefulness) to trust perceptions, which is also influenced by a person’s trust
propensity. Initial trust was measured with a uni-dimensional scale adapted from Doney and
Cannon (1997) and Jarvenpaa and Tractinsky (1999).

Trust in the C2C Marketing Context

Up to today, there exists only a small number of research contributions that explicitly target
trust in the context of online consumer interactions. However, the general interest of the
scientific community in the consumer-to-consumer (C2C) marketing communications topic is
constantly increasing (Mack, Blose & Pan, 2008; Urban et al., 2009) as the phenomenon gets
more and more evident on the web. Nowadays, it is the combined effort of marketing as well
as information systems (IS) researchers that provides a fruitful ground for future academic
insights. However, this research string also seems to suffer from the general problematic
endeavour to define trust, with the result that several scholars refrain from providing an explicit
definition (e.g., Pan & Chiou, 2011). And operationalization approaches clearly suffer from
contextual appropriateness and psychometric evaluations. For instance, researchers typically
use ad-hoc measures taken mainly from earlier research targeting online or organizational trust
(e.g., Mayer et al. 1995) and adapt them slightly in order to fit the trust object.

Nevertheless, by investigating earlier contributions, one can conclude that two schools of
thought exist today that both heavily mirror the traditional conceptualizations of trust made in
other disciplines. Representatives of the first school define trust as a generalized belief (Hsiao
et al., 2010). For example, in their work, Hsu et al. (2007) adapt the view of Mayer et al. (1995)
and state that trust is “the willingness of a party to be vulnerable to the actions of another party
based on the expectation that the other will perform a particular action important to the truster,
irrespective of the ability to monitor or control that other party” (p. 157). McKnight et al.’s
(2003) vision of decision trust has been implemented similarly in the work of Racherla et al.
(2012, p. 97), who define trust in online product reviews as “the extent to which a consumer is
willing to depend on something or someone in a given situation with a feeling of relative
security, even though negative consequences are possible”. Similarly, the understanding of trust

66
as “a willingness of a party to be vulnerable to the actions of another” advanced by Mayer,
Davis and Schoorman (1995) is stressed by other eWOM researchers (Smith et al., 2005; Smith,
2002). Shin (2010) also did not bring forward an own definition, but adapts the work of Dwyer
et al. (2007) for defining eWOM trust. These approaches imply that the inner nature of trust
remains the same, independent of the trust object.

The second prominent school of thought views trust as a set of specific implicit beliefs (Burgess
et al., 2011; Casalo et al., 2008; Chen & Hung, 2010. Among these scholars it is accepted that
trust is multi-dimensional. However, the propositions for the most important dimensions differ.
For instance, Wu et al. (2010) adopt the view of McKnight and Chervany (2002) and state that
ability, benevolence, integrity, and predictability are the key elements that reflect the
confidence of individuals in virtual communities. By implicitly referring to the earlier work of
Gefen (2002b), Hosmer (1995) and Moorman et al. (1992), Riding et al. (2002) define trust as
“an implicit set of beliefs that the other party will refrain from opportunistic behaviour and will
not take advantage of the situation” (p. 275). They further outline that the construct consists of
three distinct beliefs or factors (i.e., ability, benevolence, and integrity) (Blau 1964; Butler
1991; Giffin 1967; Mayer et al. 1995). However, they admit at the same time that these beliefs
are likely to be intertwined (Crosby et al. 1990; Ganesan 1994). Due to this condition, the
authors further conclude that the dimension of ability and the combined dimension of
benevolence/integrity appear to represent interpersonal online trust best. Such a dimensionality
has been proposed in the context of teams before (Jarvenpaa, Knoll & Leidner, 1998). Casalo
et al. (2008) define three types of beliefs, which, according to them, can be differentiated on
the levels of competence, honesty, and benevolence. A slightly different perspective is
advanced by Smith (2002), who makes an early contribution to the greater context of eWOM
trust by conceptualizing it as the specific beliefs in the trustee’s competence (i.e., the belief that
the recommender is knowledgeable) and affiliation. The latter refers loosely to the concepts of
source attractiveness (Giffin, 1967) and likeability, as well as homophily or similarity (Gilly,
Graham, Wolfinbarger & Yale, 1998). This trust dimension is operationalized by asking the
respondents whether they trust the recommender due to a perceived similar background, similar
tastes and preferences, and/or because they identify themselves with the recommender. In her
dissertation, Smith initially theorized on a third dimension: moral integrity (a construct that is
maybe also best described as benevolence). However, this has not been tested empirically. The
author states that perceptions of moral integrity usually evolve due to long lasting or in-person
relationships, which is typically not the case in virtual communities, where the peer
recommenders are typically strangers. However, it has to be pointed out that societal norms also
apply to online communities.

In reference to the measurement of trust, current research also suffers from various
shortcomings. For instance, some researchers recognize the complex nature of eWOM trust but
prefer to measure it as a single dimension (e.g., Pan et al., 2011; Pernard et al., 2010; Smith et

67
al., 2005). Others define trust in terms of a generalized belief. However, at the same time they
measure the construct later solely with items representing specific beliefs (e.g., Smith, 2002).
Lastly, there exists no research aimed at finding an adequate measure for eWOM trust. Instead,
researchers adopt existing trust scales from a diversified field of trust research. For instance,
Fang et al.’s (2010) measure mirrors a modified version of the scales earlier applied by Chiu et
al. (2006), Cook and Wall (1980), and Ridings et al. (2002). Such approaches to imprecisely
measure trust in eWOM mirror a lack of understanding of the discussed concept; however, they
still represent today’s research norm (e.g., Casalo et al., 2008; Chen & Hung, 2010; Fogel &
Nehmad, 2009; Lu et al., 2010). This thesis assumes that the conceptual heterogeneity and
wobbliness in measuring trust stems from both a lack of conceptual understanding of the trust
construct as well as the domain-specific characteristic of the social interaction.

Towards an Integrated View of Trust Components

The following section strives for an integration of interdisciplinary research on the elements
that can be considered manifestations of trust in social but impersonal communication. These
elements of trust are often referred to interchangeably as determinants, dimensions, or
principles of trust – depending on the basic rationale. In fact, the terms all refer to elements that
are likely to mirror a sense of trustworthiness or even determine whether individuals will trust
another person, object, or information. The discussion of the interdisciplinary literature has
provided some insights into the nature and general patterns of the trust concept. Hereafter, it
can assumed that the construct exists at the intersection of several trust-mirroring concepts or
trust dimensions. What follows is a proposition of two key frameworks which together enable
a categorization of these trust components. This approach will guide the further development
of an understanding of what has to be included specifically in the eWOM trust domain and what
has to be excluded.

Components of Trust

This thesis’ review on the meanings of trust has documented the various possibilities of viewing
trust and of defining the concept in the light of different disciplines. Major academic disciplines
agree that humans seem to have exquisitely complicated relationships with other people. Here,
the concept of trust is regarded a strong force of human interaction and a reasonably good
predictor of human behaviour. However, the disciplines’ manifold conceptualizations and
operationalizations have led to serious confusion among researchers concerning how to
integrate different concepts that all seem to play a crucial role for the understanding of trust as
a complex construct of the human mind. In addition, the question of whether or not the different
aspects of trust are integrable remains widely unanswered. Gefen (2003) and other scholars
have accentuated the conceptual diversity by pointing out that trust is often viewed in various
ways such as: (1) a set of specific beliefs about the characteristics of another party (Doney &
Cannon, 1997; Ganesan, 1994; Larzerle & Huston, 1980); (2) a general belief that another party

68
can be trusted (Gefen, 2000; Hosmer, 1995), sometimes such beliefs are called trusting
intentions (McKnight et al., 1998) or “the willingness of a party to be vulnerable to the actions
of another” (Mayer et al., 1995, p. 712); (3) affect reflected in the “feelings of confidence and
security in the caring response” of the other party (Rempel et al. 1985, p. 96); or (4) a
combination of these elements. Various scholars have attempted to combine specific and
general human beliefs into a single construct (Doney & Cannon, 1997), while others accentuate
this distinction and claim that specific beliefs can be viewed as determinants of general beliefs
(Jarvenpaa & Tractinsky, 1999; Mayer & Davis, 1999). Here, it is not uncommon to denote the
specific trusting beliefs as the perceived trustworthiness of the referent (Jarvenpaa &
Tractinsky, 1999) causing “overall trust”.

In literature, some articles aim to resolve at least some of this confusion. Specifically, two
articles, which were both published in the mid-to-late 1990s, are most notable (Mayer et al.
1995; Rosseau et al. 1998). By conceptualizing trust for their integrative model, Mayer and his
colleagues (1995) define trust as the willingness of a truster to be vulnerable to the actions of
the trustee based on the expectation that the latter will perform a particular action which is of
importance to the former. In a similar manner, Rosseau et al. (1998) made a cross-disciplinary
review of the concept in order to extract common themes, ideas and ingredients from the
alternative conceptual definitions. By building on the central idea of vulnerability and
interdependence, the scholars propose that trust is a “psychological state comprising the
intention to accept vulnerability based upon positive expectations of the intentions or behaviour
of another under conditions of risk and interdependence” (p. 395). It is evident that both
definitions show some recognizable communality (Colquitt, Scott & LePine, 2007). Most
notably, they agree that a conceptualization of trust has to respect the concept’s connectivity
with a person’s intention to accept vulnerability. The willingness to take risks in a relationship
is stressed by trust research in various disciplines (Boon & Holmes, 1991; Deutsch, 1958;
Moorman et al., 1993; Schlenker, Helm & Tedschi, 1973; Zand, 1972). The truster has to know
that he/she can be cheated by the trustee. For instance, a consumer can be deceived by reviewers
when providing wrong information deliberately. Nevertheless, due to a certain mental condition
(i.e., trust), the truster is willing to depend on this information when carrying out actions (e.g.,
buying products). Another important component of trust, which also stems from earlier attempts
to conceptualize trust (Barber, 1983; Boon & Holmes, 1991; Cook & Wall, 1980; Kee & Knox,
1970), is the truster’s positive or confident expectations of a set of beliefs about the presence
or absence of certain qualities of the trustee, paired with certain kinds of feelings.

Both conceptualizations of trust have been well received and accepted by numerous scholars
(Singh & Sirdeshmukh, 2000). They nowadays represent the theoretical basis for a great
proportion of social science research on the topic. However, especially Rosseau et al.’s (1998)
work has also received some criticism, as others regard their definition as too abstract to be
useful for conceptual or empirical work (Bigley & Pearce, 1998). Such voices make the claim

69
that trust has to be conceptualized in the context of the domain as well as in reference to the
specific meaning of the construct in the various disciplines (Chen & Dhillon, 2003). This thesis
basically agrees with this assertion, as it is here theorized that the underlying domain and the
context both ultimately determine the nature and the content of the trust construct itself.
However, it appears that some core overlaps between different trust constructs do exist.
Nevertheless, it is highly unlikely that completely the same ingredients are identical across
contexts or trusting objects.

The discussion about the nature of trust has led to the formation of two schools of thought.
Researchers of the first literature string insist that trust has to be treated as a unitary concept
(e.g., Rotter, 1971). Hence, they advance the thinking that trust can be measured by a single
dimension, such as reliability (Mohr & Spekman, 1994) or the trustee’s motivation (Crosby et
al., 1990). This thesis assumes that such an approach fails to capture the general nature of trust.
Others hold that trust is, by its very own nature, complex, multi-dimensional and/or a higher-
order construct. Trust researchers increasingly agree with this opinion and it seems that
meanwhile the majority of scholars are followers of this school (e.g., Butler, 1991; Ganesan,
1994; Ganesan & Hess, 1997). Therefore, this study’s author feels confident that a multi-
dimensional approach capturing the manifold nature of trust is most appropriate.

This said, there still exists little consensus on the specific dimensions that together represent
trust. In addition, there is also only a small amount of research on how the dimensions
interrelate. Obviously, trust is a very complicated and comprehensive construct. This thesis’, as
well as, for instance, Kim et al.’s (2007) review, provides some evidence that some key
concepts are regularly associated with the construct. For instance, confidence (Cook & Wall,
1980; Deutsch, 1960; Luhmann, 1979), reliance (Giffin, 1967; Schlenker et al., 1973), integrity,
honesty, expectancy (Rotter, 1967), expectations (Barber, 1983; Good, 1988; Luhmann, 1979;
Zucker, 1986), confident or positive expectations (Lewis & Weigert, 1985), vulnerability
(Zand, 1972), decision to cooperate (Gambetta, 1988), willingness to rely on (Moorman et al.,
1993), acceptance of and exposure to vulnerability (Doney, Cannon & Mullen, 1998; Rosseau
et al., 1998), belief in a person’s competence (Sitkin & Roth, 1993), cognition-based and
affection-based confidence and willingness (McAllister, 1995), perceived credibility and
benevolence (Doney & Cannon, 1997; Ganesan & Hess, 1997), intention to accept vulnerability
(Mayer et al., 1995; Rosseau et al., 1998), beliefs and intention (McKnight et al., 1998), risk,
and belief in other’s goodwill and benign intent (Burke & Stets, 1999). While some of these
concepts seem to have an overlapping meaning, an approach recognizing the manifold nature
of the concept appears most adequate. It is very likely that a variety of these concepts could
also be applied to form the conceptual basis (or content) of the eWOM trust construct. Their
apparent relevance to the meaning of trust in the traditional sense leads to the assumption that
viewing trust as a higher-order construct with various subdimensions could best capture the
breadth and complexity of the underlying construct. In order to research something that is

70
apparently multi-dimensional in its sub-dimensions, one must first understand all of its
components. Therefore, the thesis stresses two key frameworks or classifications of trust
components that lead to a better comprehension of the concept in a more systematic way.

Trust Types: Beliefs, Attitudes, and Behavioural Intentions

The first framework emphasizes the various mental constructs which together represent what
can be named trust. In line with interdisciplinary literature, it appears that trust can be best
described as something that is regularly called “attitude”. A paradigm that enables a meaningful
insight and parallels various conceptualizations of trust is provided by the school of social
psychology. More specifically, Fishbein and Ajzen (1975) propose a notable framework that
makes a clear separation of attitudes, beliefs, and intended behaviour that together reside in the
mind of a person. The authors have integrated this classification of mental constructs in a series
of models, most notably the theory of reasoned action (TRA) (Ajzen & Fishbein, 1980) and the
theory of planned behaviour (TPB) (Ajzen, 1988; Ajzen & Madden, 1986). In general, these
models postulate that behavioural intention is the direct antecedent of actual behaviour.
Behavioural intention is defined as the individual’s likelihood of engaging in the behaviour of
interest and is itself a function of attitudes and beliefs (Ajzen & Fishbein, 1980; Pookulangara,
Hawley & Xiao, 2011). The TRA and TPB both assert that attitudes mediate between beliefs
and intentions, although beliefs may similarly have a direct impact on intention. In their original
idea, Fishbein and Ajzen (1975) emphasized that these three elements can be used to
conceptualize a construct as they note: “concepts like attraction, value, sentiment, valence, and
utility all seem to be subsumed under the category of attitude; opinion, knowledge, information,
stereotype, etc. may all be viewed as beliefs held by an individual; brand image, prejudice, and
morale may relate to all three categories” (p. 13). The most insightful contribution of Ajzen and
Fishbein is the categorization of mental constructs that others have summarized or simplified
as “attitude”. By following this insight, various trust studies have proposed that trust is the
conglomerate of all three of them, namely trusting beliefs, trusting attitudes, and trusting
intention (e.g., Kim & Tadisina, 2007). Other researchers prefer a more parsimonious approach
by recognizing that at least two dimensions (i.e., trusting beliefs and trusting intentions)
adequately capture the construct’s nature (Mayer et al., 1995; McKnight et al., 1998). Most of
the studies that make such a distinction originate in investigations of interpersonal interactions,
such as those happening within a firm. In contrast to this view of trust, early studies which
emphasize ongoing economic transactions rarely make such a distinction (e.g., Doney &
Cannon, 1997; Schurr & Ozanne, 1995). Gefen (2003a) argues that this can be attributed to the
facts that trust in economic exchange relationships “is an extension, rather than direct
implementation of the original definition of interpersonal trust”. Further, in contrast to the role
of interpersonal trust as an instrument to foster solid social relationships, trust in economic
relationships serves as a mechanism to avoid opportunistic behaviour. Nevertheless, there is a
growing recognition among marketing researchers concerning the relevance of such a

71
distinction between the elements of trust; this is especially true in the context of online trust
(e.g., McKnight & Chervany, 2002; McKnight et al., 2002a; 2002b; Moorman et al., 1992;
Sirdeshmukh et al., 2002). In addition, as the type of relationship investigated in this research
can be regarded as being economical only in the broader sense (i.e., the focus of investigation
is in fact the relationship among customers and is therefore social in nature), the application of
the above-mentioned framework appears appealing. Based on this research and particularly
drawing on the work of Mayer et al. (1995), McKnight et al. (2002a) and Kim and Tadisina
(2007), this thesis adopts their insights and assumes that the concept of trust can be distilled
down to the three major elements, namely (1) trusting beliefs, (2) trusting attitudes, and (3)
trusting intention (i.e., the trust triangle). Additionally, such a conceptualization relies on the
separation between trust and actual behaviour (e.g., risk taking). This thesis argues for equal
rights of the three “types of trust” as understood in the three-component view of attitude (Hogg
& Vaughan, 2011). All three components have to exist in order to regard the trust concept as a
kind of generalized attitude. It refrains from regarding beliefs as direct antecedents of behaviour
intentions – but intertwined concepts expressing a consistent orientation towards the trust
object. Nevertheless the thesis agrees with TRA/TPB that attitudes (including intent) are
antecedents of behaviour (e.g., information adoption).

In part, trust essentially means that a person (the truster), based on some kind of judgment, finds
the object of trust as able and also willing to act in the best interest of the truster. Without this
judgment, the belief is nothing more than pure “faith”, as in God, or what is often called
“naiveté” (Fournier, 1998). The simple mental process of evaluating the critical trustee’s
characteristics and actions is defined as trusting beliefs (Becerra & Korgaonkar, 2011). Mayer
et al. (1995) focus on the same elements by naming them “factors of trust”.

These beliefs are cognitive judgments or evaluations of the trustee or trust object that result
from attribution processes made by the truster (Fishbein & Ajzen, 1975). Most authors – at least
in the field of marketing – rely upon a belief-based conceptualization of trust (Büttner & Göritz,
2008), where the conglomerate of these beliefs is often referred to as trustworthiness (e.g.,
Mayer et al., 1995). For example McKnight et al. (2002, p. 337) describe trustworthiness as “a
confident truster perception that the trustee has the attributes that are beneficial to the truster”.
In line with support in literature, this thesis proposes that trustworthiness is not a synonym for
trust. In fact, it is only a factual component of the concept. Generally, this distinction is not
always made clear (Blois, 1999). However, Corritore et al. (2003) state that trust is a personal
act of the truster. Whether the trust is well placed or not only emanates from the truster and has
nothing to do with the trustee. However, trustworthiness on the other hand is a mental
representation of certain characteristics of someone or something that can be called the object
of trust. Although trustworthiness and trust are distinct, there is a logical and strong relationship
between them (Solomon & Flores, 2001), as trustworthiness regularly instils trust or trust is
mirrored in perceptions of trustworthiness. This can be illustrated by the following statements:

72
“I trust in [an object] because it exhibits characteristics that signal its trustworthiness to me” or
“I trust an object, therefore I attribute characteristics of trustworthiness to it”. While the first
statement is probably more likely for first time interactions (initial trust) or situation-specific
trust (situational trust), the latter mirrors an individual’s general, stable orientation towards the
trust object (generalized trust).

Literature has widely recognized the trust-trustworthiness distinction. For example, for Good
(1988), trust is based on expectations of how another person will behave. These expectations
are themselves based on that person’s current and earlier implicit and explicit claims. Sheppard
and Sherman (1998) have also recognized that trust results to some degree from people’s
capacity to assess the trustworthiness of their relationship partners. For instance, if an individual
is unsure whether to trust another person met for the very first time, the individual strives to
“construct” initial trust on the basis of (rational) judgments on the trustee’s characteristics.
Hence, trustworthiness (or trusting beliefs) forms here the basis for trust (Serva, Benamati &
Fuller, 2005). Mayer et al. (1995, p. 717) similarly noted that trusting beliefs are “not trust per
se”, but they “help build the foundation for the development of [initial/situational] trust”.
McKnight et al. (1998) and later McKnight and Chervany (2001) also use (multiple) trusting
beliefs as part of their broader trust typology. It has to be noted that in many conditions, trust is
the result of a calculative (and emotional) process, as individuals need to derive their level of
reliance from an evaluation of the characteristics of the trustee. The above literature describes
such cases and, hence, refers to the first statement. However, when individuals have already
made up their mind due to a number of earlier interactions or experiences – hence, they already
possess a stable attitude towards the trust object due a social learning process – the causality is
reversed. Here, trust becomes part of the personality. Trusting beliefs (like trusting attitudes
and intentions) are here theorized to be the consequence of an underlying trust construct
forming a uniform mental dispositon (see Chapter 3 for a discussion in greater detail).

Trusting attitude is another element of trust. Attitude is sometimes described as the

predisposition of the individual to evaluate some symbol or object or aspect of this world in a
favourable or unfavourable manner (Katz, 1960). Ajzen and Fishbein (1980) propose an
alternative definition, as they define attitude as “a state of preparedness or readiness for
attention or action”. Dictionary definitions of attitude also stress this insight, as attitude is
described as “a mental position with regard to a fact or state, or an organismic state of readiness
to respond in a characteristic way to a stimulus” (Mish, 1993). Trusting attitudes include (i) the
affect for the trust object (e.g., reviews) (Thurstone, 1931) and (ii) a specific state of mental
readiness. Specifically, trusting attitudes are a person’s determined mental status which may
lead a person’s mental orientation until something shatters the status (Kim & Tadisina, 2007).
In reference to Eagle and Chaiken (1998), trusting attitudes are defined as a psychological
tendency of the truster which is expressed by judging a particular trust object with some degree
of favour or disfavour. This thesis holds that individuals can potentially have diverse forms of

73
trusting attitudes. Their existence, though, is finally caused by the specific trust-domain. For
instance, the attitudes of affinity or commitment may potentially be a force of trust in inter-
personal (and especially intimate relationships (e.g., marriages)). But they are of no importance
in trust in things such as pencils.

The third element of trust is trusting intention. It is a person’s subjective probability that he/she
will perform the behaviour in question (Fishbein and Ajzen 1975). Hence, it is a person’s active
mental status that potentially results in a subsequent actual behaviour (Kim & Tadisina, 2007).
What distinguishes behavioural intention from trusting intentions is that the latter involves a
recognizable amount of risk (Moorman et al., 1992). Consumers can face a variety of risks –
such as functional, financial, social, and psychological risk (Dowling, 1986; Jacoby & Kaplan,
1972; Stone & Gronhaug, 1993). Therefore, one can generally define trusting intentions as the
willingness to make oneself vulnerable to another entity in the presence of different forms of
risk (Kim, Ferrin, Cooper & Dirks, 2004).

Trust Elements: Cognitive, Affective and Conative Trust Components

The second framework is based on works in the field of attitude and trust research that
emphasize that trust is best described by cognitive (C), affective (A), and behavioural (B)
mental structures (e.g., Lewis & Weigert, 1985a, b): the CAB, or three-component model
(Egaly & Chaiken, 1993; Hogg & Cooper, 2003; Zanna & Rempel, 1988). It seems that a
reasonable amount of trust contributions follow the classic, tripartite view of attitude (Hilgard,
1980; Rosenberg & Hovland, 1960) by treating trust as a synonym of attitude or – at least an
attitude-like – multi-component mental construct. Followers of the CAB model define attitude
as “a relatively enduring organisation of believes [i.e., cognitive components], feelings [i.e.,
affective components] and behavioural tendencies [i.e., intentional components] towards
socially significant objects, groups, events or symbols” (Hogg & Vaughan, 2011, p. 148). This
definition is favoured by various researchers (Himmelfarb & Eagly, 1974; Krech, Crutchfield
& Ballachey, 1962; Ostrom, 1968). Accordingly, it is here assumed that trust as an attitude-like
construct manifests itself in consistent thoughts (beliefs, judgments, knowledge: E.g., “I believe
the trustee is honest.”), emotions (affect, feelings, mood: E.g., “I have a positive feeling towards
the trustee.”), and behavioural intentions (motivations: E.g., “I am willing to rely on the
trustee.”) as distinct elements of trust (Egaly & Chaiken, 1998).

A number of authors suggest that trust involves both cognition which is triggered by rational
thoughts or judgments about the trust object, and emotions towards the trust object
(Granovetter, 1985; Lewis & Weigert, 1985; McAllister, 1995). So the cognitive component
includes cognitive-based beliefs, thoughts and attributes associated with the trust object. In the
case where the formation of trust in a relationship has predominantly cognitive content, it is
called “cognitive trust”, whereas one that has a predominant emotional content is denominated
“affective (or emotional) trust” (Becerra, 2006).
74
Cognition can include attitudes such as confidence. For instance, cognitive trust is sometimes
defined as a person’s confidence in relying on an exchange partner (Moorman et al., 1992;
Rempel et al., 1985). But cognitive trust is also often equated with trusting beliefs. Here, it is
theorized that trust is typically grounded on rational expectations about the behaviour of the
object of trust (McKnight et al., 1998). Those expectations are formed by beliefs about the
characteristics of the object of trust as being beneficial to the truster (McAllister 1995;
McKnight et al. 2002a). In line with this description, trust theory describes cognitive trust as “a
truster’s rational choice that is motivated by a conscious calculation of advantages” (Komiak
& Benbasat, 2006). Johnson and Grayson (2005) claim that cognitive trust is based on the
truster’s knowledge that allows him/her to make predications with a certain degree of
confidence about the likelihood that the interaction partner will live up to his/her obligations.
This characteristic is what Rempel et al. (1985) call “predictability” and Johnson-Geoge and
Swap (1982) refer to as “reliableness”. Such knowledge may arise from previous interactions
or experience, but can also be attributed to reported reputation reflecting the trustee’s earlier
behaviours with others or initial assumptions. Here it has to be stated that, even though trust
may be influenced by a person’s knowledge, only incomplete knowledge is necessary for trust
to become operational. If a person has complete knowledge about the interaction partner’s
future behaviour, risk is eliminated and trust is redundant (Johnson-George & Swap, 1982).

Otherwise it seems intuitively appealing that besides rational thinking, feelings, emotions
and/or affect influence trusting relationships (e.g., Lewis & Weigert, 1985a, 1985b), as most
people who had a strong distrust for someone or something would agree that such relationships
also involve emotions. Emotions can be best described as mental responses that arise in
response to appraisals regarding something of relevance to one’s well-being (Bagozzi &
Dholakia, 1999). As such, emotions can be seen as indicators for the presence and quality of
trust within a relationship (Jones & George, 1998). Affective trust is a person’s confidence in
an interaction partner on the basis of feelings that can be attributed to perceptions of care and
deep concern the interaction partner shows (Johnson-George & Swap, 1982; Rempel et al.,
1985). At its heart, affective trust is the reliance on a partner based on emotions. Various authors
have followed this view by defining affective trust as an emotional bond between two parties
(e.g., McAllister, 1995; Johnson-George &0 Swap, 1982; Lewis & Weigert, 1985a, 1985b);
explicitly, Lewis and Weigert (1985a) characterized the construct as “an emotional bond among
all those who participate in the relationship” (p. 971). As these emotional bonds deepen, trust
in an interaction partner may venture beyond what is defensible by available knowledge. The
presence of affective trust in a relationship makes it less vulnerable to objective risk assessments
which lie in the focus of economic theorists. Emotional bonds between truster and trustee both
enable the making of emotional investments in order to express genuine care and concerns
towards each other, and the showing of confidence in the intrinsic value of the reciprocal
relationships (McAllister, 1995). Rempel et al. (1985) make the statement that affective trust is
closely tied to the opinion that a partner’s behaviours are intrinsically motivated. In a similar
75
vein, Johnson-George and Swap (1982) describe this psychological state as a trustee’s
benevolence towards the truster. Later, the construct of benevolence of McKnight and his
colleagues (2002a) has been classified as one of the three dimensions of cognitive trust.
However, in its essence, affective trust develops from one’s basic social instincts, intuition, or
feelings concerning whether an individual, a group of people, or thing is trustworthy (Morrow,
Hansen & Pearson, 2004).

According to literature, the type of relationship determines the relative importance of cognitive
and affective trust (e.g., Rosseau et al., 1998). For example, Sheppard and Sherman (1998) state
that cognitive trust is the single necessity for economic exchanges. Similarly, Williamson
(1993) emphasizes the importance of cognitive trust by suggesting that the trust between firms
can be best described as “calculative trust”; however, the author further states that in consumer
transactions, a “leap of faith” is a permanent feature. Therefore, trust seems to be multi-faceted
with rational and affective components (Rosenbaum, 2003), besides trusting intentions. Or, to
put it another way, it has “cognitive, affective, or behavioural dimensions with the strength of
the particular dimension being determined by the type of relationship involved between the
truster and the trustee” (Jarvenpaa & Tractinsky, 1999, p. 36). The latter view can be attributed
to a considerable amount of literature that indicates that affective processes and cognitive
beliefs as well as attitudes along with the behavioural intention of trust must be present for trust
to exist (e.g., Johnson & Grayson, 2005). This thesis adopts this opinion.

Table 2 summarizes and classifies the key trust-related concepts that regularly surfaced in prior
research on trust. Specifically, the matrix adapts the original idea of Kim and Tadisina (2007)
and categorizes the identified concepts in line with the frameworks of Fishbein and Ajzen
(1975), as well as Lewis and Weigert (1985a, b). For example, the trust ingredient of
ability/competence is positioned in the cognitive-belief cell as the construct captures the
truster’s belief or rational judgment that the trustee possesses the necessary skills to perform a
specific task for the truster and this belief is driven by a rational judgment. The following
sections discuss selected key ingredients of trust in greater detail and classifies them further.

76
Table 2: Matrix of Trust Dimensions

Trusting Beliefs Trusting Attitudes Trusting Intentions

Cognitive e.g., belief in a person’s e.g., confidence; N/A

ability and competence; reliance; expectancy;
predictability and openness; expectations; confident
integrity; honesty; expectations
reliability; dependability

Affective e.g., belief in a person’s e.g., affection; affect; N/A

benevolence and goodwill likeability

Conative N/A N/A e.g., willingness; vulnerability;

decision to cooperate;
willingness to rely on;
willingness to be vulnerable;
intention to accept vulnerability

Trusting Beliefs

Numerous authors have offered trust definitions that highlight the truster’s thoughts on or
perceptions of the trustee (i.e., trusting beliefs). McKnight and Chervany (2001, p.46) comment
on these judgments that “trusting beliefs means that one believes that the other party has one or
more characteristics beneficial to oneself”. Those authors have defined trust as a multi-
dimensional construct. They have typically included a large variety of characteristics of the
trustee in their conceptualizations that should foster trust, such as ability, integrity,
benevolence, predictability, and dependability (Mayer et al., 1995; Palmer & Bejou, 1994;
Selnes, 1998). A closer look at this research stream shows that these characteristics are often
referred to interchangeably as underlying dimensions, antecedents, or factors of trust (Wang &
Emurian, 2005). In essence, these attributes reflect different components of trustworthiness, a
concept which has also been viewed differently by various scholars (Riegelsberger, Sasse &
McCarthy, 2003). For instance, it was defined as one’s ability and willingness to act in the other
person’s best interest (McLain & Hackman, 1999). However, this thesis adds to a general
understanding of McKnight and Chervany’s (2002, p. 33) definition of trusting beliefs, and
describes the construct as “the extent to which one believes (and feels confident in believing)
that the other person is trustworthy in the situation”. Hence, trustworthiness is the conglomerate
of all critical trusting beliefs.

Research has recognized that trust results from or is mirrored by judgments of numerous
characteristics of the trustee (Barber, 1983; Covello, 1992; Lee & Turban, 2001). For instance,
the trust object’s expected fairness, honesty, and moral obligation towards the interaction
partner. Many research attempts have been advanced in order to identify these characteristics
(Butler, 1991; Butler & Cantrell, 1994; Gabarro, 1978; Mayer et al., 1995; Stickland, 1958).
For instance, Liebermann (1981) concludes that in a fiduciary relationship, trust arises from

77
two essential elements, namely the truster’s beliefs in the trustee’s competence (ability) and in
the perceived integrity. On the other hand, according to Barber (1983) and Covello (1992),
trusting beliefs deal with the knowledge and expertise, openness and honesty, and concern and
care all contribute to higher trust. Later, Stikin (1995) argues for competency, benevolence and
value-based trust representing the key elements of trust. In the same year, Mayer et al. (1995)
found that perceptions of trust depend on the three characteristics of ability, integrity, and
benevolence. The same attributes have been proposed by Lee and Turban (2001), as they
maintain that “three attributes arguably constitute the main elements of trustworthiness: ability,
benevolence, and integrity” (p. 77). Mishra (1996), however, concludes that trust stems from
the perception of competence, concern, openness, and reliability. Doney and Cannon (1997)
proposed predictability, capability, intentionality, and transferability processes for establishing
cognitive trust. McKnight et al.’s (2002a) review on 32 articles found 15 different types of
beliefs or characteristics. The authors conclude that these elements can be clustered into four
narrower dimensions: (1) competence, (2) benevolence, (3) integrity, and (4) predictability.
Bhattacherjee (2002) applied Mayer et al.’s (1995) three characteristics, and additionally found
that predictability and/or consistency appear in several studies (e.g., McKnight et al., 1998;
Sheppard & Sherman, 1998). More recently, Chen and Dhillon (2003) identify that competence,
integrity and benevolence are key dimensions of trust in an Internet vendor. This study has
reviewed a large variety of research contributions in the area of trust. Table 3 shows these
contributions and the variety of trusting beliefs deemed to be important in various contexts.
Similar to earlier approaches (e.g., Mayer et al., 1995), the identified trust beliefs can be
categorized into three dominant classes: (i) ability, (ii) benevolence, and (iii) integrity. Each
class includes a variety of highly intertwined and overlapping concepts typically used in social
research literature.

78
Table 3: Classification of Trusting Beliefs

Trusting Beliefs
Ability Benevolence Integrity Other

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

(Miscellaneous)
Responsiveness

Dependability

Predictability
Benevolence
Competence

Carefulness
Expertness

Dynamism

Credibility

Reliability
Goodwill

Morality
Integrity

Openess
Article/book

Ability; Capability

Honesty; Sincerity

Careful; Safe
Consistency
Goodness
(Ahrholdt, 2011) X X X
(Andaleeb, Lee, &
X X
Gruneisen, 1992)
(Anderson & Narus, 1990) X
(Anderson, Lodish, &
X
Weitz, 1987)
(Baier, 1986) X X
(Barber, 1983) X
(Benedicktus, Brady, Darke,
X X
& Voorhees, 2010)
(Bhattacherjee, 2002) X X X
(Blakeney, 1986) X X X X X X
(Blau, 1964) X X X
(Banoma, 1976) X X X X
(Busch & Wilson, 1976) X
(Butler, 1991) X X X X
(Casalo, Flavian, &
X X X
Guinaliu, 2007)
(Chang & Chen, 2008) X X X
(Chau et al., 2007) X X X
(Chen & Barnes, 2007) X X X
(Chen & Dhillon, 2003) X X X
(Chen & Dibb, 2010) X X X
(Cheung & Lee, 2001) X X X
(Childers, 1986) X X
(Chiou, 2004) X
(Colquitt et al., 2007) X X X

79
 Table 3 (Cont.): Classification of Trusting Beliefs
Trusting Beliefs
Ability Benevolence Integrity Other
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

(Miscellaneous)
Responsiveness

Dependability

Predictability
Benevolence
Competence

Carefulness
Expertness

Dynamism

Credibility

Reliability
Goodwill

Morality
Integrity

Openess
Article/book

(Connolly & Bannister,

X X X
2007)
(Corbitt et al., 2003) X X X
(Covello, 1992) X X X
(Crosby et al., 1990) X X
(Cummings & Bromiley,
X X
1996)
(Deutsch, 1958) X X
(Doney & Cannon, 1997) X X
(Dwyer & Oh, 1987) X X
(Dwyer et al., 1987) X X
(Dunn, 1988) X
(Fassnacht & Köse, 2008) X
(Fisher & Chu, 2009) X
(Flavian & Guinaliu, 2006) X X X
(Flavian & Guinaliu, 2005) X
(Gabarro, 1978) X X X X X X
(Gaines, 1980) X
(Ganesan, 1994) X X X
(Gefen, 2002b) X X X
(Gefen & Silver, 1999) X X X
(Fisher & Chu, 2009) X X X X
(Geyskens & Steenkamp,
X X
1995)
(Giffin, 1967) X X X
(Hart & Saunders, 1997) X X X X X
(Holmes, 1991) X X
(Hosmer, 1995) X
(Husted, 1990) X
(Jarvenpaa et al., 1998) X X X
(John & Reve, 1982) X
(Johnson-George & Swap,
X X X X X
1982)
(Kantsperger & Kunz, 2010) X X
(Kasperson, Golding, &
X X X X
Tuler, 1992)
(Kee & Knox, 1970) X X
(Koller, 1988) X X X X

80
 Table 3 (Cont.): Classification of Trusting Beliefs
Trusting Beliefs
Ability Benevolence Integrity Other
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

(Miscellaneous)
Responsiveness

Dependability

Predictability
Benevolence
Competence

Carefulness
Expertness

Dynamism

Credibility

Reliability
Goodwill

Morality
Integrity

Openess
Article/book

(Korsgaard, Schweiger, &

X X
Speienza, 1995)
(Krackhardt & Stern, 1988) X
(Kumar et al., 1995) X X
(Lee & Turban, 2001) X X X
(Liebermann, 1981) X X
(Lindskold, 1978) X X
(Luhmann, 1979) X X
(Mayer et al., 1995) X X X
(McAllister, 1995) X X
(McGregor, 1967) X
(McKnight & Chervany,
X X X X
2002)
(McKnight et al., 2002b) X X X
(McKnight et al., 1998) X X X X
(Mayer et al., 1995) X X X
(Mishra, 1996) X X X X X
(Moorman et al., 1993) X X
(Morgan & Hunt, 1994) X X
(Murphy, 2003) X X X
(Nöteberg, Christiaanse, &
X
Wallage, 2003)
(Palvia, 2009) X X X
(Pavlou, 2003) X
(Peters et al., 1997) X X X X
(Ramaswami, Srinivasan, &
X
Gorton, 1997)
(Ranaweera & Prabhu,
X
2003)
(Ratnasingham & Pavlou,
X X X
2003)
(Rempel et al., 1985) X X X X X
(Ring & Van de Ven, 1994) X X
(Rosseau et al., 1998) X
(Rotter, 1971) X
(Rotter, 1980) X
(Sato, 1988) X
(Scheer & Stern, 1992) X
(Schlosser et al., 2006) X X X

81
 Table 3 (Cont.): Classification of Trusting Beliefs
Trusting Beliefs
Ability Benevolence Integrity Other
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

(Miscellaneous)
Responsiveness

Dependability

Predictability
Benevolence
Competence

Carefulness
Expertness

Dynamism

Credibility

Reliability
Goodwill

Morality
Integrity

Openess
Article/book

(Schurr & Ozanne, 1985) X

(Selnes, 1998) X
(Serva et al., 2005) X X X
(Sichtmann, 2007) X
(Sitkin & Roth, 1993) X
(Solomon, 1960) X
(Tan & Theon, 2001) X X X X
(Thorslund, 1976) X X
(Worchel, 1979) X X
(Yamagishi & Yamagishi,
X
1994)
(Zaffane, Tipu, & Ryan,
X
2011)
(Zaltman & Moorman,
X X X
1988)

The diversity of trusting beliefs but also structure are obvious. However, it is important to note
that trust components vary from domain to domain. In order to instil trust in the other individual,
the attributes the trusted person must possess also differ from relationship to relationship and
from context to context. The fact that different contexts require different characteristics of the
trusted person or object, some generosity of the trusting beliefs is natural and adequate as
researchers look across a wide range of different situations in order to investigate trust
(McKnight & Chevany, 1996). For instance, as Giffin (1967) addresses trust in speech
communication, it is comprehensible that appearance-related attributes like dynamism and
personal attraction were more important for the author’s conceptualization of trust than other
elements, such as benevolence or honesty of the trustee. More specifically, Giffin (1967)
defines trust as “reliance upon the characteristics of an object, of the occurrence of an event, or
the behaviour of a person in order to archive a desired but uncertain objective in a risky
situation” (p. 105). For Giffin, this trust consists of three trust dimensions that also found their
recognition in Aristotle’s rhetoric: intelligence (corresponding to ability), good character (i.e.,
honesty and integrity), and goodwill (i.e., benevolence) (Gefen & Straub, 2004). In contrast,
Gaines’ (1980) research on trust in relationships between subordinates and superiors shows that
benevolence is key in the formation of subordinates’ trust. There is some recognition that the
relative importance of the belief dimensions is attributable to the social presence of others
(Gefen & Straub, 2004) (i.e., present others, imagined others, etc.). Thus, the kind of human
interaction plays a key role.

82
On the other hand, it is also very likely that some trusters may place a greater amount of
importance on one of the factors across a large array of situations than does another person
(Mayer & Davis, 1999). In literature, however, there is no agreement which or how many
characteristics influence the generation of a trust response or mirror trust best. Despite this fact,
the remainder of this section focuses on a variety of trusting beliefs that repeatedly surface in
interpersonal relationships – albeit under differing titles. This review pays tribute to the growing
acceptance of Mayer et al.’s (1995) classification of trusting beliefs (see also Table 3) which
have present the attributes of ability, benevolence, and integrity as critical components of trust.
This thesis categorizes the beliefs accordingly.

Trustee’s ability is widely accepted as being an important element of trust (Bhattacherjee, 2002;
Chen & Dhillon, 2003; Hovland et al, 1953; Lee & Turban, 2001; Sitkin & Roth, 1993). By
striving to define the concept, Mayer et al. (1995, p. 718) describe it as “the group of skills,
competencies, and characteristics that enable a party to have influence within some specific
domain”. This definition stresses the specificity of the domain of influence (Zand, 1972) by
emphasizing that the trusted object can only develop specific skills, competencies, and
characteristics in specific contexts. That is, when individuals have trust in another person, they
usually think that the trustee is capable of executing a certain task on the behalf of the truster
or is knowledgeable in a certain domain. A great array of researchers has adopted this definition
in their research (e.g., Becerra & Korgaonkar, 2011). Mayer et al. (1995) have recognized that
various scholars have discussed similar concepts as elements of trust, however, using several
synonyms. For instance, Cook and Wall (1980), Deutsch (1960), and Sitkin and Roth (1993)
commonly agree that ability is an important factor that represents trust. Lee and Turban (2001)
describe ability as a construct which manifests a company’s competence in the online shopping
business. According to Colquitt et al. (2007) the ability dimension captures the “can-do”
component of trustworthiness by describing whether the trust object has the necessary skills
and abilities required to carry out a task in an appropriate fashion. Other researchers (e.g.,
Butler, 1991; Butler & Cantrell, 1984; Kee & Knox, 1970; Liebermann, 1981; Mishra, 1996)
propose competence as a similar concept. Tan and Sutherland (2004) define this term as the
belief in the other party’s abilities, skills, and expertise within the certain domain. The concept
can be understood as the truster’s thought that the trustee will be able to accomplish the
expected level of service. Similarly, the competence has been described as a person’s belief that
the exchange partner has the needed abilities to perform a job or desired task efficiently as well
as effectively (Sirdeshmukh et al., 2002). Hovland et al.’s (1953) research on trust identified
perceived expertise as an influential element of trust. Analogously, Giffin (1967) proposes that
expertness is a critical belief that is a manifestation of trust. All of these conceptualizations are
similar to the construct of ability. When a truster puts trust in the trustee, he/she typically
justifies this decision to himself/herself by thinking that the trustee is qualified to carry out the
desired task for the truster. This judgment on the trustee is the result of a majoritarian rational
process.
83
Mayer et al. (1995) also consider the perception of benevolence to be the second key indicator
of trust. Benevolence, which includes responsiveness, goodwill and/or motives for the trustee’s
expected behaviour, relates to the attachment the trustee has towards the truster (Becerra &
Korgaonkar, 2011). The basic idea behind the benevolence concept is the assumption of the
trusting individual that the trustee does not exploit his/her vulnerability and develops a fair
attitude towards the truster. Hence, the trustee doesn’t act manipulatively nor opportunistically
(McKnight et al., 2002b). For Mayer et al. (1995), benevolence is the truster’s perception of the
trustee’s positive orientation toward him/her as they define this trusting belief as “the extent to
which a trustee is believed to want to do good to the truster, aside from an egocentric profit
motive” (p. 718) and further “benevolence is the perception of a positive orientation of the
trustee toward the truster” (p. 719). These words imply that if the trustee is perceived to be
benevolent, he/she is supposed to have some specific attachment to the truster. According to
various authors, this connection is based on affective bonds. For instance, Dunn (2000) and
McAllister (1995) characterize benevolence as an affective characteristic of the object of trust.
Likewise, Riegelsberger et al. (2003) state that affective trust is based on benevolence. It has
been recognized that benevolence can cause an emotional bond to the trustee. Here caring and
supportiveness both foster a sense of positive affect. So it’s an emotion-based thought. Various
researchers have included characteristics similar to benevolence in their conceptualizations of
trust. For example, Larzerle and Huston (1980) as well as Solomon (1960) applied the concept
in their trust framework. More recently, McKnight et al. (2002a) used the term and describe it
as “the truster’s perception of whether the trustee acts in the buyer’s best interest and/or out of
genuine concern”. According to Tan and Sutherland (2004), benevolence is the belief that the
other party wants to do something good for the customer and is not solely looking for making
a profit. More generally, Doney and Cannon (1997) state that benevolence refers to beliefs in
the trustee’s positive orientation. Other researchers have introduced concepts obviously related
to benevolence. For example, Cook and Wall (1980), Deutsch (1960), Giffin (1967), Kee and
Knox (1970), and Mishra (1996) have recognized sympathetic intentions or motives as crucial
elements of trust. These authors have in common that they assume that the trustee’s orientation
towards the truster is critical. The terms intentions and motives possibly include broader
implications than the orientation towards the truster (Mayer et al., 1995). Similarly, Frost et al.
(1978) contend that altruism leads to heightened levels of trust. Butler and Cantrell (1984)
identify loyalty as a key determinant in dyadic trust relationships. Consequently, all of these
researchers used a concept or a basic idea similar to benevolence to describe trust. In this
research, benevolence is recognized as an emotion-based cognitive belief that is triggered by
trusting feelings and especially by experiential judgments about the trustee’s altruism and
empathy. A person has more confidence in another when this person is assumed to be generous,
the truster has the feeling that his/her pains matter to the trustee and the trustee acts in an
unselfish way.

84
Benevolence and integrity – a concept described before – are sometimes referred to as being
the “character variables” of the trustee (Colquitt et al., 2007), as they both capture the “will-do”
components of trustworthiness. They both describe whether the trustee is expected to choose to
use his/her abilities to act in the best interest and on behalf of the truster. Hence, benevolence
and integrity reflect to a large extent ethical traits (McKnight & Chervany, 2002). Their
connection is also highlighted by Bhattacherjee (2002) as the scholar believes that the affect-
based component of trust is captured “within the benevolent and integrity dimension of trust”
(p. 220). A similar assertion is proposed by Grabner-Kräuter and Faullant (2008), who propose
that benevolence and integrity (along with honesty and credibility) represent the “soft
dimensions” of trust. They further state that this encompasses the intrinsic, “value-based
motivation” of the trustee to act in the interest of the truster. They further suggest that the
truster’s perception of the soft dimension mainly is affect-based. In contrast, ability (along with
competence and predictability) are the “hard dimension” of trust, for the assessment of which
the truster used his/her rational judgment. However, there is an ongoing controversy concerning
which belief dimensions can be attributed to cognitive/affective trust. For instance, various
authors propose that apart from the belief of ability, integrity has been found to represent the
cognitive side of trust, while in contrast benevolence is affect-based (e.g., Dunn, 2000;
Riegelsberger et al., 2003). This study follows the latter more established view. Here, it is
agreed that integrity beliefs are mainly triggered by the truster’s attempts to make judgments
resulting from calculative thoughts on the trustee’s honesty and reliability.

According to Mayer et al. (1995) the concept of integrity has two aspects: justice and
consistency. In his research, Gabarro (1978) states that the trustee’s “character” is essential for
trust perceptions and that integrity is one of its key components. Butler and Cantrell (1984) also
recognized the role of integrity as they integrate the concept in their trust model apart from the
elements of competence, consistency, loyalty, and openness. Likewise, Butler (1991) included
integrity, promise fulfilment, and fairness as the conditions of trust. By referring to Mayer et
al.’s (1995) earlier work, McKnight and his colleagues (2002a) conceptualize integrity
similarly, as they propose that the concept is composed of two distinct factors, that are honesty
beliefs (i.e., justice) and predictability beliefs (i.e., consistency). Chen and Dhillon (2003)
conceptualize integrity as acting in a consistent, reliable and honest manner. Lee and Turban
(2001) define integrity as the truster’s perceptions of the trustee’s honesty, keeping
commitments, being reliable, dependable, sincere, genuine, credible and adhering to an
acceptable set of principles. This definition includes various elements of integrity which also
can be found in the work of other trust researchers. For instance, Kumra and Mittal (2004)
emphasize honesty as a critical element of trust. They define it as the belief that the partner is
open and always tells the truth in communications. Other researchers have recognized that
reliability is by definition another important factor for the truster’s confidence in the
relationship partner (Delgado-Ballester & Munuera-Aleman, 2001; Rotter, 1967). Here, the
concept refers to the trustee’s characteristic of standing by his/her words. Some researchers
85
state that integrity is a very rational reason to trust someone (Colquitt et al., 2007). Hence, the
integrity dimension is assumed to be a cognitive-based thought, such as ability.

Parallel with the above-mentioned trust elements, research has less frequently identified various
other factors of trust that impact trusting perceptions. Some of these seem to be quite distinct,
such as openness, which implies a willingness to freely share pertinent information and genuine
feelings (Mishra, 1996). Others appear to have a closer relationship such as predictability,
which focuses on the belief in the other party’s consistent behaviour (Tan & Sutherland, 2004).
Gabarro (1978) defines predictability as the beliefs regarding the trustee’s consistency that may
allow the truster to predict the future behaviour of the trustee. Shapiro et al. (1992) note that
predictability refers to the trustee’s consistency that enables the truster to anticipate the trustee’s
future behaviour. However, McKnight and Chervany (2001b) notice that predictability and
integrity are similar. Nevertheless the concepts also differ, as integrity is a value-laden attribute,
while this is not true for the predictability concept. In general, trust literature tends to mix trust
dimension in measurement approaches. For instance, Ganesan (1994) conceptualizes credibility
in the sense of reliability. However, he operationalizes the construct in terms of honesty and
benevolence, by using items that target the trustee’s extent of being frank, open and not making
false claims. Additionally, reliability has sometimes been used as a synonym for competence
and expertise (Delgado-Ballester & Munuera-Aleman, 2001). However, the author is quite
confident that the true nature of interpersonal trust and all necessary elements are captured by
the classification of the three main trusting beliefs described above. It is further theorized that
these beliefs are relevant across cultures, as they represent moral universals. For instance,
trustee’s fairness is valued by humans all over the world.

Trusting Attitudes

A recent review of trust definitions provides evidence that two dimensions of trusting attitudes,
namely confidence and affect, are both critical elements of trust (Kim & Tadisina, 2007). While
some researchers differentiate trust from confidence (e.g., Mayer et al., 1995), numerous
scholars see trust as a kind of confidence and advance the concept in their conceptualizations
of trust (Deutsch, 1958; Gefen, 2000; Lewicki & Bunker, 1995; Lewis & Weigert, 1985;
Luhmann, 1979). For example, Rempel et al. (1985, p. 96) define trust as the “feelings of
confidence and security in the caring response of the other party”. The concept of confidence
is often referred to as a mutual confidence that no partner of an exchange relationship will
exploit the other’s vulnerabilities (Corritore et al., 2003). Confidence is generally referred to as
being an attitudinal dimension of trust (Fazio & Zanna, 1981; Radan, 1985), and its relevance
for the construct was revealed by various studies (Jian et al., 2000). Analogously, multiple other
researchers have considered similar concepts such as reliance (e.g., Giffin, 1967; Schlenker et
al., 1973). Rotter (1967) focuses on the attitude of expectancy in his conceptualization of trust,
as he defines trust as “an expectancy held by an individual or group that the word, promise,

86
verbal or written statement of another individual or group can be relied upon” (p. 652). In a
similar vein, other scholars have defined trust in terms of expectations (e.g., Barber, 1983; Good
1988; Rotter, 1971; Baier, 1986), confident expectations (e.g., Lewis & Weigert, 1985a), or
subjective probability (e.g., Gambetta, 1988), which all seem to capture a similar dimension of
trust as confidence does. For example, the concept of reliance describes how confident a person
is to depend on others. In parallel, confidence is often referred to as favourable expectations or
predictions of a truster that the trustee will exhibit an advantageous behaviour to the truster.
Accordingly, Barber (1983, p. 164-165) defines trust as “socially learned and socially
confirmed expectations”, which emphasizes the view of trust as a truster’s mental position or
attitude towards the trustee. Confidence is a necessity for trust and has been described as both
cognitive and affective (Luhmann, 1988; Muir, 1994). Accordingly, the concept of expectation
is discussed for being strongly related to predictability as a cognitive component, but also
having an affective component which is hopeful, future-oriented thinking (Corritore et al.,
2003). The thesis at hand, therefore, disagrees with the conceptual work of Kim and Tasisina
(2007), who regard confidence as a cognitive dimension of trust and follows the above
reasoning. Hence, it is the paper’s position that confidence is a concept that rather mirrors the
true nature of the trust construct and, hence, is inseparably tied to the trust construct.

Elements of affective trust attitudes are another part of trust. Affect-based trust – expressed in
feelings/emotions towards the trust object – seems to play a critical role in the majority of other
human relationships but also in impersonal e-commerce relationships. However, only a few
authors have included it in their conceptualizations of trust (e.g., Johnson-George & Swap,
1983; Lewis & Weigert, 1985a; Rempel et al., 1985). A major contribution has been made by
McAllister (1995), who advanced a broad understanding of affect that covers trusting beliefs as
well as emotions. In contrast to the concepts that are mainly based on cognitive processes,
affect-based trust is based on an “emotional bond” (Lewis & Weigert, 1985a) or on “emotional
ties” (McAllister, 1995). In addition, strong emotions such as likeability (i.e., a person’s positive
emotional response towards the trust object) also seem to be an ingredient of affect-based trust.
For example, Johnson-George and Swap’s (1982) conceptualization of trust includes four liking
items along with three other trust items. Accordingly, Swan et al. (1988) emphasize likeability
as a critical dimension of trust. More abstractly, Rempel et al. (1985) include faith as a trust
factor. The affect-based trust dimension seems to capture the communality among these
elements (Kim & Tadisina, 2007).

Trusting Intentions

In definitions of trust, one is frequently able to identify behavioural intention or similar terms
(Mayer et al., 1995; McAllister, 1995; Moorman et al., 1993; Rosseau et al., 1998; zand, 1972).
Such an intention is typically referred to as a willingness to be vulnerable or to take a risk (e.g.,
Kaplan & Nieschwietz, 2003; Stewart, 2003). For example, Zand (1972) argues that a trusting

87
behaviour increases the vulnerability of the truster. Hence, to trust means for the truster that
he/she must have a willingness to take a certain risk in a social relationship. Moorman et al.
(1993, p. 82) explicitly define trust as “the willingness to rely on an exchange partner”. These
scholars further advance the insight that trust is only meaningful in situations that are affected
by vulnerability and uncertainty. Mayer et al. (1995, p. 712) also define trust as “the willingness
of a party to be vulnerable to the actions of another party” under the condition that the truster
has no control over the trustee. McAllister (1995) conceptualizes trust as both the truster’s
confidence and his/her willingness to act. For the scholar, trust “enables people to take risks”
(p. 25). Rosseau et al. (1998) point out that trust becomes operational in situations that are risky
and the relationship among the interaction partners is characterized by dependency. Similarly
to above-cited works, Rosseau et al. (1998, p. 395) define trust as “the intention to accept
vulnerability”. Especially meaningful for the underlying thesis is the work of McKnight and his
colleagues as they separate trusting beliefs from trusting intentions. McKnight and Chervany
(2001, p. 46) refer to trusting intentions by stating “trusting intentions means that one is willing
to depend on, or intends to depend on, the other party even though one cannot control that
party”. Similarly, McKnight and Chervany (2002, p.27) later define trusting intentions as “the
extent to which one party is willing to depend on the other party in a given situation with a
feeling of relative security, even though negative consequences are possible”. These definitions
embody some crucial elements concerning a truster’s behavioural intentions. First, negative
consequences may emphasize the importance of trust. Second, dependence and reliance both
are essential factors supporting the trusting intention. That is, trusting intention (as an
intentional state) is the readiness of a person to depend on the interaction partner in a given
situation. The centrality of this readiness for trusting intentions is mirrored by various authors
(e.g., Giffin, 1967; Lewis & Weigert, 1985a; Ring & Van de Ven, 1994). In this context, to
depend means that the trustee has to do something on the behalf of the truster. Third, the feeling
of security makes the trusting intention stronger. Fourth, trusting intention is context-specific,
as one is willing to depend on the other party in a specific domain as well as person-specific
(Gabarro, 1978; Sitkin & Roth, 1993). Fifth, the willingness of the truster is not based on the
possession of control or power over the other party. Instead, the truster has to “rely on trust”
when he/she trusts another party. Here, McKnight and Chervany relate trusting intentions to the
power literature, since the concept is implicitly defined in terms of dependence and control. In
their subsequent work, McKnight and his colleagues (McKnight et al., 2002a) accordingly
define trusting intentions as the truster’s secure willingness to depend or intention to depend on
the trustee. Here, the authors emphasize that trusting intentions is based on two distinct sub-
constructs, namely the willingness to depend (i.e., “volitional preparedness to make oneself
vulnerable to the trustee”) and subjective probability of depending (i.e., “the perceived
likelihood that one will depend on the other”) (p. 337). They assume that the latter is the more
concrete sub-construct as it goes beyond a stated willingness to rely on another, as it represents

88
a stated intention of relying on them in specific ways. Hence, the truster’s motivation to take
specific risks seems to be an important component of the definition of trust.

Risk-taking Behaviour

For some researchers, such as Lewis and Weigert (1985a, p. 971), actual behaviour is
considered as an aspect of trust, as they mention that “to trust is to act as if the uncertain future
actions of others were indeed certain in circumstances wherein the violation of these
expectations results in negative consequences for those involved”. In contrast, various other
scholars reject this perspective by separating trust from trusting behaviours. For instance,
Mayer et al. (1995) assume that trust (i.e., the willingness or behaviour intention) is a precursor
of actual behaviour (i.e., risk taking). The scholars emphasize that when people trust (i.e., they
exhibit a willingness to be vulnerable), there is no risk taken, but risk is implicitly included in
the behavioural manifestation of the willingness to be vulnerable. The scholars further state that
one does not need to risk anything in order to trust. Nevertheless, an individual has to take a
risk for engaging in a trusting action. They emphasize the critical difference between trust and
trusting behaviours as “trust is the willingness to assume risk” and “behavioural trust is
assuming of risk” (Mayer et al., 1995, p. 724). A similar distinction is also made by McKnight
and Chervany (2001a), who state that taking risks is a behaviour while trust is the willingness
to depend on another, based on the expectations or beliefs of the truster about the object of trust.
Rosseau et al. (1998, p. 395) also follow this separation, as they state “trust is not a behaviour
(e.g., cooperation) or a choice (e.g., taking a risk), but an underlying psychological condition
that can cause or result from such actions”. This view also recognizes that behaviour may be
related to other factors as well as trust. For example, actual consumer trusting behaviour (e.g.,
cooperation or reliance) can be influenced by other variables such as situational (e.g., power of
control) and personal factors rather than trust in others (Moorman et al., 1992). Thus, trust is
an imperfect predictor. According to Ripperger (1998), trusting behaviour can be best
characterized as the risky advance concession in the hope for a favourable outcome, without
any explicit contractual security or control measure against possible opportunistic behaviour.
The author recognizes that the probability that an individual makes a risky advance concession
in a specific context depends on trust, but also other circumstances. Including trusting
behaviour, which maybe is not based on trust alone, would therefore result in a measurement
problem. Therefore, this thesis views the willingness to act on the trustee as a more appropriate
indicator of trust in a relationship than actual trusting behaviour. Further, trust is regarded as
one of the antecedents of risk-taking behaviour in any relationship, in which the form of risk-
taking, according to Mayer et al. (1995), is dependent on the situation. Behaviours, therefore,
appear to be inadequate for measuring general orientations of a person. This approach also
parallels the established literature on attitudes, which generally postulates a separation of the
constructs but also their relatedness (Hogg & Cooper, 2003; Hogg & Vaughan, 2011).

89
Conditions of Trust

In literature, there seems to be a general agreement that expectation, confidence, risk,

vulnerability and exploitation, as well as interdependence (amongst others), are conditions for
trust to develop and become operational (e.g., Corritore et al., 2003). Some of these concepts
have been already discussed. That is, various authors have focused on the concept of
expectation in their definitions of trust (e.g., Deutsch, 1962; Rotter, 1971; Baier, 1986). On the
other hand, others view trust as some kind of confidence (e.g., Deutsch, 1958; Lewicki &
Bunker, 1995). For instance, Barney and Hansen (1994) conceptualize trust as “the mutual
confidence that no party to an exchange will exploit another’s vulnerabilities” (p. 177). This
definition is valuable, since it guides this thesis’ investigation of trust to the necessary
conditions that cultivate trust in human affairs.

Most scholars agree that risk is a condition that must exist for trust to arise and to influence
choice as well as behaviour. The central role of risk in the conceptualization of trust is
emphasized in psychology, sociology, and also in the economic view on trust (Coleman, 1990;
Doney & Cannon, 1997; Lewis & Weigert, 1985; Rotter, 1967). Hence, risk has also been
considered as an important element in various trust definitions. For instance, Deutsch (1962)
describes trust as “the willingness of an individual to behave in a manner that assumes another
party will behave in accordance with expectations in a risky situation” (cited in Corritore et al.,
2003, p. 741). The author states that if there is no risk, then the truster is not forced to make an
assessment of the trustworthiness of the trustee (Deutsch, 1958). Analogously, Mayer et al.
(1995) conclude in their interdisciplinary review that there is no need for trust if there is no risk
in a situation. Risk can be defined as the decision-maker’s interpreted, perceived probability of
loss (Chiles & McMackin, 1996). The causal relationship between trust and risk is due to a
reciprocal relationship. A key function of risk is that it leads to an opportunity to form trust,
which then leads to risk acceptance. Moreover, risk taking buttresses a sense of trust when the
expected behaviour materializes (Coleman, 1990; Rosseau et al., 1998). The close relationship
between trust and risk is also evident in the typical conceptualization of risk, as it has been
interpreted to have two distinct dimensions: (1) uncertainty and (2) seriousness of potential
outcomes (Bauer, 1967). Trust scholars often refer to these two components when they
explicitly or implicitly argue that trust has three major constituents: (1) uncertainty about the
outcomes of an interaction; (2) personal harm as a possible outcome of the interaction; and (3)
lack of influence on the outcomes (Petermann, 1996). Similarly, Deutsch (1960) suggests that
trust arises under conditions when the following conditions are fulfilled: (1) the course of action
in the future is ambiguous; (2) outcome occurrence relies on the behaviour of others; and (3)
strength of the harmful event is greater than that of the beneficial event. Accordingly, most
scholars agree that trust would neither develop nor become operational if one’s actions could
be undertaken with complete certainty and, hence, without risk (Lewis & Weigert, 1985).
Interaction agents (e.g., consumers) face multiple sources of uncertainty in social interactions;

90
for example, uncertainty about the quality of information, the consequences of a transaction
(Cox & Rich, 1964; Gefen & Straub, 2003), the amount at stake (Cho & Lee, 2006), and the
uncertainty about whether the interaction partner will abuse the trust given to his/her person or
not (Rosseau et al., 1998).

Closely related to risk is the concept of vulnerability, as it has been regarded as a function of
the former concept (Lee & Soberon-Ferrer, 1997). The first author who emphasized the
centrality of vulnerability (and/or exploitation) in trust definitions was Deutsch (1958), who
has presented that the individual “perceives that he/she will be worse off if he/she trusts and
his/her trust is not fulfilled than if he/she does not trust”. Therefore, trusters would lay
themselves open to the trustee and allow themselves to be vulnerable, while hoping that the
trustee will meet their promise (Tan & Sutherland, 2004). According to Smith and Martin
(1997), vulnerability refers to susceptibility to injury or to being taken advantage of by another
party. Vulnerability means that the truster must be exposed in some way (Corritore et al., 2003).
For instance, while online shopping, a consumer can be exposed because of his/her lack of
knowledge or expertise, or the inability to purchase goods without the assistance of others.
Some trust authors argue that while expectation, confidence, and risk are all essential conditions
of trust, they alone would not be sufficient for trust to arise (Luhmann, 1979; 1988; Muir, 1994).
These scholars suggest that vulnerability, with the concomitant possibility of exploitation, has
to be an integral part in reasonable trust definition (Corritore et al., 2003; Deutsch, 1962; Mayer
et al., 1995; Zand, 1972). For example, Sabel recognizes the centrality of the concept of
vulnerability in his trust conceptualization, as he defines trust as “the confidence that no part to
an exchange will exploit another’s vulnerabilities” (Sabel, 1993, p. 1133). Luhmann (1979, p.
42) states “one fundamental condition of trust is that it must be possible for the partner to abuse
the trust; indeed it must not merely be possible for him to do so, but he (she) must also have a
considerable interest in doing so”. Additionally, besides one’s own vulnerability to the
misconduct of others, the dependence upon other people impacts a person’s need for trust
(Deutsch, 1958, Gefen, 2000; Luhmann, 1979; Rosseau et al., 1998). In essence, these scholars
agree that the greater the dependence upon other people, the greater is the need for trust in
general. Hence, these and various other trust scholars have recognized that interdependence,
interpreted the way that one party’s interests cannot be achieved without reliance upon another
party, is another necessary condition for trust. Rosseau et al. (1998) states that although both
risk and interdependence are required for trust to develop, the nature of risk and trust changes
as interdependence increases. The authors further recognize that as risk and interdependence
over the course of a relationship change, both impact the level and potentially the form that
trust takes between those parties. That is, the trust a firm places in a temporary worker is quite
distinct from trust associated with a veteran worker (Hudson, 2004).

91
Conclusions
The prior review shows that a multitude of trust conceptualizations and operationalizations have
evolved over the last decades and although variations are obvious, several conclusions can be
drawn reflecting some general consistency among trust researchers about the nature of the
construct.

First, trust is essential for all relational exchanges and is a critical precondition that enables
people to deal with the physical as well as social world (e.g., Mayer et al., 1995; Ring, 1996;
Rosseau et al., 1998). Specifically, trust is a mental mechanism that takes away complexity,
risk, and uncertainty from interactions and triggers human behaviours that otherwise would be
prohibited. As complexity arises in multiple aspects of life, humans can develop trust towards
subjects as well as objects. While the basic nature of the trust concept seems to be stable across
contexts, the kind of trust object as well as the kind of relationship to it both ultimately
determine what has to be included in the trust concept. This thesis was able to identify several
regularities that typically arise when people have interpersonal trust.

Second, it is a subjective, psychological state which is associated with a number of mental

structures or components that are necessities for trust to exist. Missing one of the important
elements of trust means missing the true nature of the concept. The relative importance of the
components is context and relationship specific; however, trust is regularly a product that is
mirrored in positive expectations and trusting beliefs as well as positive trusting attitudes
towards the trust object. While expectancies refer to the future orientation of trust, beliefs
reflect the critical role that judgments about the other party play in trust. Hence, trust can be a
generalized expectation about others’ reliability (Rotter, 1967); a commonly shared set of
expectations among social agents that social rules and the institutional environment work
properly (Barber, 1983; Lewis & Weigert, 1985; Shapiro, 1987; Zucker, 1986); and the truster’s
expectation that the trustee will behave appropriately and in the best interest of his/her
interaction partner (Gambetta, 1988; Rempel et al., 1985. Additionally, to trust means that one
has the positive expectation that a future event will not be disadvantageous to the trustee
(Deutsch, 1958). Hence, there is an obvious connection between trust and an individual’s
optimism that a beneficial event will take place (Golembiewski & McConkie, 1975). The
beliefs can be combined in beliefs towards the trustee’s (i) abilities, (ii) integrity, and (iii)
benevolence.

Third, trust is strongly linked with and inseparable from confidence. Among scholars there is a
general consistency that trust arises from the truster’s confidence in the trustee’s characteristics,
i.e., his/her trustworthiness (McAllister, 1995; Moorman et al., 1993; Morgan & Hunt, 1994,
as well as his/her confidence in the relationship strength and quality (Rempel et al., 1985; Swan
et al., 1988).

92
Fourth, two different kinds of trust basically exist, as trust can represent a relative stable,
generalized attitude developed from prior personal experiences (i.e., trust as a personal
characteristic) but can also be build up situationally/contextually through cognitive processes
and thoughts without any prior familiarity with the trust object (i.e., situational trust). Both
positions – that trusting beliefs and attitudes about another party’s trustworthiness and
confidence in the relationship are formed based on knowledge or earlier experiences with the
object of trust (Hosmer, 1995), as well as the assumption that trust can also arise due to initial
interaction (e.g., Chen & Barnes, 2007; Fisher & Chu, 2009; Fisher & Chu, 2009 – are well
supported in literature. This accords with the personality vs. the situational school of thought
of social psychology (Hogg & Vaughan, 2011).

Fifth, another critical trust component is the individual’s willingness to be vulnerable under
conditions of risk and interdependence. One of the most frequent definitions, namely that of
Mayer et al. (1995), defines trust as the “willingness to be vulnerable”. Other authors mean the
same thing but with different words. For instance, Doney, Cannon and Mullen (1997) refer to
the truster’s “willingness to rely” on another. Literature suggests that risk, vulnerability,
uncertainty, and interdependence are necessary conditions for trust (e.g., Rosseau et al., 1998;
McKnight & Chervany, 2001). For instance, Deutsch (1958) proposes that one has no reason
to trust an interaction partner in situations characterized by complete information and control
over the other. In such situations, one would simply rationally predict the other’s future actions
rather than trust the other (Lewis & Weigert, 1985a). Trust is only required in an environment
that is uncertain and risky (Wang & Emurian, 2005). Deutsch (1958) states that trusting another
person means that the truster must be willing to make him-/herself vulnerable in order for trust
to become operational. The truster has to accept the risk of losing something that is important
to him/her and relying on the trustee means not to be exploited out of this vulnerability.
Accordingly, trust can be characterized as a behavioural intention to accept vulnerability based
on favourable expectations about the intentions and behaviours of the interaction partner in a
risky situation. Analogously, interdependence is another condition for trust, where the interests
of one party cannot be attained without reliance upon someone else (Rosseau et al., 1998).
Therefore, trust is likely to be established in any context which is characterized by risk,
uncertainty and interdependent relationships. In fact, the need for trust increases as individuals
have less information and experience in dealing with the object of trust (Aldrich & Firol, 1994).

Sixth, trust has cognitive, affective, and conative elements. Even though various definitions of
trust are strongly influenced by earlier conceptualizations which heavily emphasize on the
truster’s rational thinking about the characters that make an interaction partner to be received
as trustworthy, a growing literature string emphasising the three-component view of trust has
provided evidence that also affective processes and behavioural intentions to trust, together
with cognitive beliefs, are a necessity for trust to exist (e.g., Johnson & Grayson, 2005). This
conforms to the classic three-component view of attitude (Hogg & Vaughan, 2011). For

93
instance, Lewis and Weigert (1985a, b) advance that trust manifests itself by both feelings as
well as rational thinking. The exclusion of either one of these two dimensions would
misconceptualize trust, since cognitive-based trust without an emotional base is nothing else
than rational prediction, while emotion-based trust without cognition is pure faith. Cognition
and affection are, however, likely to be intertwined. Additionally, Morgan and Hunt (1994)
propose that beliefs as well as the willingness to rely on the truster both have to be present.
Otherwise, trust would be limited. Hence, eWOM trust is likely to be composed of cognitive-
and affective-based trusting beliefs and attitudes towards the object of trust, combined with the
willingness to rely on the information conveyed in such messages.

Seventh, the type of the social relationship among trust protagonists is likely to determine the
relative importance and strength of the cognitive versus the affective elements of trust (Lewis
& Weigert, 1985). While emotion-based trust (e.g., judgments about trustee benevolence,
feeling of likeability) is most influential in long-established primary-group relationships (e.g.,
among members of a kin) (Johnson-George & Swap, 1982; Rempel et al., 1985), cognitive-
based trust is more extensive and contributes more to the development of trust in relationships
with secondary groups (e.g., unknown individuals, social systems).

In the next chapter, the domain-specific nature of trust in eWOM is discussed in detail and the
construct is conceptualized. This is followed by the study’s research questions and hypotheses.

94
Chapter 3

Conceptualization of the Construct,

Research Questions and Hypotheses

In this chapter – after discussing the role of trust for online review impact – this thesis presents
a literature-based definition and conceptualization of eWOM trust. Hereafter, thirteen research
questions and fifteen hypotheses are proposed in order to assess multiple forms of construct
reliability and validity, including translation-, criterion-, and nomological validity. For testing
the latter, a comprehensive theoretical framework of eWOM trust including its antecedents,
consequences, and correlates is presented. Subsequently, this chapter outlines additional
assessments of scale generalizability. The chapter closes with a discussion of the applicability
of the new eWOM trust scale (eWT-S). Here, it is proposed that the new measurement
instruments offers – together with an assessment of consumers’ online ad trust – a valuable tool
to segment online information recipients.

Evidence for Generalized eWOM Trust

Trust has been found to be an important factor that determines the nature of interpersonal
relationships and enables its conduct (Ridings et al., 2002). Additionally, its critical role for
social exchanges in the presence of risk and uncertainty is emphasized throughout academic
literature (Gefen, 2000; Jarvenpaa et al., 1998). Trust is an important facilitator of information
sharing among individuals as well as a primary cause for the development of new relationships.
This is true for face-to-face interactions that take place every day (e.g., Coppola, Hiltz & Rotter,
2004; Gabarro, 1978; Piccoli & Ives, 2003), but also for impersonal online interactions (e.g.,
Gefen, 2000; Gefen et al., 2003; McKnight et al., 2002a). Here, during the last decade the
research interest has been focused on the topic of consumer-vendor relations (Pavlou &
Fygenson, 2006). However, more recently a new topic has emerged that targets the significant
role of trust in the information given in eWOM reviews and recommendations (e.g., Hsu et al.,
2007; Ridings et al. 2002; Urban et al., 2009).

95
© Springer Fachmedien Wiesbaden 2017
W. Weitzl, Measuring Electronic Word-of-Mouth Effectiveness,
DOI 10.1007/978-3-658-15889-7_3
Trust in online social interactions (eWOM) can be best viewed in this context as an
interpersonal relationship or relational exchange of communication content. For various years,
consumer reviews are an inseparable element of the online shopping experience. OCR have
become a socially significant object. This enabled the development of generalized behavioural
patterns (e.g., eWOM usage, adoption) but also of relatively stable attitudes such as eWOM
trust towards this form of online market communication. Before discussing the functions and
key role of eWOM trust, it shall be noted that this thesis assumes that eWOM trust is not trust
in reviewers per se, but a generalized orientation towards the information provided by these
reviewers. That is, trust in the product of the source. Hence, it includes trust in other’s product
experiences, evaluations, and recommendations. However, it is assumed that eWOM trust is
inseparable from trust between people (interpersonal trust) (Rotter, 1971) or what Luhmann
(1988) calls “personal trust”. This means that consumers of eWOM develop a uniform attitude
towards the product (i.e., reviews and recommendations) and the source (i.e., reviewers or the
review community). While social shopping, a consumer normally doesn’t interact with a
specific other when exchanging market information; instead he/she interacts with a group of
mostly unknown shoppers that form the community of earlier/present/potential customers of a
product. Hence, the consumer interacts passively with this group, when he/she reads customer
reviews from a group of persons, but also potentially actively when he/she posts reviews to the
general public or responds to specific reviews. In both cases, trust is a necessity. Therefore,
trust in eWOM has to develop between the individual and this community in order to enable
and foster social exchange. That is, for a person who wants to engage actively in or consume
passively eWOM communication, trust has to exist on the generalized collective (Ridings et
al., 2002) or institutional level (i.e., trust in reviews in general) rather than on the individual
level (i.e., trust towards single reviewers). For instance, writing reviews and giving personal
recommendations to fellow shoppers is an apparent social action which bears the potential of
being embarrassed by others in the form of negative comments. Trust in the community relieves
the writer from this burden or minimizes the perceived risk which otherwise would deter
him/she from involving in relational exchange. eWOM trust is therefore a key facilitator of this
interpersonal relationship. This corresponds to various research contributions that have applied
the notion of interpersonal trust to collective entities (e.g., Jarvenpaa & Leidner, 1998) or
institutions (e.g., Soh et al., 2009) earlier – OCR are a socially constructed institution.

For consumers, eWOM trust has an important purpose. It is an imperative that in cases where
strong trust between people exists, they are also more willing to get involved in cooperative
interaction or shared activity (Gambetta, 1988; Nahapiet & Ghoshal, 1998). In the context of
online reviews and recommendations, this cooperative interaction is represented on the side of
the review reader by his/her tendency to collect, use and accept eWOM information and adapt
their purchase according to the new insights granted by the reviewers. It’s common sense that
consumers who think that customer reviews are both trustworthy and reliable also attribute a
higher information value to this kind of market information, which further inclines them to
96
collect information from this source (Fang & Chiu, 2010; Ridings et al., 2002). Consequently,
this trusted and valued information is very likely to influence subsequent purchasing decisions
(e.g., Pan & Chiou, 2011) – especially in the absence of alternative information. This claim is
well supported in literature. For instance, for Blau (1964), knowledge-sharing activities result
from social exchange relationships which are influenced essentially by trust. Other scholars
agree with Blau that there exists a causal relationship between trust and information sharing
(e.g., Adler, 2001; Chen & Hung, 2010; Dyer & Chu, 2003, Fang & Chiou, 2010). For Dodgson
(1993), trust relationships are substantial for effective communication. The same is assumed to
be true for eWOM communication where eWOM trust enhances the quality of the dialogue and
the persuasiveness of the eWOM messages.

Individuals develop trust in eWOM because this saves cognitive energy (Hogg & Vaughan,
2011). It gives us a basic orientation towards the trust object. More specifically, eWOM trust’s
basic function is its vital role for reducing risks typically associated with customer review
information and hence social complexity (Luhmann, 1979). In online social shopping,
consumers typically look for cues which help them to select the right product, vendor etc. The
positive/negative experiences of others with market offerings are an important aid and often the
key reason for selecting/not selecting it. Similarly, recommendations are an important decision
aid, often triggering the final purchasing decision. Depending on these reviews and
recommendations therefore automatically implies accepting a set of various risks with respect
to the product (e.g., financial risk) but also – and inseparable from this – risks associated with
the reviews themselves (e.g., risk of deception, risk of being manipulated). Hence, the risks and
dependencies that consumers face are typically not insignificant. And consumers apply
mechanisms such as trust to reduce these risks. At its heart, to trust means that the truster thinks
that the trustee will not exploit his/her vulnerabilities (Casalo et al., 2008; Corritore et al., 2003).
Various authors emphasize that a reduction of perceived risk is an important result of the trust-
building process (Mitchell, 1999). Implicitly, the same scholars acknowledge that for the
development of trust, uncertainties about the social relationship have to be present that lead to
certain perceptions of risk, vulnerability, and interdependence (Casalo et al., 2011). Trust’s
function as risk reducer is especially important in the context of virtual knowledge exchange,
as it shall “rule out” opportunistic behaviours (Ridings et al., 2002).

The risks associated with OCR originate from their specific characteristics. For instance,
customer reviews are typically provided by strangers with whom the consumer has never
interacted before and very likely will never meet in future. Due to the reviewers’ and
recommenders’ anonymity it is reasonable for consumers to have doubts about the
trustworthiness of these messages and about the creators’ basic motivations. Is the reviewer
driven by true altruism or are there intentions to steer other shoppers wrong? Besides the lack
of face-to-face contact, the identities of the reviewers may be suspect. For instance, employees
of firms can impersonate regular reviews and fake the content of reviews easily. As they purely

97
communicate information that is in the interest of the product’s producer, the reviews are no
longer independent, objective and credible (Bray & Schetzina, 2006; Burgess et al., 2011). The
opportunity to falsify customer reviews has been picked out by earlier research contributions
(Litvin et al., 2008), but also by the media, which has uncovered unethical business tactics to
disguise advertising as proper eWOM communication (e.g., Gee, 2011; Moe, 2012; Streitfeld,
2011; Tuttle, 2012). Therefore, consumers have good reasons to be sceptical towards eWOM
and refuse this kind of information because of its latent risks. Believing and applying eWOM
for purchase decisions can have serious negative effects. For instance, the recommended
product turns out to perform poorly (functional risk) or the usage of a recommended product
puts social pressure on the consumer (social risk).

The theoretical basis for explaining the need for a psychological construct which can be named
(generalized) eWOM trust is supported by various disciplines but particularly by social
psychology. In addition, there is some evidence that consumers tend to develop trust in eWOM
as a way to reduce involved risks, interdependencies and vulnerabilities. For instance, according
to a study realized by Nielsen (2012) 70% of consumers at least say that they in general trust
and use online customer reviews. Other research institutions provide similar figures (e.g., Cone,
2011). Various studies (cited before) state that trust in reviews is the key trigger to show positive
behaviours towards them and buying the recommended products – overcoming review and
product risks.

We have seen that the characteristics of eWOM lead to conditions of risk, uncertainty, and
dependence which makes the decision of the consumer how to implement review information
in the purchasing process quite complex. Trust is an efficient mental mechanism of the human
mind to reduce this otherwise overwhelming complexity. Since shoppers are not able to
diagnose the quality and the background of every single review, they are urged to develop a
constant mental orientation – or in other words – a certain level of generalized trust towards
this kind of market communication. This orientation is triggered by generalizations about
reviews but also their creators (i.e., reviewer stereotypes) that are reasonable in the eye of the
consumer. eWOM trust determines the way of dealing with the eWOM world in a consistent
manner and is, as such, deeply rooted in the consumer’s personality.

As eWOM trust ultimately determines communication persuasiveness as well as consumer

reactions towards reviews and recommendations (buying behaviours), it is reasonable to
measure the concept adequately. Academic as well as managerial studies usually apply
imprecise operationalizations of the eWOM trust construct and are not able to detect varying
degrees of trust. For instance, the above statistics about eWOM trusters may include persons
who may label online reviews as sincere but are still not willing to rely on them. Not
discriminating between those two types of groups may lead to a too optimistic impression about
the effectiveness of eWOM messages in general. Others may reject eWOM completely, as they

98
find this kind of information totally unbelievable, while others describe themselves as not
trusting but unconsciously take a moderate stance. A precise measure of eWOM trust enables
us to differentiate among those consumers and to anticipate the impact of reviews and peer
recommendations.

Conceptualizing Trust in eWOM

eWOM trust is understood in this research as a latent construct which precedes consumer
eWOM behaviours and acts as a guidelined related choices. Diamantopoulos et al. (2008, p.
1204) define latent constructs (or variables) as “phenomena of theoretical interest which cannot
be directly observed and have to be assessed by manifest measures which are observable”.
According to Nunnally and Bernstein (1994, p. 85), “such a variable is literally something that
scientists 'construct' (put together from their own imaginations) and which does not exist as an
observable dimension of behaviour”. Hence, a comprehensive conceptualization of the
construct of interest as a first step is critical in any measure development process (DeVellis,
2012; Kaplan, 1964; Netemeyer et al., 2003; Schwab, 1980). By following the notion of James
and Jones (James & Jones, 1974), Schwab (1980) emphasizes that the definition of the
conceptual meaning and the theoretical boundaries of a construct should have supreme priority
in substantive research endeavors. Albers and Hildebrandt (2006) state that the importance of
a clear conceptual definition often remains unrecognized by researchers. Hence, in order to
overcome this problem, the following section aims to clarify the domain of the discussed
construct, the construct itself and its composition. By building on theoretical insights and
conclusions of the previous literature review, this thesis attempts to provide a conceptual
framework that enables a multi-layer analysis of eWOM trust and allows the integration of key
trust concepts and different types of trust that have their theoretical origins in different academic
disciplines.

Construct Definition and Domain

At the center of this thesis is its definition of eWOM Trust that recognizes the complex higher-
order, multi-dimensional (Butler, 1991; Ganesan, 1994) as well as the domain-dependent
(Luhmann, 1979; Rosseau et al., 1998) nature of the trust construct. A deep investigation into
prior trust literature led to the identification of the concept’s critical ingredients which seem to
be essential for a profound definition of the eWOM trust concept. Hence, in the identification
phase of this research, eWOM trust was preliminarily defined as a consumer’s general
confidence that the information conveyed in online customer reviews is reliable. More
specifically, eWOM trust is formally defined as a five-dimensional, second-order construct
capturing the extent a consumer has the belief that information given in online customer reviews
is (a) honest, (b) useful, and (c) benevolent; (d) has a favourable attitude towards this kind of
market information; and (e) is willing to rely on online customer reviews in general. Hence,

99
eWOM trust is broadly regarded as a person’s disposition toward having belief in online review
claims and recommendations in general.

eWOM trust is a generalizable trait-like attitude that predisposes individuals to rely on the
opinions and consumption recommendations expressed by fellow shoppers on diverse
platforms enabling consumer-generated content (e.g., retailer sites, review sites, blogs, online
newsrooms, forums). According to Allport and Odbert (Allport & Odbert, 1936) traits are
“generalized and personalized determining tendencies – consistent and stable modes of an
individual’s adjustment to his environment” (p. 26). Accordingly, eWOM trust is recognized
as an individual difference variable and a single consistent response tendency. eWOM trust is
a constant, long-lasting and internally caused psychological construct that can be observed
across a wide range of situations and contexts (John & Srivastava, 1999). This means that
consumers with a low eWOM trust, regardless of the situation, may be impossible to influence
by any means by the information given or arguments conveyed in online customer reviews,
because they would not believe any stated claims (Obermiller & Spangenberg, 1998).
Conceptualizing eWOM trust as a kind of personality trait coincides with the acceptance of the
psychological claim that some mental structures are stable over time. There exists support in
trust and personality literature to assume that if a person is a high truster at the present point of
time he/she will likely be a high truster later. However, the correlation is not assumed to be
perfect and at least some change is supposed to happen. Therefore, the person could also
become less trusting as an outcome of a social learning process. However, a person’s timely
ordered earlier position on the eWOM trust continuum will be significantly related to where the
person finds himself/herself in the future. This justifies describing eWOM trust as a relatively
stable trait. “Relative” as it is acknowledged that various conditions (e.g., gathering of negative
reviews experiences) can lead to changes or variations in the consumer’s orientation and the
time span of this change being presumably shorter compared to other personality traits.

This said, it has to be noted that the conceptual domain of eWOM trust is limited to text-based
online customer or consumer reviews as well as recommendations of products, brands,
manufacturers, or retailers. It is not this thesis’ intention to quantify trust in any form of peer
communication on the Internet per se, such as informal conversations among consumers in
online forums or social platforms. However, it is recognized that reviews – as defined earlier –
can appear in diverse forms and on miscellaneous platforms. Further, eWOM trust is understood
as a form of implicit interpersonal trust (Rotter, 1971), which exists between an individual (i.e.,
the consumer) and a collective entity of fellow shoppers that are typically unknown to the
consumer. Specifically, it represents a generalized trust toward the information or
recommendations provided by others (i.e., the reviewer community) and not toward a specific
person or review (Lu et al., 2010; Ridings et al., 2002). This ultimately determines the
components and scope of the construct.

100
It can be shown that this thesis’ trust definition is consistent with literature’s earlier
conceptualizations. First, this study postulates that confidence is an integral ingredient of its
trust definition. In trust literature, the concept of confidence frequently surfaces, as many
scholars have decided to include this aspect in their definitions and conceptualizations of trust
(e.g., Deutsch, 1960; Gefen, 2000; Johnson-George & Swap, 1982; Moorman et al., 1993).
Barber (1983), for instance, views confidence as favourable expectations or predictions held by
the truster that the trustee will benefit the truster. Confidence has been regularly regarded as an
attitude (Fazio & Zanna, 1981; Raden, 1985) based on cognition and affect. Here, it is
sometimes argued that confidence implies a person’s readiness toward a specific object and
therefore the concept parallels the narrow definition of attitude (Kim & Tadisina, 2007).
Numerous authors also included similar concepts, such as expectations (Baier, 1986; Barber,
1983; Good, 1988; Rotter, 1971), confident expectations (Lewis & Weigert, 1985, or subjective
probability (Gambetta, 1988). As eWOM trust is conceptualized as a positive expectation that
stems from a cognitive as well as emotional evaluation of internal beliefs and attitudes, as well
as a positive intention towards peer information, the concept of confidence (i.e., the attitude
that relying on eWOM information is safe) seems to be appropriate to mirror the true nature of
the construct. Trusting eWOM users are best characterized by their sanguine attitude that the
information given in reviews is authentic and acting on this basis of knowledge is safe and
advantageous.

More specifically, in this thesis, eWOM trust is defined as the generalized confident expectation
that online customer reviews are reliable. Here, reliability is defined as the consumer’s belief
that eWOM consistently keeps its promises and hence is trustworthy. In literature, reliability
has often been operationalized in terms of dependability, credibility, and consistency (Butler &
Cantrell, 1984; Johson-George & Swap, 1982). For instance, McGregor (1967, p. 164) refers
to reliability by noting “inconsistencies between words and actions decrease trust” and also
Ouchi (1981) conceptualized trust in terms of an individual’s beliefs in the consistent or reliable
behaviour of the trusted party. In marketing literature, Swan et al. (1988) identified reliability,
which they labeled dependability, as one of the four distinct dimensions of trust. A consumer
perceives eWOM to be reliable, if eWOM trusters believe that OCR is sincere about the claims
they make about the discussed or recommended object (e.g., product, service, brand, and
vendor) and they expect that these claims are very likely to be fulfilled by this object later.
When an information recipient trusts ORC, this means – in general – that he/she is confident
that his/her reliance is not capitalized.

Second, this thesis recognizes that the trust construct has more than one conceptually
distinguishable facet and, therefore, defines eWOM trust as the psychological collectivity of
relatively stable confident beliefs, favourable attitudes, and well-meaning behavioural
intentions towards online customer reviews under conditions of uncertainty, risk and
interdependence that are inherent in online social interactions of this kind. In trust theory, a

101
two-way literature stream of trust conceptualizations typically emerges which mirror such an
understanding. While the first approach concentrates on trust as an expectation, belief or
attitude towards the current or future behaviour of the object of trust (Barber, 1983; Luhmann,
1979), the alternative approach associates trust with a person’s acceptance of and/or exposure
to vulnerability (Doney et al., 1998; Mayer et al., 1995; Rosseau et al., 1998). The thesis at hand
integrates both views by including five dimensions of trust that the literature has most
frequently identified: (1) integrity/honesty; (2) ability/usefulness; (3) benevolence; (4)
likeability; and (5) willingness to rely on. The first four factors of trust represent specific trusting
beliefs and attitudes. On the other hand, the willingness to rely on dimension deals with the
consumer’s acceptance of a risk inherent in any eWOM message posted from more or less
unknown other consumers on the Internet. High eWOM trust should be manifested in each of
these dimensions.

Third, this thesis’ definition also shares the similarity with established trust conceptualizations
that trust consists of cognitive (i.e., the knowledge or belief component), emotional (i.e., the
affective response towards the trust object), as well as conative elements (i.e., behavioural
tendencies) (e.g., Barber, 1983; Lewis & Weigert, 1985; Luhmann, 1979; Mishra, 1996). These
scholars heavily draw on the original conceptualization of attitude which follows the “three-
component” or “tripartite” view (Katz & Stotland, 1959; Krech et al., 1962; Ostrom, 1968;
Rosenberg & Hovland, 1960). High eWOM trusters therefore base their trust on cognitive
judgments, emotional bonds and the readiness to use OCR in risky shopping conditions.

The proposed definition also separates eWOM trust from similar constructs, such as the more
general attitude toward eWOM. Later it is argued that both constructs are conceptually
separated but nevertheless related. For instance, a consumer may dislike eWOM in general
because he/she does not trust this form of market information. Hence eWOM trust is seen as a
critical correlate of positive/negative eWOM attitude.

While being conceptualized as an individual difference factor, it is recognized that an

individual’s trusting stance in online reviews and recommendations is influenced by the human
socialization process (Fishbein & Ajzen, 1975), including prior consumer experiences that
make a consumer think and feel that he/she can rely on the words of others. Instrumental
conditioning is one behavioural approach to explain the formation of trust. When the reliance
on OCR (i.e., purchasing a socially recommended product) is followed by positive
consequences (e.g., the product meets the consumer’s expectations), this trust is reinforced and
trusting behaviours are more likely to be repeated. However, when the reliance on OCR is
followed by negative consequences (e.g., the product doesn’t meet the consumer’s expectations
as the reviews turned out to be manipulative, false and opportunistic), trust decreases (Hogg &
Vaughan, 2011). However, changes in the level of trust only appear slowly. People have a
natural tendency to trust eWOM – like their natural tendency to trust/distrust other people – and

102
hence it is quite fixed. However, this disposition coexists with situational eWOM trust when
consumers face online reviews. Both types of trust may be influenced by the situation and both
are assumed to have an interdependent relationship. For instance, it is accepted that the impact
of generalized trust in eWOM on situational trust may vary to some degree due to situational
variables that a consumer faces in a consumption actuality (e.g., perception of reviewer’s
motives) – thereby induce a temporary state of trust (this view is consistent with the related
works of Forehand and Grier (2003); Obermiller and Spangenberg (1998)). Hence, as such it
can be temporarily, briefly and externally influenced (John & Srivastava, 1999). Nevertheless,
a person’s generalized tendency to trust eWOM remains relatively stable and dominant in the
most part and essentially determines situational trust (i.e., the consumer’s reliance in the
specific review) as it sets the level of potential trust (McKnight et al., 1998). That is, a person
who has a general low propensity to trust eWOM will never show the same level of trust in a
specific review as a person who is generally more trusting. It is agreed that individual reaction
to reviews is ultimately determined by situational trust. However, situational trust is itself only
a mediator of internalized trust. Human reactions to eWOM are to a large degree determined
by an individual’s global orientation towards it.

What follows is a more detailed discussion of the various components that are theorized to be
responsible for the very own nature of a construct that can be called generalized eWOM trust.

Construct Composition

This thesis strives to integrate key trust concepts under a single, common definition of
interpersonal trust. Here, the framework advanced by Kim and Tadisina (2007) is regarded as
a suitable basis for this objective, as it has identified the essence of trust across interdisciplinary
research. By adopting this approach to the context of eWOM information and recognizing this
thesis’ earlier discussions, Table 4 summarizes the key components of eWOM trust.

103
Table 4: Theorized Key Components of eWOM Trust

Types of Trust Dimensions Definition Elements

Trusting Beliefs Ability/ The consumer’s belief that eWOM content is a Cognitive
Usefulness useful, knowledgeable and adequate source of
purchase-relevant information and/or
recommendations.

Integrity/ The consumer’s belief that the information Cognitive

Honesty conveyed in eWOM is believable, truthful,
verified, and adheres to moral standards.

Benevolence The consumer’s belief that eWOM is guided or Emotional

motivated by the reviewer’s favourable and
positive intentions toward the consumer’s welfare
and not by the reviewer’s self-interest only.

Trusting Likeability The consumer’s positive emotions toward eWOM. Emotional

Attitudes

Trusting Intention Willingness to The consumer’s willingness to rely on eWOM and Behavioural
rely on accepting potential loss by applying eWOM- (Conative)
conveyed information for purchasing decisions.

Types of eWOM Trust

Earlier research typically proposes a trust typology that comprises different types of trust that
commonly manifest a person’s level of trust. Within literature, trust researchers distinguish
between the constructs of people’s (1) beliefs (i.e., trusting beliefs) and (2) attitudes (i.e.,
trusting attitudes) about the object of trust, as well as their (3) willingness to rely on the trustee
for making a decision or performing a specific action (i.e., trusting intentions), coupled with a
sense of vulnerability or risk in cases where the trust is violated (Luhmann, 1979; Mayer et al.
1995; Rosseau et al., 1998). For example, Gefen et al. (2003) note that trust should be best
viewed as a combination of a set of distinct beliefs, attitudes or “feelings of confidence and
security” (Rempel et al., 1985, p. 96), and “the willingness of a party to be vulnerable to the
actions of another” (Mayer et al., 1995, p. 712). This thesis has earlier recognized this view by
referring to it as the concept of the “trust triangle” (see Chapter 2). The trust literature provides
sound evidence that numerous and often-cited researchers have conceptualized trust as a
composition of trusting beliefs which represent the truster’s expectations about the trustee,
based on certain characteristics which the object of trust possesses or lacks (Becerra &
Korgaonkar, 2011; Bhattacherjee, 2002; Kee & Knox, 1970); McKnight & Chervany, 2002)
and trusting intentions (Lewis & Weigert, 1985; McAllister, 1995; McKnight et al., 1998;
Moorman et al., 1993). The latter can be understood as the extent to which a person is willing
to be vulnerable or take a risk in a relationship. Or, in other words, an individual shows trusting
intentions when he/she is willing to depend on or intends to depend on the trusted party in a
given situation (Deutsch, 1962; Kaplan & Nieschwietz, 2003; Mayer et al., 1995; Stewart,
104
2003). For example, according to Zand (1972), trusting refers to an individual’s willingness to
take a risk in the relationship with another person. In a similar manner, Moorman et al. (1993,
p. 82) define trust as the “willingness to rely on an exchange partner” and conclude that trust
becomes more important under conditions of vulnerability and uncertainty. In their often-cited
work, Mayer et al. (1995, p. 712) define trust as “the willingness of a party to be vulnerable to
the actions of another party”. As a result of their interdisciplinary review, Rosseau et al. (1998;
p. 395) conceptualize trust as an individual’s intention to accept vulnerability in situations
characterized by risk and interdependency. McKnight et al. (1998) emphasize a two-side nature
of trust by viewing the concept as consisting of trusting beliefs and trusting intentions. Here,
the authors define trusting intentions as the truster’s readiness to depend on the trustee in a
given situation. Hence, a truster’s willingness (i.e., trusting intention) along with trusting beliefs
both seem to be a key aspect of the definition of trust. While the topics of uncertainty,
vulnerability, risk, etc. are all recognized by the discussed researchers, none of these advances
a more profound insight into the various forms, kinds and levels of risk. Therefore, the current
knowledge is still shallow.

Recent research recognizes that trusting attitudes are another critical component of trust. In
general, the concept of attitude represents a simple mental process of evaluation of the trust
object’s attributes mirrored in a person’s feelings or perceptual judgments about the trusted
party (Kim and Tadisina 2007). According to Eagly and Chaiken (1998), attitude represents a
psychological tendency or status expressed by evaluating a specific entity with some degree of
favor or disfavor. In the more narrow sense, attitude is often regarded as a state of preparedness
or readiness for behaviour (Ajzen & Fishbein, 1980). Jung’s (1971, p. 687) definition of attitude
emphasizes this perspective, as he views the concept as the “readiness of the psyche to act or
react in a certain way”. While being impersonal, eWOM nevertheless represents indirect human
interaction that is likely to be mirrored in consumers’ readiness to favor this kind of market
information. It is very likely that trust manifests itself in various attitudinal dimensions or forms
of trusting attitudes.

With such theoretical distinction of the key concepts of trust in mind, the current thesis adopts
most influential prior conceptualizations as it proposes that trust in eWOM is composed of (1)
trusting beliefs and (2) attitudes towards the characteristics of the information conveyed in
online reviews (these two types of trust together represent something that various authors would
call the concept of eWOM trustworthiness, i.e., the perception of eWOM as a trust object itself),
as well as (3) the consumer’s willingness to act and rely on other consumers’ opinions and
recommendations under the conditions of vulnerability, uncertainty and risk (i.e., trusting
intention). Hence, in accordance with literature, trusting intention is defined as a person’s active
mental status or motivation that may result in actual behaviour. By applying the three
underlying trust concepts to the context of eWOM, a person is said to have high disposition of
trust in online reviews when he/she thinks that eWOM is honest, useful, benevolent etc. – that

105
is, has positive characteristics (i.e., certain beliefs); is confident about and likes eWOM (i.e.,
positive attitude); and wants to use the information and to follow its advice (i.e., intention).
That is, the person has a consistent opinion towards OCR which is also stable over time. The
collectivity of these mental structures facilitates dealing with social information and simplifies
judgments about the usage of eWOM information for purchase-related decisions through
generalizations which are independent of the situation.

Here, one has to note that the conceptualization of eWOM trust at hand incorporates
behavioural intention (i.e., the consumer’s willingness to rely on eWOM information) rather
than actual or consequent behaviour (i.e., relying on eWOM information). In trust literature,
there exist two divergent views of the topic (see Chapter 2 for a more detailed discussion).
Some scholars following the first literature stream consider actual behaviour as an aspect of
trust which has to be integrated in any trust definition. For example, in their work, the
recognized trust researchers Lewis and Weigert (1985a, p. 971) state that “to trust is to act as if
the uncertain future actions of others were indeed certain in circumstances wherein the violation
of these expectations results in negative consequences for those involved”. Trust as an actual
behaviour is also mentioned in the definitions of trust proposed by other scholars (e.g., Deutsch,
1962). In contrast, the second stream of literature which is also followed by this research
refrains from including actual behaviours from the definition of trust. This view surfaces, for
instance, in the work of Rosseau et al. (1998, p. 395), who make the statement that “trust is not
a behaviour (e.g., cooperation), or a choice (e.g., taking a risk), but an underlying psychological
condition that can cause or result from such actions”. This approach is well supported in the
appropriate trust literature (McLain & Hackman, 1999). Among others, Mayer et al. (1995) also
separate trust from trusting behaviours as they treat trust (i.e., the willingness or behavioural
intention to take a risk) as an antecedent of actual behaviour (i.e., taking of risks). This thesis
advances the thinking that the latter approach makes more sense, because a central argument of
this school of thought is that trusting behaviours (e.g., cooperation or reliance) may be related
to other factors besides trust. That is, for instance, some behaviours may result from situational
factors (e.g., power or control) and not from true trust in others based upon positive expectations
or beliefs in the trustworthiness of the interaction partner (McKnight et al., 1998; Moorman et
al., 1998; Pearce, 1974). A basic example illustrates this argument in the context of customers’
eWOM usage: if a consumer is faced with a lack of an alternative source of purchase
information (e.g., Consumer Reports, commercials), he/she might be forced to use and accept
information that is provided in the form of eWOM, even though the consumer does not have
confidence in the trustworthiness of the eWOM-conveyed information. Separating trusting
intentions from actual behaviour and excluding behaviour from a trust definition also helps to
eliminate an inherent measurement problem. That is, if risk-taking behaviour is not solely based
on trust, trusting behaviour is simply not a dimension of trust (Kim & Tadisina, 2007).
Therefore, this study proposes that trusting intention is a fundamental part of eWOM trust and
a more appropriate indicator of the construct than actual trusting behaviour. Additionally, the
106
research at hand assumes that actual behaviour may be influenced by trust (i.e., trusting
intention) - besides other factors. The separation of attitude and behaviour is also well
established in literature (Hogg & Vaughan, 2011).

Dimensions of eWOM Trust

As discussed earlier, researchers have regularly conceptualized trust as a complex, multi-

dimensional construct. This can be attributed principally to an inclusion of trusters’ various
critical beliefs about specific characteristics of the trusted party in the researchers’ trust
definitions. For instance, McKnight et al. (1998) defined trust in terms of beliefs a person holds
about the trustee’s benevolence, competence, honesty, and predictability in a given situation.
Fundamentally, their definition of trust is similar to that of trustworthiness by Mayer et al.
(1995), who defined trustworthiness as attributes or characteristics of the trustee. According to
them, trustworthiness is primarily established by distinct beliefs about the ability, integrity, and
benevolence of the interaction partner. Since then, numerous authors have agreed with the
three-dimensional conceptualization of trust (Blau, 1964; Butler, 1991; Chen & Dhillon, 2003);
also in the context of eWOM trust (Yu & Tang, 2010). However, there is evidence that these
dimensions are often intertwined (Crosby et al., 1990; Ganesan, 1994; Gefen, 1997). Empirical
identification therefore can be threatened. An extensive review of trust literature had identified
that even though the terms of the dimensions vary between authors and some dimensions can
be specified in more detail, their definitions and operationalizations suggest that this
fundamental set of trustworthiness beliefs represent key dimensions of trust. However, due to
the dominant cognitive-based approach towards trust, a number of trust researchers have
neglected additional important aspects of trust, such as the role of trusting attitudes and
intentions. For instance, the ad-hoc eWOM trust measure of Casalo et al. (2011) understands
the construct as being multi-dimensional; however, this conceptualization is only limited to the
cognitive beliefs of (i) honesty; (ii) benevolence; and (iii) competence. On the other hand,
Ridings et al. (2002) conclude that ability, on the one hand, and a combined benevolence and
integrity dimension, on the other, seem to represent best the concept of trust.

In accordance with the literature’s predominant multi-dimensional idea of trust and the aim to
include a preferably broad basis of potentially relevant trust concepts, this study has identified
five dimensions of trust that have most frequently appeared in related trust research: (1) ability
(usefulness); (2) integrity/honesty; (3) benevolence; (4) likeability; and (5) willingness to rely
on. These key concepts have been applied in the identification stage of the research at hand in
order to extract an initial pool of items based on current academic knowledge and their
definitions helped to gain additional insights in expert and consumer interviews. MacKenzie,
Podsakoff and Podsakoff (MacKenzie, Podsakoff & Podsakoff, 2011, p. 300) note, “if a
construct is multi-dimensional then it is important to define each of the sub-dimensions with
the same care that was used in the case of the focal construct itself”. Hence, in the remainder of

107
this section, the diverse dimensions are discussed in reference to the study’s conceptualization
of eWOM trust and formal definitions of each sub-dimension are provided.

Including the concept of ability in trust definitions is not new, as it represents a critical
component in a great number of trust conceptualizations (Bhattacherjee, 2002; Chen & Dhillon,
2003; Hovland et al., 1953; Lee & Turban, 2001; Mayer et al., 1995; Mishra, 1996; Sitkin &
Roth, 1993). For instance, by defining trust as “socially learned and socially confirmed
expectations”, Barber (1983, p. 164-165) included “competence trust” (i.e., the belief that the
interaction partner exhibits a technically competent role performance) as well as goodwill in
his trust conceptualization. Likewise, Nooteboom (1996) notes that trust may be affected by an
interaction partner’s ability to perform according to earlier agreements. In their empirical study,
Swan et al. (1988) found that competence, a similar concept, principally contributes with the
second dimension for trustee’s honesty and responsibility to perceptions of trustworthiness. In
general, ability is considered to be strongly linked to competence and expertise within a specific
domain, as well as to all characteristics that are required to successfully accomplish a given task
or job (e.g., Barber, 1983; Lee & Turban, 2001). For Mayer et al. (1995, p. 718), ability is “the
group of skills, competencies, and characteristics that enable a party to have influence within
some specific domain”. Mayer et al.’s (1995) view is subsequently picked up by various authors
– also in the context of eWOM (e.g., Ridings et al. 2002). Similarly, Gefen (2002) considers
ability as the truster’s beliefs about the skills and competencies of the trusted party, and for Tan
and Sutherland (2004) it’s the truster’s thought that the trustee is able to accomplish a task.
These conceptualizations are also applicable to the context of eWOM. Consumers who are
accessing eWOM have a mutual interest in making better purchasing decisions and hence are
typically concerned about the ability of online reviews and recommendations with respect to
this mutual concern. From the consumers’ perspective, the primary functions or tasks of eWOM
are to furnish consumers with meaningful information about the true/real characteristics of a
market offering (e.g., product) and consumption recommendations that enable them to make
better purchase-related decisions. Trusting consumers think that online reviews furnish them
with important and relevant product insights gained from own consumption experiences which
would not been accessible otherwise. Readers of OCR are interested in viewing the market
offering from the consumer perspective and want to find out whether purchasing the product,
service, etc. enables them to reach the desired consumption goal. A key question therefore is:
are reviews and reviewers both capable of providing these insights? Therefore, in the context
of eWOM, this study defines (eWOM) ability as the consumer’s belief that eWOM is a useful,
knowledgeable and adequate source of purchase-relevant information and/or recommendations.

Research has identified the concept of integrity/honesty as being another critical dimension of
relational trust (Gabarro, 1978; Mayer et al., 1995). Here, most scholars tend to provide quite
broad and multifaceted definitions of the construct; however, the very nature of the concept is
evident. Integrity is commonly conceptualized as the trusted party’s moral character, honesty,

108
truthfulness, and sincerity (Butler & Cantrell, 1984; Larzerele & Huston, 1980; McKnight &
Chervany, 2002). For example, according to Lee and Turban (2001), integrity captures the
truster’s beliefs of the trustee’s honesty, dependability, reliability, credibility, and also
adherence to an accepted set of principles. Chen and Dhillon (2003) describe integrity as acting
in a consistent, reliable, and honest manner. Based on earlier insights of Mayer et al. (1995),
McKnight et al. (1998) also found integrity to be a critical component of trust and furthermore
conceptualize integrity as two distinct constructs, namely honesty and predictability
(consistency). For Büttner and Göritz (2008), integrity implies that the trustee follows a set of
desirable principles. Similarly, Gefen (2002b) defines this trust dimension as the truster’s belief
that the trusted party adheres to accepted rules of conduct, norms or standards, such as honesty
and keeping promises. Consistently with prior conceptualizations (e.g., Soh et al., 2009), the
thesis at hand defines (eWOM) integrity/honesty as a consumer’s belief that the information
conveyed in eWOM is believable, truthful, verified and adheres to moral standards. Integrity
applies to eWOM communication because it is the believability or honesty that is assumed to
be a key expression of trust in online reviews and recommendations that allows consumers to
rely on this kind of information. Without reliability, their information value would be zero.

Benevolence captures another important element of trust identified in previous studies (e.g.,
Larzerele & Huston, 1980; Solomon, 1960; Strickland, 1958). In literature, it seems to be agreed
that benevolence implies a perception of trustee’s responsiveness, goodwill, positive intent
and/or good motives. Further, the concept relates to the attachment the trustee has towards the
truster. For instance, according to Mayer et al. (1995), benevolence is the truster’s perception
of the trustee’s positive orientation toward his/her relationship partner. A similar definition is
advanced by Doney and Cannon (1997). In their work on trust, McKnight et al. (2002a)
conceptualize benevolence as the perception of the trusting party of whether the trustee acts in
the interaction partner’s best interest and/or out of genuine concern. The positive motive aspect
is emphasized by Lee and Turban (2001), who define benevolence as “the extent to which the
trusting party believes that the trusted party wants to do good things […]” – a view that is also
advanced by subsequent research (e.g., Cheung & Lee, 2006). According to Büttner et al.
(2008), benevolence denotes that the trustee is interested in the truster’s well-being. By
referring to earlier conceptualizations, this thesis defines (eWOM) benevolence as a consumer’s
stable belief that eWOM is guided or motivated by the reviewer’s favourable and positive
intentions toward the consumer’s welfare and not by the reviewer’s self-interest only. A
consumer views eWOM-conveyed information as benevolent in the case that he/she has the
perception that the source has favourable motives towards other consumers. Benevolence is an
essential element of eWOM trust, since without positive reciprocation and reviewer’s altruism,
the truster would view reviews and recommendations not to be genuine and reviewers to be
careless or unsupportive. Benevolent reviews and recommendations are not believed to mirror
opportunistic or manipulative intentions. Further, trusting consumers do not regard eWOM
creators as striving to accomplish a selfish goal but believe that they are driven by their
109
generosity towards others. Consequently, if consumers think that eWOM intends to benefit
them, benevolence is generally present.

A specific dimension of trusting attitude, namely likeability, has been recognized only
infrequently by researchers in their notion and definitions of trust. However, since some
scholars have found emotional attitude trust as one of the dimensions of trust (Johnson-George
& Swap, 1982; Lewis & Weigert, 1985; Rempel et al., 1985; Swan et al., 1985) that potentially
exerts meaningful influence on the level of perceived trust, this study regards likeability (as a
key indicator of affect) to be another key component of trust. McAllister (1995) is one of the
leading authors who developed a comprehensive model of affect that includes both trusting
beliefs as well as trusting attitudes. Here, the author emphasized that the emotional content of
trust is based on “emotional ties” between the interaction partners. A similar notion (i.e.,
emotional bonds) is used by Lewis and Weigert (1985). Subsequent research has shown that
the strong emotion of liking represents another aspect of affect-based trust. For example,
Johnson-George and Swap’s (1982) research finds that one of the identified trust factors is
composed of four liking items which parallel three additional trust items. Similarly, Swan et al.
(1988) conclude that likeability is truly a component of trust as they found that to trust means
developing feelings of liking the trust object. Often, likeability has been measured in terms of
perceived attractiveness, feeling of liking, enjoyability, and positive affection (Ganesan, 1994;
Haley, 1996; Ohanian, 1990; Young & Albaum, 2003). Hence, this study defines the dimension
of (eWOM) likeability as the consumer’s positive emotions toward eWOM. OCR
communication regularly shares the same content as person-to-person interaction: it contains
personal feelings, opinions, normative recommendations and also emotions of the reviewer.
Therefore, it is likely that this kind of information also results in affective responses of the
consumer which finally mirror the degree of perceived trust. This is especially true in the long
run.

The consumer’s willingness to rely on (or trusting intention) the trustee represents the
behavioural component of trust. As referred to earlier, to trust means that the truster has a
willingness to be vulnerable and/or take a risk by relying on the trustee’s words or behaviours
(Kaplan & Nieschwietz, 2003; McKnight & Chervany, 2001; Stewart, 2003). In the context of
eWOM, there are multiple forms of uncertainties, vulnerabilities, and risks. For instance,
according to Racherla, Mandviwalla and Connolly (2012), consumers have to deal with two
different uncertainties. First, for consumers it is often hard to infer the true characteristics of
the reviewed product. If the consumer is not able to identify adequate market information,
diverse consumption risks may arise because of undesirable outcomes of relying on eWOM for
making purchasing decisions. For instance, consumers have to bear the risk that when following
the advice of eWOM recommendations, they may suffer from a financial loss due to
unsatisfactory products or services and/or ordering at an unreliable vendor. The functional risk
is more incalculable in the context of hedonic, experience and credence goods, where it is

110
difficult to quantify the product in terms of features and functionalities. On the other hand, a
social-emotional loss may arise in cases where the consumer’s peer group has a negative
judgment about the selection or use of the wrong brand. Second, readers of eWOM have to deal
with the uncertainty regarding the integrity and motives of fellow shoppers that provide the
reviews (Racherla et al., 2012). For instance, Chevalier and Mayzlin (2006) note in their
research that marketers have numerous incentives to disguise promotional reviews as legitimate
consumer recommendations in order to directly influence consumers’ evaluations of their
products. This uncertainty and the inherent risk of misleading market communication is likely
emphasized by various researchers (e.g., Mayzlin, 2006; Pan & Chiou, 2011; Racherla et al.,
2012). Hence, (eWOM) trusting intention can be defined as the consumer’s willingness to rely
on eWOM and accepting potential loss by applying eWOM-conveyed information.

Elements of eWOM Trust

This thesis’ investigation into the nature of trust has spotted that a considerable broad string of
literature exists that emphasizes that both affective processes or emotional bonds towards the
object of trust and behavioural intentions, along with cognitive thoughts or rational
expectations, have to be present in order to provide insight into the level of trust. In other words,
trust is a matter of thoughts, feelings, and behavioural intentions (e.g., Clark & Payne, 1997;
Cummings & Bromiley, 1996; Granovetter, 1985; Riegelsberger et al., 2003). Hence, based on
these conceptualizations, this thesis proposes a multifaceted concept of trust that incorporates
cognitive, emotional, and behavioural elements in its definition. This is also in accordance with
the three-component view of attitude (Himmelfarb & Eagly, 1974; Krech et al., 1962). As
shown in Chapter 2, Lewis and Weigert (1985) have introduced a framework of trust that
provides the basis for categorization of the trust dimensions along these three underlying
elements of trust. According to the scholars, trust may arise based on “a cognitive process
which discriminates among persons and institutions that are trustworthy, distrusted, and
unknown” (Lewis & Weigert 1985, p. 970). They further note that a truster can have “cognitive
trust”, which is “good rational reasons why the object of trust merits trust” (Lewis & Weigert
1985, p. 972). Typically, the cognitive dimension deals with the truster’s rational evaluation of
specific characteristics of the trustee, such as his/her/its integrity and honesty. Hence, in this
research the first two trusting beliefs (ability/usefulness and integrity/honesty) represent in part
the predominant cognitive base of eWOM trust. In literature, ability is commonly associated
with a rational evaluation of the trustee’s competencies and skills. Integrity and reliability, on
the other hand, are cognitive beliefs that reflect ethical traits and deal with the belief that the
partner will keep to a given promise or will not lie.

Emotional trust in eWOM is based on the belief of benevolence, as well as feelings of affect
(likeability). Benevolence is related to a party’s altruistic motives and the trustee’s goodwill
towards the truster (Mayer et al., 1995). Even though some researchers state that the concept

111
represents a kind of cognitive judgment, a great number of scholars stress that the construct is
a predominant emotional or affect-based characteristic of the object of trust (Dunn, 2000;
McAllister, 1995; Riegelsberger et al., 2003). In this thesis, benevolence is recognized as a
basically cognitive belief which is triggered by emotions. Trusting consumers identify some
kind of affection between them and reviewers, who are seen as wanting to do well to them. As
discussed earlier, Lewis and Weigert (1985), amongst others, advance the view that likeability
is another critical element of trust. Unlike benevolence, likeability represents “pure emotion or
affect” without any cognitive involvement. The potential importance of these emotional
elements can be deduced from literature, as it is commonly agreed that the type of relationship
determines the basis of trust (Rosseau et al. 1998). In general, consumers use eWOM in order
to make better purchasing decisions – that is, identify the best product. This purely economic
goal which parallels the impersonal form of eWOM communication makes cognitive elements
of trust more likely to be important. However, the social character of this interaction, the need
to exchange with similar minded persons, the inherent dependence of the consumer on the
reviewer community, and the potential perception of the consumer that information seeking is
itself a source of pleasure (Mathwick & Rigdon, 2004; Mudambi & Schuff, 2010) make
emotional trust elements likely to be also an integral part of the eWOM trusting stance.

eWOM Trust and Similar Concepts

In order to have a profound understanding of what eWOM trust truly is, it is reasonable to also
clarify what eWOM trust is not. There are many related concepts that have been often used
synonymously with the focal construct. This regularly results in an obfuscation of the original
concept and conceptual confusion. Here, the four concepts of eWOM distrust, scepticism,
credibility and attitude in general are discussed in greater detail. At its heart, this thesis
postulates that, even though all constructs have some facets in common, the conceptualization
of eWOM trust includes other dimensions that usually do not reflect the nature of these related
concepts. The identified constructs are hypothesized as being separable but nevertheless related.
Scholars regularly use further terms like cooperation, faith, and competence (e.g., eWOM
helpfulness), whose meanings have to be clarified. The following section differentiates trust
from these concepts by reviewing earlier traditional as well as domain-specific contributions.

eWOM Distrust

Some early trust researchers view distrust as merely the opposite of high trust and argue that
trust and distrust (also called mistrust) are two extreme values along the same dimension (e.g.,
Rotter, 1967). For example, Luhmann (1979), one of the most thorough theorists in the field of
trust research, states that trust and distrust are simply one and the same construct, as they
represent distinct functional equivalents that impact human thinking and behaviour separately.
The scholar further suggests that trust cannot exist without distrust and that a change in either
without a corresponding change in the other can be harmful (Benamati, Serva & Fuller, 2006).

112
In contrast, a more recent literature stream views distrust as qualitatively independent from trust
(e.g., Lewicki, McAllister & Bies, 1998; Sitkin & Roth, 1993). In essence, these studies
commonly suggest that both constructs coexist simultaneously and have differential effects on
behaviour (Kramer, 1999; Lewicki et al., 1998). Specifically, researchers of this school of
thought also see distrust having a more negative impact on an individual’s risk-taking decisions
than could be attributed to a simple impairment of trust (Ou, Abratt & Dion, 2006). One of the
broadest definitions of trust is the notion of the truster’s willingness to become vulnerable to
the trustee, having taken into consideration certain attributes of the trustee (Mishra, 1996;
Rosseau et al., 1998; Zand, 1972). Such a definition can be applied to three types of trustees, as
the truster could show a willingness to become vulnerable to (1) specific others, e.g. another
person (i.e., interpersonal trust); (2) a context, e.g. the Internet, OCR (i.e., institutional trust);
and/or (3) general others (i.e., dispositional trust) (McKnight et al., 2004). By assuming that
distrust is the opposite of trust, as Lewicki et al. (1998) argue, distrust can be defined as the
unwillingness to become vulnerable to a trustee – another person, institution, or people
generally – having considered characteristics of that trustee (McKnight et al., 2004).

Lewicki et al. (1998) extensively argue for a conceptual separation of the two constructs. The
scholars categorically state that “high distrust is not the same thing as low trust” (p. 444).
Lewicki et al. offer three reasons for this claim, as they advance the insight that both constructs
have different consequences and causes as well as factor separately. In their explanation, they
conceptually illustrate various situations where combinations of high-low trust and distrust may
co-exist. Later, these arguments found support by empirical evidence (McKnight et al., 2004).
McKnight et al. (2001, p. 40) agree on the separation of the constructs by stating “distrust is the
distinct opposite of trust”. They argue that trust researchers often define trust in terms of specific
feelings of security (e.g., Lewis & Weigert, 1985a; Rempel et al., 1998) or comfort (Eayrs,
1993). Trust focuses on more favourable emotional reactions (like hope, confidence, and
assurance) towards another person (or thing) and is, as a consequence, positive-valent. In
contrast, distrust is based on more negative emotions, such as suspicion, wariness and fear
(Deutsch, 1958; Lewicki et al., 1998) and hence is negative-valent (Benamati et al., 2006).
Accordingly, Lewicki et al. (1998) describe that trust is best specified by hope, faith, assurance,
confidence, and initiative. In contrast, distrust could be described with terms like fear, doubt,
scepticism, cynism, wariness and watchfulness (Ou et al., 2006). In parallel, in one of the most
influential trust studies McKnight et al. (2002a, p. 341) argue that “negatively worded trust
items tend to factor separately into distrust, which is conceptually separate from trust […] we
used all positively worded [trust] items”. McKnight et al. (2001) conclude by referring to
Luhmann (1979) that if a person distrusts, he/she experiences a more frantic or emotionally
arousing state which leads to this psychological reaction. Hence, according to the authors, the
emotional intensity of distrust clearly distinguishes it from trust. McKnight and Chervany
(2001) assume that trust as well as distrust both have cognitive and affective elements. This
contrasts McAllister’s (1995) approach, which proposes that trust is either cognitive or
113
affective. Nevertheless, behind both constructs lie similar functional mechanisms, as trust and
distrust both reduce social complexity: While trust fulfils this function by compelling a person
to undertake behaviours that expose the truster to risk, distrust reduces complexity by
compelling an individual to take protective action to minimize risk (Luhmann, 1979; Benamati
et al., 2006).

In accordance with the literature discussed above (e.g., McKnight et al., 2004) this thesis
defines eWOM distrust as a consumer’s unwillingness to rely on information and
recommendations conveyed in online customer reviews having considered the motives of peer
reviewers and the characteristics of reviews in general. It is assumed that this concept is
separated from and the opposite of eWOM trust, as it is a facilitator of stronger negative
attitudes and behaviours that cannot be triggered by a person’s lack of trust in online customer
reviews alone. Therefore, it is theorized that persons with high levels of eWOM distrust are, for
instance, characterized by a recognizable level of fear of being exploited by reviewers and they
also perceive unique negative emotions towards online reviews in general.

eWOM Scepticism

A concept related to eWOM distrust is eWOM scepticism. Earlier research has shown that
consumers are generally sceptical about personal experiences made by others concerning
product performance and quality (Ford, Smith & Swasy, 1988; Nelson, 1974). Similarly, it can
be assumed that when consumers read online reviews they may ask themselves whether or not
the reviewer who recommends or refrains from doing so receives an incentive from the
producing company or another interested party. As a consequence, consumers may show a
certain degree of scepticism, as they don’t know whether they can believe this kind of
information or not. Especially the contributions of Obermiller and Spangenberg provide
theoretical as well as empirical insights to the evident fact that consumers are often sceptical
towards advertising in general (Obermiller & Spangenberg, 1998, 2000). This is also true for
advertisement that is disguised as peer information.

Obermiller and Spangenberg (1998) describe consumer scepticism as a market belief which is
stable over time and reflects the implicit view of how the marketplace works. Specifically, it is
the generalizable tendency towards the disbelief of claims made in ads (or other forms of
marketing messages). This definition makes the connection to the trust concept evident as ad
(eWOM) scepticism can be understood as a consumer’s lack of belief in advertising (eWOM)
(Boush, Friestad & Rose, 1994; Mangleburg & Bristol, 1998). The inclination to disbelieve
marketing messages is the result of a socialization process and earlier purchasing experiences,
for instance, with the Internet which is likely to produce some scepticism towards the available
information (Sher & Lee, 2009). While Obermiller and Spangenberg regard consumer
scepticism as uni-dimensional, other researchers question this approach. For instance, Boush et
al. (1994) find that the construct has two dimensions, namely mistrust of advertiser motives and
114
disbelief in ad claims. Ford et al. (1990) show in their empirical research that an individual may
have different levels of scepticism towards claims made in advertising and the motives of
advertisers. Their work leads to the assumption that consumer scepticism is both a function of
the product type as well as the individual’s personality. Here, the direct connection to the
concept’s opponent, eWOM trust, becomes more obvious. However, eWOM scepticism solely
deals with the negative expressions concerning claim believability (i.e., the construct is
regarded as the opposite of perceived eWOM integrity/honesty), while trust is a richer concept
which introduces additional elements (e.g., ability). While the issue of the dimensionality of
consumer scepticism is not completely resolved yet, Obermiller and Spangenberg’s (1998)
conceptualization remains the point of reference for eWOM researchers (e.g., Sher & Lee,
2009), which also guides this thesis’ standpoint that eWOM trust and scepticism are likely to
be negatively related. This is also mirrored by the items usually used to measure consumer
scepticism (Obermiller & Spangenberg, 1998). According to this literature, eWOM scepticism
is best defined as a single dimension. This separates this concept from eWOM distrust, which
is multi-dimensional and includes further (negative) evaluations of the trust object.

Credibility

Trust and credibility have been sometimes used interchangeably. However, these two constructs
should not be confused or considered synonymous. Tseng and Fogg (1999, p. 41) state that trust
is different from credibility, because “trust indicates a positive belief about the perceived
reliability of, dependability of, and confidence in a person, object, or process”. The scholars
continue by emphasizing that trust typically refers to dependability, while credibility is roughly
a synonym for believability. This view contains the implicit assumption, based on the earlier
insights of this thesis, that credibility is likely to cover only distinct elements of trust concerning
its meaning. Additionally, when defining trust, scholars characterize the concept as a set of
beliefs, predispositions, and behaviours which are related with the acceptance of risk and
vulnerability. In contrast, credibility refers to the quality of the information or a source which
may (or may not) result in trusting intentions and/or behaviours (Rieh & Danielson, 2007).
Credibility can be regarded as an antecedent of trust. Hence, credibility provides a reason to
trust but is not trust itself (Corritore et al., 2003). Given this conceptual distinction, the
credibility concept has received wide attention in marketing communications research as well
as in studies on eWOM (e.g., Bae & Lee, 2011; Brown et al., 2007; Huang & Chen, 2006).
People can develop credibility perceptions towards different subjects and objects. The reminder
of this section emphasizes the conceptual nature of two major credibility constructs in the
context of eWOM – namely (i) reviewer credibility and (ii) review credibility – and discusses
their conceptualizations.

115
Reviewer Credibility

Credibility has been explored from three distinct perspectives: (1) source credibility; (2)
message (content) credibility; and (3) medium credibility. Among these, source credibility has
been researched most extensively (Yang, 2007). The researchers’ interest is due to the general
assumption that a highly credible source is commonly found to induce more persuasion than a
low-credibility one (Pornpitakpan, 2004b). Source credibility is a term typically used to suggest
that a communicator’s positive characteristics can increase the value of information conveyed
in a message and this, as a consequence, affects the likelihood that the receiver subsequently
accepts the message (Anderson, 1971; Dou, Walden, Lee & Lee, 2012; Hovland et al., 1953).
Source credibility refers to a message recipient’s perception of the credibility of a message
source, reflecting however nothing about the message itself (Chaiken, 1980). A landmark
among these attempts is surely the contribution of Hovland et al. (1953), who postulate that
credibility is a critical attribute associated with the communicator which exerts influence on a
recipient’s message acceptance. The authors identify expertise and trustworthiness as the two
outstanding determinants of source credibility. Subsequent research increased the confidence
in this conceptualization as it regularly identifies the same two dimensions. For example,
Ohanian (1990) as well as Pornpitakpan (2004) both conclude that these dimensions are seen
as the most important source characteristics. Similarly, Dholakia and Sternthal (1977) find that
expertise and trustworthiness determine source credibility. Expertise is defined as the extent the
communicator is regarded as representing a source of valid assertions and has the ability to
provide accurate information coming from the knowledge, experience, training or skills that the
source possesses (Hovland et al. 1953). On the other hand, trustworthiness (here best
understood as a synonym for the source’s integrity/honesty) refers to the receiver’s confidence
or belief that the source will provide information that is both unbiased and honest (Ohanian,
1991). Hovland et al. (1953) provide evidence that a source’s trustworthiness significantly
affects the acceptance of a message and enhances changes in opinions and attitudes.
Accordingly, McGinnies and Ward (1980) conclude that communicators who are in the
possession of these characteristics cause the greatest change in others’ opinions. In a similar
vein, Petty and Cacioppo (1986) define source credibility as the extent to which the information
source is assumed to be believable, competent, and trustworthy by the information recipients.

In marketing research, source credibility has been studied mainly in the context of endorser and
advertiser credibility. Research on endorser credibility is to a large extent influenced by the
source credibility model advanced by Hovland and his colleagues (1953). Therefore, it comes
as no surprise that this literature string has also identified trustworthiness and expertise as the
most critical elements of a communicator’s credibility. This is also supported by a recent meta-
analytic study by Amos et al. (2008), who have found in their study that these two components
bear more influence on the effectiveness of ads than any other source characteristics, including
celebrity performance and celebrity/product match. This further supports a multi-dimensional

116
understanding of the concept of source credibility. Ohanian (1990) introduces source
attractiveness as an additional dimension, which is defined as the receiver’s affection for the
source due to its physical appearance and behaviour. Endorser credibility is typically measured
with semantic differential items. For instance, expertise is often measured in terms of
“knowledgeable-unknowledgeable”, “qualified-unqualified”, and “experienced-
inexperienced”. In contrast, for measuring a source’s trustworthiness, researchers often use
terms like “trustworthy-untrustworthy”, “honest-dishonest”, and dependable-undependable”
and for attractiveness “beautiful-ugly”, “attractive-unattractive”, and “plain-elegant” (Lafferty
& Goldsmith, 1999; Ohanian, 1990).

Another literature string puts its emphasis on the credibility of the advertiser who produced the
advertised product being a valuable source of information, i.e., advertiser credibility (Lafferty
& Goldsmith, 1999; MacKenzie & Lutz, 1989; Newell & Goldsmith, 2001). Here, in the same
tradition, expertise and trustworthiness have been identified as components of credibility. While
expertise is generally described as the extent to which consumers have the perception that the
company/advertiser has the knowledge or skill to fulfil its stated claims, trustworthiness is
defined in terms of whether the company can be trusted to tell the truth or not (Newell &
Goldsmith, 2001). Van de Berge et al. (1981) propose that an advertiser’s credibility is best
described in terms of trustworthiness, prestige, competence, familiarity, competitiveness along
with the affective elements of attractiveness and likeability. Similarly, Haley (1996) found that
consumers perceive advertisers as credible when they show adequate levels of trustworthiness,
expertise, and attractiveness.

Research on the credibility of individuals who state their opinions and comments online (i.e.,
reviewer credibility) heavily draws from the traditional view of source credibility. It is
commonly understood as a consumer’s attitude towards the message source (Gunther, 1992),
which dictates how much the consumer believes in the message sender and consequently the
message itself. When a source is perceived to be credible, the level of belief in what the source
claims is positively influenced Ling & Liu, 2008; Wu & Wang, 2011). In an analogy, this thesis
views reviewer credibility as an (situational) antecedent of eWOM trust. If a consumer holds a
positive orientation towards the credibility of eWOM endorsers, he/she will be more likely to
trust the information provided by this group of fellow shoppers.

The work of Hovland et al. (1953) seems to be most influential, as reviewer source credibility
is recognized as an important factor impacting persuasion effectiveness. Additionally, most
researchers recognize that the source’s trustworthiness as well as expertise are both factors that
have been most frequently studied. For instance, the works of Brown et al. (2007), Huang et al.
(2006), and Luo et al. (2013) all agree that source credibility of online reviewers depends on
these two factors. By drawing on the work of Sternthal et al. (1978), Luo et al. (2013) equally
define the construct with its dimensions, as they state that source credibility is an information

117
reader’s perception of the expertise and trustworthiness of a source. Consequently, the construct
has been regularly measured by applying the semantic-differential scale derived from Ohanian
(1990, 1991). For example, Dou et al. (2012) use “dependable-undependable”, “honest-
dishonest”, “reliable-unreliable”, “sincere-insincere”, and “trustworthy-untrustworthy” to
measure trustworthiness, and “expert-not expert”, “experienced-inexperienced”,
“knowledgeable-unknowledgeable”, “qualified-unqualified”, and “skilled-unskilled” to
quantify the expertise dimension. Wu et al. (2001) recognize that source attractiveness is also
an important determinant. However, they refrain from measuring this dimension, because
according to their opinion eWOM senders cannot be directly contacted by the consumer and
therefore attractiveness cannot be assessed. Each of the remaining dimensions includes five
semantic-differential items adopted from Ohanian (1990). Similarly, Rabjohn et al. (2008)
measure trustworthiness and expertise with two items each. By applying a Likert scale, Lis
(2013) quantifies reviewer expertise with five items (e.g., “The reviewer is an expert.”, “The
reviewer is qualified.”), and the same amount of items for measuring reviewer trustworthiness
(e.g., “The reviewer is honest.”, “The reviewer is sincere.”). All these measures are taken again
from Ohanian (1991).

Review Credibility

In traditional marketing communications research, it is generally agreed that the credibility of

the specific advertisement is an important attribute of effective communication, since
consumers are not likely to behave in the desired manner in cases where they do not accept or
believe the claims made in this communication vehicle. In this context, credibility of specific
ads is typically defined as “the extent to which the consumer perceives claims made about the
brand in the ad to be truthful and believable” (MacKenzie & Lutz, 1989, p. 51). Various
approaches have been put forward to measure the credibility of single message content in
advertising. For example, without assuming any dimensionality among the items, Beltramini
and Evans (1985) developed their believability scale with ten bi-polar adjective pairs:
“trustworthy”, “convincing”, “believable”, “reasonable”, “credible”, “honest”, “conclusive”,
“unquestionable”, “authentic”, and “likely”. This scale has been used, for instance, by O’Cass
(2002) in order to measure the believability of political advertising. In the context of online
reviews and C2C communications, research similarly has concentrated on the believability
dimension of credibility.

Various authors (e.g., Cheung, Lee & Rabjohn, 2008; Hong & Park, 2012) have all introduced
the credibility concept to the academic discussion of eWOM, as review credibility is often
evaluated as an important ingredient of the information adoption process and a strong predictor
of various behaviours, attitudes, etc. Meanwhile, considerable research has demonstrated the
relationship between review credibility and adoption. The aforementioned literature stream
adopts the interdisciplinary view that credibility is often understood as or equated with

118
believability. Here, the work of Hovland, Janis, and Kelley (1953), which advances the insight
that credibility is an attitudinal construct that typically refers to the believability of some
information and/or its source, seems to be most influential. In adaptation of Hass’ (Hass, 1981)
original work and related works (Flanagin & Metzger, 2000; Fogg et al., 2002; McKnight &
Kacmar, 2007) in the eWOM context, credibility is often defined as the extent to which one
perceives the review/recommendation as unbiased, believable, true, or factual (e.g., Cheung et
al., 2009; Lee & Koo, 2012; Park et al. 2011). In reference to the earlier work of Tseng and
Fogg (1999) and parallel contributions in eWOM research (e.g., Cheung et al., 2009), Park et
al. (2011) as well as Lee and Koo (2012) both define review credibility as the “extent to which
one perceives sources of information provided as unbiased, believable, true, or factual” (p.
1076). Mackiewitz (2007) adopts Ohanian’s (1990, p. 41) more general view of the concept by
defining it as the “characteristics that affect the receiver’s acceptance of a message”.

It shall be noted that in this research, eWOM or review credibility refers to the information
conveyed in specific online reviews and recommendations itself and not to the messages’ source
(i.e., reviewer), which is sometimes confused in literature. However, literature outlines the
general function of credibility as an important determinant of the level of message acceptance
and learning. In cases where the review is perceived as having bias or representing messages
that are motivated by illegitimate purposes other than consumer experiences, opinions, and
recommendations, the credibility of the message is degraded (Hass, 1981; Lee & Koo, 2012).
Additionally, if the reader perceived the review’s credibility as not ensured, he/she will
generally resist the persuasive intent of the review. In contrast, if consumers think that incoming
information is credible, they will be more confident to adopt the eWOM comments and apply
this information in order to make purchasing decisions (Tseng & Fogg, 1999). In essence, if a
consumer perceived a eWOM message as credible, he/she will learn from and use this review.

Park et al. (2011) measure review credibility with four items on a five-point agreement scale:
“I believe the online review which has been read a lot.”, “I believe the online review which is
believed by others.”, “I believe online review is important and credible information.”, “I
believe the online review is written with responsibility.”. Instead, Lee and Koo (2012) measured
credibility with two items taken from Cheung et al. (2009): “I think the review is accurate.” and
“I think the review is credible.”. Lis (2013) also uses Cheung et al.’s scale but also includes the
item “I think the review is factual”, and measures a single credibility construct. In contrast, Xie
et al. (2011) agree with Kelman and Hovland (1953) that trustworthiness and expertise are the
two key components of review credibility. However, the authors measured each dimension with
a single question (i.e., “To what extent do you consider the travelers’ reviews you just read as
trustworthy?” and “To what extent do you consider the travelers’ reviews reflect the reality of
hotel [x]?”). Also Bae and Lee (2011) understand eWOM credibility (in specific reviews) as a
multi-dimensional construct consisting of believability, fairness, accuracy and depth of
information. However, these approaches seem to stem from a confusion of review credibility

119
and reviewer credibility, as they do not represent the norm for measuring the credibility of a
piece of information. The majority of this research stream interprets review credibility of a
specific review as its believability (Bambauer-Sachse & Mangold, 2013; Luo et al., 2013), and
therefore as a uni-dimensional construct. For example, Qiu et al. (2012), who define general
review credibility by drawing on the work of Tseng and Fogg (1999) as the extent to which a
piece of information is perceived as true and valid, measured review credibility with three items
targeting solely the believability dimension.

In parallel to the literature discussed above, various scholars have investigated general
credibility perceptions in the context of eWOM. This literature string heavily draws from offline
research. For instance, in the context of marketer-endorsed communication, MacKenzie and
Lutz (1989, p. 51) define advertising credibility as the “consumer’s perceptions of the
trustfulness and believability of advertising in general, not simply the particular ad in question”.
The scholars applied three bi-polar adjective pairs in order to operationalize the construct:
“believable-unbelievable”, “biased-unbiased” and “convincing-unconvincing”. MacKenzie
and Lutz’s definition as well as the measurement of the construct both imply that advertising
credibility describes a consumer’s general beliefs in the integrity of claims in advertising. In
the majority of cases, researchers applied a single item measure to quantify the construct (Calfee
& Ringold, 1994) or scholars have adapted a sub-scale to measure the consumer’s attitude
toward offline (or online) advertising in general (Alwitt & Prabhaker, 1994; Ducoffe, 1996;
Yang, 2003). Among these scales, one can find items such as: “ads usually present a true
picture”, “most advertising is believable”, and “ads are reliable sources of information”. In a
similar vein, Thorson et al. (2010) focused on consumer-generated communication and
measured review credibility in general (RCred) with a single credibility dimension. By using
items taken from Meyer (1988), the authors used a scale comprised of six semantic-
differentials: “fair-unfair”, “biased-unbiased”, “accurate-inaccurate”, “doesn’t tell the whole
story-tells the whole story”, “cannot be trusted-can be trusted”, and “balanced-unbalanced”.
Also Hong and Park (2012) used five items to describe the believability of online reviews: i.e.,
to me, overall reviews are … “biased-unbiased”, “trustworthy-untrustworthy”, “accurate-
inaccurate”, “believable-unbelievable”, and “complete-incomplete”. By drawing mainly from
the work of Lee and Koo (2012), (eWOM) review credibility can be defined as the extent to
which a consumer holds the general orientation to perceive the information given in online
customer reviews and recommendations as unbiased, believable, true, and factual.

At the heart, the approaches referring to the subject of online reviews itself focus on eWOM’s
integrity/honesty (and hence a single dimension), while approaches targeting the peers behind
these reviews emphasize both eWOM integrity and competence. As such, reviewer credibility
appears multi-dimensional. Nevertheless, popular approaches typically neglect the
measurement of consumers’ evaluations of the benevolent intentions of the reviewers.
Additionally, even though original source credibility literature sometimes proposes the

120
inclusion of affect as an additional “attractiveness” dimension (Freeman & Spyridakis, 2004),
no research approach has included this factor in the context of review credibility. Additionally,
another factor, namely the consumer’s willingness to rely on eWOM, has never been used in
order to conceptualize credibility in the offline or online context.

In line with these findings, this thesis claims that review (eWOM) review credibility (in general)
and its measurement represent only a distinct part of eWOM trust. In essence, review credibility
is understood as a positive characteristic of the trusted object (more specifically, a positive
belief in or a judgment of the believability of eWOM messages), whereas trust is a
psychological reaction toward the object which is partly based on credibility evaluations.
Accordingly, earlier research has identified credibility as being an (situational) antecedent of
trust (Colquitt et al., 2007; Lowry, Vance, Moody, Beckman & Read, 2008; Mayer et al., 1995)
and it is likely that the construct represents one of the necessary requirements for trust to evolve
in the context of online reviews. However, it would be inappropriate to make the claim that a
person who shows a high level of eWOM credibility automatically displays the same levels of
trust in eWOM. For instance, a consumer can perceive eWOM opinions as highly truthful and
accurate; however, due to missing willingness to rely on this information, the consumer does
not trust them.

Attitude towards Reviews in General

Another construct which is closely related to trust in eWOM is attitude towards reviews in
general (review attitude or RAtt). In reference to the appropriate advertising literature (Lutz,
1985), this thesis defines the construct as the learned predisposition to respond in a consistently
favourable or unfavourable manner to online customer reviews in general. By adopting the
argumentation of early contributions to the field (e.g., Obermiller & Spangenberg, 1998) but
also more recent academic literature (e.g., Soh et al. 2009), this thesis agrees that eWOM trust
may be a critical antecedent (or correlate) for some aspects of consumers’ attitudes toward
online reviews and recommendations. For instance, a lack of eWOM trust may be a basis for
generally disliking eWOM communication. Nevertheless a consumer’s positive or negative
attitude typically depends on many other dimensions which are not typically related to the trust
concept itself. For example, customers enjoy funny customer reviews, but at the same time
readers often doubt the truthfulness of extremely humorous reviews. At the same time, the
majority of consumers tend to evaluate truthfulness as something that is positive, and falseness
negatively, which leads to the assumption that one should expect the eWOM trust scale (eWT-
S) to overlap in some aspects with a consumer’s overall attitude toward reviews in general.

A review of several scales provides evidence that trust and attitude toward eWOM are related
but conceptually separated constructs (Obermiller & Spangenberg, 1998). In fact it appears that
in the conceptualizations of review attitude, few items are included that appear to reflect general
trust, but on the other hand, include many other items that seemingly measure other dimensions
121
of attitude that capture specific aspects unrelated to trust in eWOM (e.g., Park et al., 2007). For
instance, a variety of beliefs concerning positive as well as negative outcomes of marketing
communications (e.g., informational value, entertainment, interruptive nature, ethics,
deception, and falsehood) are typically included to measure a consumer’s generalized
attitudinal tendency (Bauer & Greyser, 1968; Mehta, 2000; Mehta & Purvis, 1995). The
majority of this set of beliefs is not an integral part of eWOM information trust per se, but
overlaps do exist. This is given as various researchers choose to include the factors of
believability (integrity) and information usefulness (ability) in their general attitude measures
(e.g., Olney, Holbrook & Batra, 1991; Sandage & Leckenby, 1980). On the other hand, as
mentioned earlier, eWOM trust is conceptualized as including cognitive, emotional as well as
behavioural elements of trust. In contrast, measures of RAtt typically focus on the consumers’
cognitive and/or more rarely on emotional beliefs (e.g., Casalo et al., 2011), but none
recognized both or behavioural intention in the conceptualization of attitude. Therefore, this
thesis argues that while some similarities exist, eWOM trust typically covers some attitudinal
aspects that are not typically mirrored by a consumer’s general attitude towards reviews.
Review attitude is hence a very broad and unspecific attitude towards OCR in general. Due to
its inclusion of more relevant aspects for information adoption (namely trusting attitudes,
beliefs, and behaviours), it is reasonable to assume that eWOM trust is in contrast a better
predictor of review and shopping-related behaviours (Hogg & Vaughan, 2011).

Other Constructs

Aside from the aforementioned constructs, scholars often confuse various other (more general)
concepts with trust. The discussion should guide this thesis to separate these constructs from
trust in the eWOM context. For instance, one conceptual difficulty with the studying of trust is
that it has often been confused with cooperation (Mayer et al., 1995). Among game theorists,
it is usual to view trust as a synonym for cooperation. Deutsch (1962), for example, applies the
concept when referring to cooperation within groups. Additionally, Gambetta (1988, p. 217)
states that to trust somebody means “the probability that he [she] will perform an action that is
beneficial or at least not detrimental to us is high enough for us to consider engaging in some
form of cooperation with him [her]”. However, cooperation cannot be viewed as being the same
as trust. In essence, cooperation is either a consequent manifestation of trust rather than trust
itself (Corritore et al., 2003; Good, 1988). Mayer et al. (1995), for instance, note that
cooperative behaviour can frequently be caused by trust; however, trust is not a necessary
prerequisite for cooperation to exist. According to the authors, this can be attributed to the fact
that cooperation does not necessarily put a party at risk. Kee and Knox (1970) also emphasize
a distinction between the two constructs, as they come to the conclusion that there are numerous
reasons why individuals may engage in a cooperative or competitive behaviour that are not
related to the level of trust within the relationship. Rosseau et al. (1998) also state that there are
a variety of reasons unrelated to trust (e.g., coercion) that may elicit cooperative behaviour.

122
Mayer et al. (1995) put it this way: “you can cooperate with someone who you don’t really
trust”. Similarly, a person can “use” (read) reviews without trusting it, as he/she consults them
only for entertainment purposes or looks for information that confirms his/her aversion towards
them. Therefore, the presence of knowledge exchange in the context of eWOM communication
cannot be completely attributed to the presence of trust.

Trust is also not the same as faith. Both dictionary definitions as well as everyday usage of the
terms suggest that the two concepts can be considered as synonyms. For instance, we may
commonly say, “I have faith in you” to mean “I trust you”. However, faith is the opposite of
reason – and trust, according to its conceptual meaning, always encompasses reasons (Corritore
et al., 2001). To trust somebody or something means that one makes a decision which is based
on reasons to make oneself vulnerable in the conditions of risk and uncertainty. In contrast, to
have faith in somebody or something means to make a leap that is not fully supported by reasons
(Corritore et al., 2003; Macy & Skvoretz, 1998).

Literature also makes no clear distinction between trust and competence (or what is here called
eWOM ability). However, previous trust literature suggests that competence (or ability) is only
one of various cognitive cues for the trustee’s trustworthiness that indicates trust (e.g., Dunn,
2000). For instance, a person’s generalized trust in eWOM is, in part, mirrored in his/her beliefs
about the reviews’ usefulness (Corritore et al., 2003). But trust is a concept that goes beyond
this belief and aggregates cognitive and affective elements as well as behavioural components
(Lewis & Weigert, 1985). A person who states that he/she deems OCR useful does not imply
automatically that he/she is a high eWOM truster.

Definition summarized

As the aim of this research is to develop a reliable as well as valid measure for the construct of
eWOM trust, this section was striving for a literature-based definition and conceptualization of
the construct. This thesis defines eWOM trust as a relatively stable trait-like attitudinal concept
anchored in the consumer’s personality that predisposes individuals to rely on online customer
reviews and recommendations which mirror prior consumption experiences, opinions, as well
as claims of fellow shoppers. The proposed definition mirrors the notion of trust as a complex,
higher-order construct with multiple underlying dimensions (see Figure 2). Here, it is assumed
that the construct is associated with three conceptually distinct types of trust, namely (i) trusting
beliefs, (ii) trusting attitudes, and (iii) trusting intentions. A typology of trust provides the
theoretical basis that these trust types can be distinguished in cognitive, emotional, and
behavioural elements of trust. The conceptualization of eWOM trust distinguishes five
dimensions of trust (i.e., ability, integrity/honesty, benevolence, likeability, and willingness to
rely on). It is assumed that the relative importance of these components depends on the kind of
social relationship resulting in a dominance of cognitive components. The economic goal of
making the right product decision is likely to be the dominant motive in using OCR, while
123
hedonistic motives play a casual role. Finally, this thesis proposes that eWOM trust is
conceptually different from similar constructs, such as review credibility and attitude. In the
remaining part of this section, the research questions and the hypotheses guiding the scale
development process are presented.

Figure 2: Preliminary Conceptualization of Generalized eWOM Trust

eWOM TRUST

Ability Affect Willingness

Integrity Benevolence
(Competence) (Likeability) to rely on

l1 l2 l3 ... C1 C2 C3 ... B1 B2 B3 ... L1 L2 L3 ... W1 W2 W3 ...

TRUSTING BELIEFS TRUSTING ATTITUDE TRUSTING INTENTION

Research Questions and Hypotheses

Until the present day, no actions have been taken to systematically develop a valid, reliable,
and generalizable multi-item scale to measure trust in eWOM. In awareness of this fact, the
present study’s aim is to introduce and test a general measure of this construct. In order to
achieve this thesis’ dedicated goals, thirteen research questions and fifteen hypotheses are
presented to further assess the previously proposed conceptualization of trust in eWOM and
operationalize it as a higher-order construct. Further, these questions and hypotheses target the
evaluation of the measure’s key psychometric properties by testing for various types of validity
(e.g., translation, criterion, nomological), reliability (e.g., test-retest reliability), and provide
first insight into the new scale’s generalizability. Due to the complex nature of the construct,
these assessments are guided by referring to well-acknowledged scale development standards
(e.g., Churchill, 1979; Gerbing & Anderson, 1988; Netemeyer et al., 2003) but specifically
draw on recent studies in the field of domain-specific trust-scale development (e.g., Soh, 2007).
Additionally, the scale’s application as a valuable segmentation tool for online consumers
should advert to the construct’s relevance for present communication challenges faced by
companies (but also consumers) on the Internet. In order to address these key issues, the
proposed research questions and hypotheses follow.

124
Construct Identification and Dimensionality

It is agreed that establishing the construct’s dimensionality is an important first step of the scale
development process (Spector, 1992). The earlier review on the concept of trust has presented
the notion of eWOM trust as a higher-order construct that it is further divided into several
coexisting sub-constructs. As this thesis’ conceptualization and operationalization are
consequently derived from a careful investigation into the interdisciplinary trust research, it
seems reasonable to argue that the multifaceted content of the eWOM trust construct can only
be adequately covered by a scale which includes multiple sub-scales; that is, measures its
subdimensions adequately. However, as the empirical evidence for such a domain-specific
understanding is scarce, further empirical support for the proposed definition, including its
structure and components, is needed. Earlier research on the concept of trust has shown that,
depending on the context, all or at least some dimensions of trust seem to be highly
intercorrelated or should be combined due to their inseparability (Crosby et al., 1990; Ganesan,
1994; Gefen, 1997), while other scholars have found trust to consist of distinct factors (Gefen,
2002b; Gefen & Straub, 2004). The following questions should lead to adequate answers,
including the empirical determination of the validity of the proposed definition of eWOM trust,
its dimensionality, and the construct’s critical factors which together determine its nature.

RQ 1: (a) Is eWOM trust a complex, higher-order construct best described by several

facets or dimensions? (b) Are cognitive, affective, and conative elements best suited
to measure consumer’s generalized tendency to trust customer reviews?

Validity

Generally, construct validity refers to the question of how well a scale actually measures the
construct it is intended to measure (Hair, Black, Babin & Anderson, 2010). By referring to
Calder, Phillips, and Tybout (1982), Netemeyer et al. (2003) note that construct validity can be
best described as any efforts that aim to evaluate the degree to which a measure truly reflects
the concept being investigated or efforts that strive to investigate the covariation in and between
constructs that can be interpreted in terms of theoretical constructs. In order to assess the
validity of a measure, a series of simultaneous tests of hypotheses about the scale and the
construct itself are required (Spector, 1992). The following research questions address the
various validity types that can be used to establish a valid measurement scale. This thesis’
discussion on validity is guided by Netemeyer et al.’s (2003) understanding and classification,
which categorizes the various validity types in accordance with other scholars (e.g., Haynes,
Neslon & Baline, 1999; Trochim, 2012). Here, the authors describe three general types of
validity: (i) translation validity (i.e., the degree to which a construct is translated into the
operationalization of the construct), (ii) criterion validity (i.e., the degree to which the
construct’s operationalization behaves the way it should given to theory), and (iii) nomological
validity (i.e., the extent to which predictions from a formal theoretical network containing the

125
concept under investigation are confirmed (Campbell, 1960)). As the scale’s translation validity
is discussed in Chapter 4, the remainder of this section is dedicated to the examination of the
measure’s criterion-related and nomological validity.

Criterion-Related Validity

In order to test for a scale’s criterion-related validity, its operationalization has to be assessed
against some criterion (Trochim, 2012). Therefore, based on trust theory, some predictions are
proposed about how the eWOM trust scale will perform in the presence of such a criterion. In
order to be in accordance with the standards of judgment (Churchill, 1979; Churchill &
Iacobucci, 2005), multiple research questions are used to target different sub-types of criterion-
related validity.

Concurrent Validity

An important issue of concern is concurrent validity. According to Netemeyer et al. (2003),

evidence for concurrent validity is provided by considerably strong statistical correlations
between the construct measure under development and one or more criteria, which are measured
at the same point of time or “concurrently”. A measure’s concurrent validity is assumed to be
present in the case where the newly developed measure exhibits statistically significant
relations with a measure that previously has been validated. These two measures may quantify
the same construct or two distinct but presumably related constructs. This thesis assumes that
evidence for concurrent validity is provided in cases where the new measure is able to
distinguish between groups of consumers that it should be able to distinguish (Soh, 2007). The
fact that the new scale is able to distinguish consumers that indicated high levels of eWOM
Trust on a single-item scale from consumers that stated a low degree of trust on the same scale,
should provide confidence that the scale possesses validity. Thus, the following research
question is addressed:

RQ 2: Does the developed measure of eWOM show the ability to distinguish groups of
consumers that have indicated varying levels of eWOM trust on a single-item scale?

Convergent Validity

Another way to evaluate the eWOM trust scale’s validity is to assess the extent to which the
proposed operationalization is similar to (or “converges on”) alternative but theoretically
similar operationalizations (Trochim, 2012). In other words, convergent validity is defined as
the degree to which two independent measures which are designed to measure the same
construct are related (Netemeyer et al. 2003). Evidence for convergent validity of the developed
measure is offered by significant and strong correlations between responses obtained by
maximally different methods of measuring the same construct (Churchill, 1979; Netemeyer et

126
al., 2003; Peter, 1981). Hence, in order to investigate the convergent validity of the eWOM trust
scale, the following research question is proposed:

RQ 3: Does the developed measure of eWOM trust correlate significantly and

considerably with other methods to measure eWOM trust?

Discriminant Validity

The aim of discriminant validity evaluations is to ascertain that a measure is indeed novel and
does not simply reflect some other variable (Churchill, 1979). More specifically, discriminant
validity is the investigation into the degree to which the proposed operationalization of the
construct is not similar to (or “diverges from”) alternative operationalizations which should
measure a similar, but conceptually different construct (Netemeyer et al., 2003; Trochim, 2012).
Accordingly, evidence for discriminant validity is indicated where there are “predictably low
correlations between the measure of interest and other measures that are supposedly not
measuring the same variable or construct” (Heeler & Ray, 1972, p. 362). Discriminant as well
as convergent validity are both operational types of validity (Peter, 1981) and are typically
investigated simultaneously by using the Multitrait-Multimethod Approach (MTMM Matrix)
proposed by Campbell and Fiske (1959) – and/or alternative approaches (e.g., Anderson &
Gerbing, 1988). In the previous section, this thesis has discussed eWOM trust construct’s
relationship with review credibility (RCred) and attitude towards reviews in general (RAtt). This
thesis has concluded that these concepts are related but conceptually distinct. In order to verify
the discriminant validity of the developed measure of eWOM trust, the scale should not
correlate very highly with the measures of review credibility and attitude which represent
similar or overlapping but different constructs. Therefore, the following question addressed:

RQ 4: Is the developed measure of eWOM trust significantly different from the

measurements of (a) review credibility and (b) attitude towards reviews in general?

This thesis assumes that the new trust measure is constrained to measure confidence in OCR
and not in other trust objects. Hence, it does not possess the ability to measure a more general
tendency to trust other people or generalized others (i.e., a person’s disposition to trust) nor to
quantify trust in other forms of market communications (e.g., salespersons, advertising). Hence,
it is domain-specific and individuals develop different kinds of generalized trust depending on
the trust object. In order to evaluate whether people develop trust that is specific to customer
reviews, the following research question emphasizes the construct’s relationship with other
trust objects:

RQ 4: (c) Is the developed measure of eWOM trust significantly different from the
measurement of dispositional trust? (d) Do consumers develop trust that is specific
to online customer reviews?

127
Known-Group Validity

In order to argument for construct validity, the measure has to be able to identify groups that
are known to show differences in the focal construct (MacKenzie et al., 2011). According to
Netemeyer et al. (2003), known-group validity targets the extent to which a measure differs as
predicted between groups of people who should score low and high on the focal trait. Hence,
evidence for this kind of validity is provided by significant differences in mean scores of the
trait across these groups. In literature, diverse hints exist that for most market communications
increases in psychographic as well as socio-demographic variables such as intelligence, self-
esteem/self-confidence, education, income, and age (for a discussion see later) are all likely to
have a negative effect on consumer persuasion. In line with earlier contributions in the field of
advertising research (Obermiller & Spangenberg, 1998), it is assumed that such a resistance
against persuasion is indistinguishable from eWOM scepticism which translates into minor
levels of eWOM trust and potentially eWOM distrust. The author feels confident that the
conglomerate of heightened levels of self-confidence, and earlier experiences (e.g., with other
consumers, purchasing), amongst others, together shape a consumer’s generalized tendency to
rely on review-conveyed information. Due to a lack of an adequate knowledge base, limited
analytical abilities and alternative forms of socialization processes, younger consumers are
theorized to have significantly higher levels of trust in OCR. This is also mirrored by earlier
research that concludes that younger Internet users rate information on the Internet as more
trustworthy than older users (Johnson & Kaye, 2002). Likewise, older Internet users are more
sceptical towards the Internet in general (Metzger, Flanagin, Eyal & Lemus, 2003), which may
result in lower levels of information trust. The fact that consumer scepticism is likely to increase
with age is also supported in the offline literature (Boush et al., 1994). Various authors suggest
a negative relationship between age and persuadability in general (Tyler & Schuller, 1991;
Visser & Krosnick, 1998). Therefore, this thesis argues that eWOM trust is limited among
faculty members compared to students because of several reasons. For instance, members of
the faculty are typically older and as a consequence have other and more marketplace
experiences (Obermiller & Spangenberg, 1998). They have learned to use the Internet
differently from younger consumers. Additionally, they possess a higher educational level and
are socialized during an era where the Internet was not available or was in its early phase.
Younger people are used to socializing via the Internet, apply Web 2.0 applications more
regularly and hence have more direct experiences with the trust object. Further, it is assumed
that faculty members have a higher level of self-confidence and feel more comfortable in
making critical product judgments on their own. In addition, faculty members are theorized to
have a higher self-esteem and posses on average higher intelligence. Both circumstances should
cause limited persuadability (McGuire, 1968; Rhodes & Wood, 1992) and hence eWOM trust.

RQ 5: Do students and faculty members differ significantly in their trust in customer

reviews?

128
Predicitve (Postdictive) Validity

According to psychometrics literature, predictive validity is defined as the extent to which a

new measure predicts scores of a criterion measure (Cronbach & Meehl, 1955). Netemeyer et
al. (2003, p. 86) define this validity type as “the ability of a measure to effectively predict some
subsequent and temporally ordered criterion”. In the context of this thesis, this criterion is
referred to as trusting behaviours (i.e., diverse forms of eWOM usage). Prior research shows
that trust and trusting or risk-taking behaviours are closely related (Gefen et al., 2003; Mayer
et al., 1995; McKnight et al., 2002b). This is supported by the Theory of Reasoned Action
(TRA), which advances that attitudes indirectly and behavioural intentions directly (i.e., both
elements of trust) are the most influential predictors of behaviour (Fhishbein & Ajzen, 1975).
An essential assumption of the TRA (as well as the Theory of Planned Behaviour (TPB) (Ajzen,
1991)) is that there is a positive relationship between beliefs, intentions and behaviours.
Accordingly, numerous empirical studies – in different contexts – have found considerable
correlations among these concepts confirming the presence of such a positive, strong
relationship (Hansen, 2005; Sheppard & Sherman, 1998; Vijayasarathy, 2004). In a similar
vein, trust researchers have augmented these findings and provide theoretical as well as
empirical support that trust and trusting behaviours (i.e., risk-taking behaviours) are positively
related (see earlier). In the context of eWOM, trust has been shown to elicit various forms of
trusting behaviours. For instance, when trust is present, individuals liberally exchange, seek,
and collect knowledge gained through online reviews and recommendations (Chen & Hung,
2010). Trust between the members of a virtual community has been demonstrated to affect
members’ behaviours, such as obtaining and contributing information (Kankanhalli, Tan &
Wei, 2005; Lu et al., 2010). Based on these prior findings, this thesis is interested in
investigating the new measure’s ability to effectively predict subsequent and temporally
ordered behaviours related to eWOM communications. Similarly, the measure should be able
to postdict prior behaviours. Hence, the following research question is asked:

RQ 6: Does the eWOM trust measure significantly predict (postdict) eWOM-related

consumer behaviours (e.g., opinion seeking, adoption, giving, passing)?

Nomological Validity: A Social Shopping – Trust Framework

The scale development paradigm postulates that the validity of a measure is also determined by
its “lawful” fit into a set of hypothesized interrelations of the intended construct with other
constructs (MacKenzie et al., 2011; Netemeyer et al., 2003; Spector, 1992) or a “nomological
network” (Cronbach & Meehl, 1955). Nunnally and Bernstein (1994) also note that an
important aspect of the scale-validation strategy is to evaluate the extent to which a measure
operates within a set of theoretical constructs and their respective measures. A novel measure
has to show its nomological validity in order to be accepted as valid (Peter, 1981). The
assessment of validity type is based on investigations on the construct and its measure by
129
formulating hypotheses derived from theory about the construct’s causes, correlates, and
consequences of the construct under scrutiny. Then, nomological validity of the eWOM trust
scale is evidenced if the eWOM construct is not only theoretically related to other constructs
but also empirically, and behaves in a manner that is consistent with prior theory (i.e., the
measures correlate significantly in the predicted direction) (MacKenzie et al., 2011).

To examine nomological validity and to demonstrate the construct’s relevance, this thesis
embeds the eWOM construct into a network of intertwined constructs. While in the literature
on trust it seems usual to use the model “antecedents-trust-consequences” to investigate trust
(Shankar et al., 2002), the proposed framework specifically adopts the work of Obermiller and
Spangenberg (1998) and extends it to the trust context. This approach demonstrates the focal
construct’s role in eWOM persuasiveness and additionally enables the profiling of eWOM
trusters in respect to various attitudinal, behavioural and socio-economical characteristics. The
relationship of eWOM trust with its antecedents, correlates, and consequences is illustrated in
Figure 3. In this research antecedents are defined as constructs that are theorized to cause the
focal construct. In contrast, consequences are constructs that are themselves caused by the focal
construct and correlates are constructs whose conceptual definitions overlap with the focal
construct (MacKenzie et al., 2011).

This study defines eWOM trust as a relatively stable and a generalizable disposition towards
online customer reviews and more specifically as a personal difference characteristic of the
truster, describing his/her tendency towards relying on information conveyed in eWOM. By
acknowledging the construct as being a internally-caused variable of the human being, this
research agrees with the psychological literature that eWOM trust results from more basic
personality traits, prior consumption experiences, introspective and extrospective observations,
and social-emotional attitudes, such as an individual’s tendency to associate and to bond with
similar others. By now, several frameworks have been proposed to classify the antecedents of
trust (Gefen et al., 2003; Kim et al., 2008; Zucker, 1986). In reference to these contributions,
the eWOM trust antecedents are categorized as:

(1) Personality-based trust antecedents (i.e., truster’s stable personality characteristics, such as
disposition to trust, general scepticism, cynism, susceptibility and openness to experience). For
instance, cynism expresses an individual’s belief in the untrustworthiness of others which leads
to an orientation of being hostile to others (Williams, Barefoot & Shekelle, 1985). Hence, since
this trait is transferred to any life situation, it is likely that cynics also do not have trust in
assertions made by fellow consumers in online reviews and recommendations. Hence, they may
be less responsive and persuable to eWOM in general as they simply don’t trust it.

(2) Experience-based trust antecedents which focus on trust building through earlier
consumption experiences (e.g., experiencing selfish or opportunistic behaviours of reviewers,
age, education) as well as repeated interactions and familiarity with the object of trust. For
130
instance, consumers who had only positive experiences in the past (e.g., because reviews helped
them to shape their buying decision and guided them to a good shopping outcome) are likely to
generalize such experiences and anticipate them to the present. Hence, they are more willing to
rely on eWOM claims.

(3) Cognitive-based trust antecedents refers to trust building through cognitive evaluation
processes and is associated with self-observations (e.g., a generalized attitude towards eWOM
information quality), self-perceptions (e.g., perceptions institutional reliability, schemata about
reviewer and channel characteristics), and self-interests (e.g., eWOM image). For example, it
is theorized that eWOM trust is caused by various general perceptions or attitudes of the
reviewer community (e.g., reviewer credibility) which result from a social learning process or
the observation of others’ behaviours.

(4) Affect-Social-based trust antecedents refers to generalized emotional and social bonds
between the reader and the reviewer. This category includes attitudes as well as traits related to
interpersonal relationships. For example, literature agrees that social similarities contribute to
trust building. Various scholars (Duhan et al., 1997; Feick & Higie, 1992; Gilly et al., 1998;
Smith et al., 2005) have suggested that perceived rapport and similarity (i.e., homophily)
between a consumer and an information agent (reviewer or recommender) with respect to
demographic characteristics, lifestyles, tastes, attitudes, personality and interests may also
influence the amount of trust a consumer places in the information agent. This view is strongly
supported in eWOM literature.

Hence, various determinants or underlying conditions are likely to shape the consumer’s
orientation to have confidence in relying on eWOM. On the other hand, the focal psychological
construct is also associated with related beliefs and attitudes towards the marketplace. For
instance, eWOM trust is a correlate of consumer’s generalized (e.g., eWOM attitude) or narrow
attitude towards online reviews (e.g., eWOM helpfulness), eWOM distrust or his/her
involvement in peer information search. eWOM trust may also impact an individual’s attitude
toward the Internet as a search or shopping tool. Consumers who have a lack of trust in eWOM
are inclined to dislike eWOM for that reason and thus as the Internet nowadays is more or less
a synonym for interactive social interaction and shopping, they are inclined to dislike Internet
product search in general.

Further, it is hypothesized that eWOM trust impacts responses to eWOM appeals and has
ultimate consequences on product/brand attitudes and information-gathering behaviours. In
contrast to traditional WOM, which involves personal communication between friends, family
and affiliates, eWOM is characterized by the anonymity of the interaction partners. Hence,
consumers tend to doubt the information in general. In such a situation, the importance of the
individual’s disposition to trust eWOM is heightened as different levels of trust may lead to
different attitudes toward an object (Kim & Benbasat, 2003). Therefore, in accordance with
131
prior findings (Wang, 2010; Yu & Tang, 2010) it is expected that generalized eWOM trust
indirectly moderates the relationship between eWOM appeals and consumer reactions (e.g.,
product/brand attitude) through situational eWOM trust. This thesis assumes that if a
consumer’s trust in eWOM is high, he/she should attend more to eWOM, should be less likely
to counter argue, and should form beliefs that are consistent with eWOM claims. For instance,
if an individual has a strong belief in positive eWOM, he/she will form more positive attitudes
toward the product and will feel good about using the product. Therefore, high trusters are more
likely to use eWOM for product information and should be more persuaded by online reviews
and recommendations in general. In contrast, low trusters are less likely to be convinced by
positive eWOM, which means that they will not form positive product/brand attitudes which
mirror the positive opinions of others.

eWOM trust’s moderating role is itself affected by individual factors (e.g., consumer’s
knowledge, motivation, involvement) which have been shown to determine persuasion effects
but also by additional situational factors that are likely to impact consumers’ acceptance of
eWOM claims (Obermiller & Spangenberg, 1998). Hence, eWOM message perceptions may
vary depending on the shopping occasion (e.g., product type) but also the situational
consideration of characteristics of the reviews just read. For instance, some reviews may be
perceived as more manipulative than others. This perception may influence the situational trust
towards the just read reviews; this results in specific reactions towards the product/brand.
Hence, as described earlier, a person’s reactions to specific reviews is mainly guided by
situational trust aspects. However, generalized eWOM trust has a great indirect influence on
these reactions through situational trust. In general, individuals are not able to evaluate whether
or not specific reviews are trustworthy (e.g., cues for the source’s reputation are missing).
Hence, they use generalizations as a proxy to determine their level of trust. Generalized eWOM
trust is especially influential when peripheral route processing is used. It functions as a schema
or cognitive short-cut.

Based on this nomological network, selected relationships between eWOM trust and its
theoretically related constructs should be tested empirically. Accordingly, the following
hypotheses and propositions are introduced. While the hypotheses target a further empirical
assessment in this research, the aim of the presented propositions is to demonstrate theorized
relationships that may be investigated empirically in future empirical research.

132
 ANTECEDENTS RELATED CONCEPTS CONSEQUENCES

PERSONALITY-BASED TRUST
ANTECEDENTS eWOM Appeals
(PERSONALITY TRAITS) e.g., Valence

e.g., Disposition to Trust

Skepticism
Extravision, Sociability
Openness, Susceptibility Generalized Situational
Agreeableness eWOM TRUST eWOM TRUST
(Personality) (Environment)
EXPERIENCE-BASED TRUST Product Beliefs
Figure 3: The Nomological Framework

ANTECEDENTS
(CONSUMPTION EXPERIENCE)
e.g., Indiednces of
opportunistic Behavior RELATED BELIEFS eWOM Information
Product Attitudes
Familiarity AND ATTITUDES Processing
Age (CORRELATES)
Education
e.g., eWOM Attitudes
eWOM Distrust eWOM
COGNITIVE-BASED TRUST eWOM Involvement Persuasiveness
ANTECEDENTS
Self Observations
Self Perceptions
Self-Interests
INDIVIDUAL FACTORS SITUATIONAL FACTORS
AFFECT-SOCIAL-BASED TRUST e.g., Knowledge/Expertise e.g., eWOM Characteristics
ANTECEDENTS Motivation Product Type
Product Involvement Message Context
e.g., Tie Strength
Homophily
Interpersonal Orientation
SITUATIONAL INFLUENCES

133
eWOM Trust Antecedents

Disposition to Trust

eWOM trust is inseparably linked to the individual’s personality. Disposition to trust (TDispo),
also called trust propensity, is a personality trait that basically refers to the generalized tendency
of a person to trust others (Ridings et al., 2002). Specifically, these authors describe the
construct as “a general willingness based on extended socialization and life experience to
depend on others” (p. 278). Hofstede (1980) has also found that this trait is related to the
individual’s cultural background, personality type, and developmental experience. People’s
difference in their tendency to trust others is not based upon experiences with or knowledge of
a specific trusted party, but is instead the consequence of ongoing life experiences and
socializations (Fukuyama, 1995; Kim et al., 2008; Kramer, 1999; Rotter, 1971). In line with
prior and subsequent research, disposition to trust describes an individual’s tendency to believe
or not to believe others (Gefen et al., 2003). This is regardless of who the interaction partners
are and the actual situation (Casalo et al., 2008). Disposition to trust results from the
individual’s inclination to display faith in humanity and to adopt a trusting stance towards
others, independent of time and space (Gefen, 2000; McKnight et al., 1998) (for a detailed
review see Chapter 2).

Various research provides evidence that disposition to trust is related to (interpersonal) trust
development in a variety of situations (Gefen, 2000; Mayer et al., 1995). That is, it is often seen
as an antecedent of this construct and it has been discussed as being specifically effective when
the trusting parties are still unfamiliar with one another (Liu & Zhang, 2010; McKnight et al.,
2004; Utz et al., 2012). Here, people apply disposition to trust as a kind of heuristic to overcome
situational complexity. This has been proven, for instance, by investigations into the
development of initial trust in a business-to-consumer environment, where it has been shown
that the trust of individuals who had quite rare interactions with a selling party are
predominantly affected by their tendency to trust others in general. Likewise, disposition to
trust positively influences the initial trust of a consumer in a website (Lu & Zhou, 2007). In the
context of e-commerce, dispositional trust impacts a person’s trust perceptions in other trusting
entities (e.g., Gefen, 2000; McKnight & Chervany, 2002). The very same influence is also likely
to be present in the context of online reviews. A distinctive characteristic of eWOM is that the
reviews and recommendations often stem from persons with whom the reader is not familiar
and the interaction parties have no or little prior relationship with one another (Lee & Koo,
2012). Ridings et al. (2002) argue similarly by stating that in virtual communities it is very
likely that individuals do not have much information about one another. Accordingly, the
authors are able to show that if an individual is predisposed to trust others in general, the same
willingness also guides him/her to trust other community members. Additional support for a
close relationship comes from the work of Casalo et al. (2008). Given the anonymous character

134
of the eWOM communication, a person’s disposition to trust should be a good predictor of
generalized eWOM trust. Therefore, it is hypothesized that:

H 1: A person’s trust in eWOM is significantly and positively related to the individual’s

disposition to trust.

Self-confidence

An individual’s trust in eWOM may be affected by a number of personal predispositions and

personality traits. One of these personality traits is the individual’s level of self-confidence or
self-efficacy (the term is here used interchangeably). The construct has been identified as a
critical factor for explaining individuals’ behaviours in a social context (Reynolds & Darden,
1971) and exerting influence in the context of spreading and seeking of WOM information. The
close relationship between trust and self-efficacy can be drawn from the social cognitive theory
(SCT). Research addressing this theory has heavily investigated the topic of how people
perceive themselves and others in a variety of social situations. A key assertion of SCT is that
individuals have specific goals – each one of them is associated with an expectation of a certain
outcome. In order to achieve these goals, individuals constantly process environmental
information and adapt their behaviour accordingly (McCormick & Marinko, 2004). In this
context, individuals develop their own uncertainty control strategies (Kim et al., 2008). Their
reactions towards and their evaluations of the external environment are guided by the
individuals’ personal standards and their perceived self-confidence (Wood & Bandura, 1989.
Bandura (1977b) states that the concept of self-efficacy (self-confidence) can be generally
described as a person’s perception of how easy (or difficult) it would be to perform a certain
behaviour. The author (Bandura 1982, p. 122) defines self-efficacy as “how well one can
execute courses of action required to deal with prospective situations”. Some years later,
Bandura (1986, p. 391) adapted his view and redefined the concept as “people’s judgments of
their capabilities to organize and execute courses of action required to attain designated types
of performance”. Research views self-confidence as a major variable that underlies human
motivations (Venkatesh et al., 2003), affects perceptions and indirectly influences final
behaviours (Hernandez et al., 2009). Accordingly, Bandura (1991, p. 257) notes that “people’s
beliefs in their efficacy influence their choices, their aspirations, how much effort they mobilize
in a given endeavor, and how long they persevere in the face of difficulties and setbacks”. In
parallel, the concept should not be confused with being a measure of objective skills. Rather,
self-efficacy should be understood as the extent to which a person believes that he/she is able
to perform a specific task by applying his/her skills (Eastin & LaRose, 2000). In a similar vein,
it has been regarded as a basic belief that an individual is capable of activating a certain
motivation which is critical to overcome problems of a specific situation (Wood & Bandura,
1989).

135
It is widely agreed that there exist two different forms: (1) specific self-confidence and (2)
general self-confidence. According to Lau and Ng (2001), the former is related to a person’s
belief to perform a specific task, while the latter is aimed to describe a person’s overall self-
confidence.

Self-confidence can be developed in the face of various skills and across different situations
(e.g., economic self-confidence, social self-confidence, legal self-confidence, Internet self-
confidence) (Kucuk, 2009). For instance, the social aspect of self-confidence is highlighted in
the work of Mourali et al. (2005, p. 309), who state that self-confidence refers to the “differing
ways in which individuals relate to each other. Some people are inherently more comfortable
than others in interpersonal interactions”. General self-confidence reflects these facets and is
expected in this research to have a relationship with a consumer’s trust in eWOM. The concept
has been defined as “the extent to which an individual believes himself to be capable,
significant, successful, and worthy” (Locander & Herman, 1979, p. 270). More specifically,
this thesis narrows the notion of self-confidence down to the consumption context and adapts
the view of Bearden (2001) of a consumer’s self-confidence in marketplace decisions by
describing it as “the extent to which an individual feels capable and assured with respect to his
or her marketplace decisions and behaviours” (p. 122) and expects a clear relationship between
a consumer’s general self-confidence and eWOM trust. In this research it is assumed that
consumers with a low level of self-confidence, may be more likely to engage (passively) in
eWOM communication and may place more trust in information from others (Levy, 2012).
Self-confidence is regarded as a critical factor that determines whether a consumer seeks word-
of-mouth communication or not (Bansal, 1998). Research in the offline world has made some
investigations into the relationship between self-confidence and opinion seeking. For instance,
the research of Kiel and Layton (1981) provides the insight that information search is negatively
related with self-confidence. This essentially means that consumers who have low levels of
self-confidence put a lot of time and effort in the collection of external information. They do
not trust themselves. Likewise, it has been shown that higher self-esteem, another personality
trait which is closely related to the concept of self-confidence, guides individuals to lower levels
of opinion seeking (Pornpitakpan, 2004a). Wood and Stanger (1994) argue that individuals who
are more confident in their own judgments are consequentially less influenced by others'
opinions. At its heart, this thesis argues that consumers with low self-confidence are more likely
to consult the opinions of fellow shoppers in order to find out more about the thoughts and
recommendations. This enables them to make better-informed purchasing decision and to get
confirmation from others. Collection of external information furnishes them with confidence
that they possess an adequate level of information which they otherwise would not have. The
work of Mourali et al. (2005) supports this view by stating that the consumers with “good” self-
confidence are more likely to rely on their own knowledge as well as skills when making
decisions and do not rely on others. General self-confidence has also been shown to be a
determinant of subjective product knowledge (Park et al., 1994), which makes persons who
136
exhibit heightened degrees of self-confidence less likely to collect information from others, but
to rely on their own person. Similarly, Cox (1962) finds that persons with high self-confidence
regularly reject advice because they feel that they do not need it and instead prefer to trust their
own judgment. Later, this is also supported by Webster (1968) and Cox and Bauer (1964) who
conclude that persons who are low in self-esteem are generally more susceptible to
interpersonal influence. Various authors emphasize the nature of the construct of self-
confidence by generally stating that persons with high self-confidence are most likely to believe
that difficult tasks (e.g., making a purchase decision) are something that can be mastered on
their own (Banbura, 1982) – without the support of others. Hence, it is more likely that those
consumers challenge themselves and their purchase knowledge in order to solve difficulties in
an uncertain purchase environment (Kucuk, 2009; MacDonald & Uncles, 2007). In essence,
self-confident persons trust themselves more than others, especially when they belong to an
anonymous crowd. Hence, the following hypothesis is proposed:

H 2: A person’s trust in eWOM is significantly and negatively related to the individual’s

self-confidence.

Risk Propensity

A key consideration of risk literature is the assumption that one’s tendency to take (avoid) a
risk across different domains and variety of situations is a stable personality trait (e.g., Wang et
al., 2009). As a consequence, persons can be classified as having risk-seeking or risk-averse
decision styles – which can be regularly observed in respective research (Bromiley & Curley,
1992; Farmer, 1993; MacCrimmon & Wehrung, 1990). In this thesis, the author adopts the view
of Sitkin et al., who define risk propensity as an individual’s current tendency either to take or
avoid risks (Sitkin & Pablo, 1992). Similarly, MacCrimmon and Wehrung (1990) define the
concept as a person’s generalizable willingness to take risks. The idea that this propensity may
have a significant influence on decision making under conditions of risk and uncertainty is
grounded in various previous studies (e.g., Gosh & Ray, 1992; Kim, 1992; Taylor & Dunnette,
1974). An often-emphasized aspect of risk propensity is its impact on the relative salience of
contextual perils or prospects and, hence, results in altered risk perceptions (Brockhaus, 1980;
Vlek & Stallen, 1980). For instance, risk-averse individuals are theorized to put more weight
on potentially negative outcomes than positive ones (Schneider & Lopes, 1986). This leads to
an overestimation of the probability of loss relative to potential gain. Research has demonstrated
that a reverse relationship is also true for risk-seeking people (March & Shapira, 1987).
Individuals with an enhanced orientation towards the acceptance of risks are also predisposed
to taking riskier decisions than others (Brockhaus, 1980) and more positive attitudes. As
demonstrated earlier, trust is deeply related with risk. Regularly, trust is defined as the
willingness of an individual to rely on something or, in other words, to take some kind of risk.
In the context of eWOM, the acceptance of personal vulnerability is equalized to this risk. In

137
line with risk literature, it is theorized that risk-seeking consumers are more likely to
overestimate the positive outcomes of relying on this kind of information in contrast to risk-
averse individuals. They think that they considerably benefit from others’ experiences, as the
reviews provide them with a picture of consumption reality. In other words, risk seekers tend
to trust eWOM more easily than others because they generally perceive very limited risk – like
the risk of being deceived and manipulated. Thus, the following hypothesis is proposed:

H 3: A person’s trust in eWOM is significantly and positively related to the individual’s

risk propensity.

Extraversion

Another personality trait likely to affect an individual’s level of eWOM trust is extraversion, as
the concept has been found to be closely related to the quality of social interactions (Costa &
McCrae, 1989). Extraversion appears in one form or another in most personality inventories
often being attributed as a part of the “big five” personality traits (Norman, 1963) – or even
forming with neuroticism the “big two” (Wiggins, 1968). In general, extrovert persons are said
to be focused on the outside world (Walczuch & Lundgren, 2004). Because of this, they are
more sociable, open to others, like to be in other people’s company, as well as being interested
in the well-being of other individuals (Fahr & Irlenbusch, 2008). Additionally, extrovert people
can be characterized as lively, spontaneous, careless, and enthusiastic. Accordingly, Barnes et
al. (2007) characterize extroverts as talkative and sociable. As using eWOM (communicating
with the world) is a truly outward directed social behaviour, it can be argued that an extrovert
person is generally more likely to trust eWOM communication compared to individuals scoring
low on this trait. This is also supported by previous reseach (Gholamisaman, 2012), that
identifies extraversion as a major factor of opinion seeking and giving in social networks.
Matzler et al. (2006) find in their research that extroversive consumers have a higher level of
socialization and are also more willing to take other people’s suggestions into consideration.
This may be due to the fact that extrovert persons also have a higher degree of affiliation, which
makes them more likely to have warm and friendly feelings towards others and value close
interpersonal relationships with fellow consumers (Lee, 2002). People scoring high on
gregariousness, a factor that is often described as being an integral part of the concept of
extraversion, are said to have strong social needs and have trouble being alone (Wiggins, 1994).
Arndt (1967) also provides evidence that sociable persons are more likely to engage in
traditional WOM communication than those who are less sociable. Sociable persons are
outgoing and are therefore more likely to actively participate in online social networks and be
influenced by these networks as a result (Valck et al., 2009), as well as perceive trust in similar
information such as OCR. In parallel, extroversion may have a positive impact on an
individual’s intention to form trust in general (Tan & Sutherland, 2004), and therefore also may
lead indirectly to higher levels of eWOM trust. Hence, this thesis proposes:

138
 P 1: A person’s trust in eWOM is significantly and positively related to the individual’s
extraversion.

Agreeableness

The personality trait of agreeableness is often described in terms like social adaptability,
likeability, friendly compliance, and love (John & Srivastava, 1999). Individuals who have a
higher level of agreeableness are more likely to form friendly relationships with others (Digman
& Inouye, 1986) and have higher trust (Tsao & Chang, 2010). Agreeable persons have, in
general, positive beliefs towards others and appreciate their values and convictions (Walczuch
& Lundgren, 2004). Individuals who score high on agreeableness are best described with terms
like friendly, helpful, altruistic, modest, trusted, and straightforward (e.g., Picazo-Vela et al.,
2010). In contrast, less-agreeable people are found to be more suspecting and less trusting
(Pervin et al., 2004). Additionally, these people have little respect for others’ interests and well-
being and do not care much about social norms (Walczuch & Lundgren, 2004). In literature, it
is widely agreed that individuals who feel respect for others also believe that others have respect
for them. Hence, it is assumed that consumers who are high in agreeableness are more easily
trusting and influenced by the words of others online.

P 2: A person’s trust in eWOM is significantly and positively related to the individual’s

agreeableness.

Openness (to Experience)

Matzler et al. (2006) state that besides the personality trait of extraversion, openness should be
examined because of its potential relevance for trust perceptions. Literature provides several
hints for describing open people. For instance, Tsao and Chang (2010) characterize individuals
who are more open to experience as having more imagination and curiosity, like variability,
and less likely to be prejudiced (e.g., Barrick & Mount, 1991). These authors further notice that
persons who are more open to experience are more willing to consider various viewpoints and
others’ opinions, and seek out new opportunities to learn new things. Hence, it can be assumed
that these individuals are also more likely to use and trust the recommendations and views of
fellow shoppers in the online medium. Open people represent a specific kind of consumer that
generally tends to be more knowledge-seeking than others, preferring alternative routes of
information flow – aside from corporate information. These consumers do comparison
shopping on the Internet and are constantly looking for the latest and the greatest or the best
values (Tsao & Chang, 2010). The concept of openness to experience is able to explain
behaviours and intentions of individuals who are curious, insightful, and also have a broad
range of interests (Ranjbarian & Kia, 2010). More openness leads to a greater willingness to
take risks (Diehm & Armatas, 2004) and more carelessness with respect to new situations and
experiences. Therefore, open people are expected to be less aware of the potential risks and

139
dangers present in mostly anonymous eWOM communication. It is assumed that people with a
high openness to experience are also more likely to trust eWOM:

P 3: A person’s trust in eWOM in general is significantly and positively related to the

individual’s openness.

Cynicism

Literature suggests that cynicism should show a negative relationship to eWOM trust.
According to Kanter and Wortzel (1985, p. 6), cynicism is defined as “the suspicion of other
people’s motives, faithfulness, and goodwill”. In their research, the authors directly linked the
concept to the perceived believability of advertising claims. Kanter and Wortzel (1985) argue
further that cynics are less likely to believe in information from any source and are especially
likely to attribute advertising claims to (nebulous, one-sided) selling motives rather than strict
honesty and passion towards the consumer (Obermiller & Spangenberg, 1998). This thesis
assumes that a similar pattern is true for eWOM communication. Here, consumers with a
heightened cynical personality are more likely to consider the “dark sides” of reviews and
recommendations. They generally assume that reviewers are driven by selfish or purely
economic motives that distort the content of their communication. Reviews mirror the interests
of clever marketers that want to steer consumers wrong. Cynics may regard themselves being
“used” by the reviewers or being manipulated. This view corresponds also with the description
of the cynicism concept by Kanter and Mirvis (1989, p. 3), who define it as “a sense of being
let down or of letting oneself down, and more darkly, the sense of being deceived, betrayed or
used by others”. Cynicism is often described as a learned, unfavourable attitude that can be
broad or specific in its focus. It is a complex concept that is represented by cognitive, affective,
and behavioural components (Stanley et al., 2005). Others view it as a “negative affect” (Dean
et al., 1998). Cynicism is related to feelings of manipulation or ethical violation and individuals
– who can be best described as having a cynical personality – typically assume that others
exploit them for their own interest (Chaloupka, 1999). In general, cynicism is commonly related
to terms including suspicion, distrust, scepticism, dissatisfaction, alienation, and resistance or
even hostility towards the agent (Chylinski & Chu, 2010). At the heart of the cynicism concept
is the consumer’s assumption about the others’ pretence of unselfishness to mask selfish goals
(Helm, 2004). In parallel, this is associated with the general criticism about eWOM messages
as well. In turn, it is proposed that:

P 4: A person’s trust in eWOM is significantly and negatively related to the individual’s

cynicism.

140
Consumer Experience

Various trust researchers agree that trust develops over time as it is built up in a gradual manner
through a series of interactions (Hart & Saunders, 1997; Kumar, 1996; Ramawami et al., 1997).
Blau (1964), for instance, explains that familiarity or experience are antecedents of trust as he
states that trust is built through beneficial earlier interactions with the trusted party. Through
experience, the trusting party acquires beliefs concerning the characteristics (i.e., the
trustworthiness) of the trusted party (Gefen, 2000; Giffin, 1967; Mayer et al., 1995). This
ultimately determines general orientations towards the trust object. The person learns from
instrumental conditioning. Hence, in the eWOM context, it seems also reasonable that
consumers who had positive experiences with eWOM in the past (that is received positive
reinforcement) tend to have a more positive orientation towards it. And as a consequence, they
are more likely to place more trust in this kind of information at the present. Positive
experiences with eWOM arise when consumers have the impression that the usage of eWOM
enabled them to be more efficient during the buying process, as this kind of information was
perceived to be a valuable source of purchase information. They mean that they did better in
their purchase decisions by trusting OCR. Additionally, persons who place higher trust in
eWOM are more unlikely to be exposed to negative experiences such as opportunistic
behaviours of the reviewers and information that steered them wrong, leading to negative
reinforcement. Hence, it is argued:

P 5: A person’s trust in eWOM in general is significantly and positively related to the

individual’s prior positive eWOM experience.

Reviewer Credibility

Online consumers’ trust in eWOM is likely to be influenced by a set of assumptions about the
creators of eWOM as a social group; that is, reviewer stereotype and role schema (Hogg &
Vaughan, 2011). Here, one factor that is likely to form a constant level of eWOM trust is the
reader’s general attitude towards the reviewers’ credibility (SCred). Source credibility has been
repeatedly regarded as a “trust facilitator” in a variety of contexts (Barber, 1983). Recently, the
concept was also introduced to eWOM research (Ku et al., 2012). In general, it is here assumed
that consumers who view reviewers as highly credible are also more likely to have more trust
in the information conveyed in the reviews written by these reviewers. Consequently, research
has demonstrated that opinions of credible or reputable sources are perceived as more likely to
be adopted by the members of the consumer community (Komiak & Benbasat, 2006). In the
eWOM context, two different aspects can be identified to be related to the credibility of eWOM
reviewers (see also previous discussion). First, sender credibility is based on his/her
trustworthiness (i.e., believability) as the concept of reputation mirrors a collective measure of
reliability of the reviewer population. Various research contributions support the view that
sender trustworthiness is a critical determinant of eWOM acceptance (Lee et al., 2011). The
141
second major factor that determines reviewer credibility is the expertise of the reviewer. The
latter refers to the extent to which readers of eWOM believe a source to have a high degree of
skill in or knowledge of a particular subject (Martin & Lueg, 2013). The degree to which a
source is said to have expertise is determined by the evaluation of multiple aspects of the source,
such as knowledge, ability, or skills (Erdogan, 1999). Generally, persons who are trained or are
in the possession of significant information in a specific area are viewed to have greater
expertise than others. If eWOM readers perceive the sender’s expertise as unacceptable because
the reviewer is perceived as having inadequate experience with the brand, he/she will appear
less credible or reputable. In contrast, if the expertise level of the contributing reviewer appears
adequate, the sender is deemed to be more reputable and knowledgeable. Related research has
shown that source expertise exerts positive influence on consumer behaviour. In line with this
literature string, it is assumed that the reviewer’s opinions and recommendations are more likely
to be adopted by consumers when they generally assume that the information source possesses
a high level of expertise. This view is supported by various research approaches that all agree
that such sources exert greater influence over consumers than do other sources (Sweeney et al.,
2008; Woodside & Davenport, 1974). For instance, proficient consumers have been found to
be often consulted by others (Gilly et al., 1998). The sender’s expertise is an influential factor
of information (Solomon et al., 2006) and the receiver’s search extent (Bansal & Voyer, 2000).
Hovland and Weiss (1951) identified sources who possess relevant product knowledge or
expertise as significantly more influential than other sources. The author also showed that
changes in a consumer’s attitude and opinion can be attributed to this concept. Hence, various
studies provide sound evidence that a sender’s level of expertise and trustworthiness both have
a positive influence on influence of the endorser (Feick & Higie, 1992; Gilly et al., 1998;
McCracken, 1989). Reviewer credibility perceptions been found to make shoppers more trustful
in the online peer recommender (Chang et al., 2008; Smith et al., 2005). Hence, this thesis
proposes that generalized perceptions of reviewer credibility are likely to impact the level of
trust that consumers place in eWOM in general:

H 4: A person’s trust in eWOM is significantly and positively related to perceived

reviewer credibility.

Reviewer Conscientiousness

The concept of conscientiousness is often described with terms like dependability, task interest,
willingness to achieve, impulse control, and dedicated work (John & Srivastava, 1999). A
recent literature review conducted by Picazo-Vela et al. (2010) describes highly conscientious
individuals as being thorough, responsible, self-disciplined, organized, self-motivated,
achievement-oriented and task-oriented (Barrick & Mount, 1991; Costa & McCrae, 1992b;
Goldberg, 1993), but also cooperative (Molleman et al., 2004). A similar description is
provided by Walczuch et al., who emphasize that persons who score high on conscientiousness

142
are thought to be responsible, dutiful, and trustworthy (Walczuch & Lundgren, 2004).
Conscientious individuals tend to be more serious and are cautious in making their purchase
decisions. Previous research on this characteristic states that persons scoring high on the trait
are purposeful as well as dependable (Witt et al., 2002) and are striving for the accomplishment
of their goals (Costa & McCrae, 1992a). These attributes are generally valued by eWOM
readers. It is hence theorized that consumers who perceive reviewers and recommenders as
conscientious persons are more likely to trust the information provided by such individuals.

P 6: A person’s trust in eWOM is significantly and positively related to perceived

reviewer conscientiousness.

Reviewer Altruism and Egoism

When consumers place trust in eWOM, it is likely that this willingness is influenced by the
belief in the senders’ basic motivations that are beneficial to the reader. One of these beneficial
motivations is altruism. In general, altruism denotes voluntary helping actions with the ultimate
goal to increase the welfare of other individuals or the community (Bickart & Schindler, 2001;
Fang & Chiu, 2010). Altruism is basically an act of doing something for others and serving the
public good without anticipating any extrinsic reward. The motive of the public good is linked
to empathic emotions (Cheung & Lee, 2012). Altruistic persons are characterized by their
generosity and kindness towards others. Empathy, a concept that mirrors feelings of sympathy,
compassion, and tenderness, is viewed as the source of altruism. Empathy is a moral emotion
which mirrors the person’s feeling that the pain of others matters to him/her. Previous research
provides evidence that empathetic individuals are more willing to help others (Cheung & Lee,
2012). Hence, altruism is likely to lead to knowledge-sharing behaviour out of the passion of
being helpful to others and the desire to improve their welfare. Various researchers have
regarded altruism as a critical factor that explains individuals’ willingness to share their
knowledge (Ba, 2001; Davenport & Prusak, 1998; Palmer, 1991) and it has been often described
as an important motivation that drives people to actively participate in offline/online
communities or social spaces (e.g., Dellarocas et al., 2007; Eagly & Chaiken, 1993) as well as
accept message content. For instance, Sundaram et al. (1998) find altruism being a motive for
WOM communication. The researchers conclude that informants are driven by their passion to
enable others to have good experiences as they have perceived on their own or they try to
prevent others from experience the problems they have encounter earlier. This research follows
the work of Dichter (1966), who identified altruism as one of the major motivations for
engaging in verbal information exchange. By studying the knowledge-contribution behaviour
in electronic networks, Wasko and Faraj (2005) identified altruism among other motives (i.e.,
reputation, general reciprocity, and community interest) as a significant factor determining
information exchange. Similarly, Kwok and Gao (2004) also highlighted the motivation of
altruism as one of the key determinants of knowledge contribution in peer-to-peer communities.

143
In their research, Hennig-Thurau et al. (2004) state that the concern for others is closely
intertwined with the concept of altruism (Jeong & Jang, 2011). They conclude that consumers
furnish others with product information in eWOM because of the person’s desire to help fellow
consumers in their buying decision, to save other persons from negative experiences, or both.
Therefore, this thesis assumes that consumers are more willing to trust online reviews and
recommendations when the senders of this information are considered to act on the basis of an
altruistic motivation and not due to other (e.g., selfish) backgrounds. This assumption seems to
be well supported by trust literature, which suggests that altruism contributes to the creation of
trust (Frost et al., 1978). Altruism is regarded as being a related but separate concept from trust.
While Hosmer (1995) is not clear on this, others disassociate trust from naiveté, stupidity, and
altruism (Friedland, 1990). The latter argument seems to be the dominant view in trust literature
today. The same literature string however also supports a close relationship between altruism
and components of trust. For instance, the truster’s belief of benevolence is typically seen to be
mainly based on perceptions of the altruistic motivations of the trustee (Mayer et al., 1995; Wu
et al., 2010). Hence, it seems reasonable to argue that perceived sender altruism is a critical
determinant of trusting beliefs and therefore a likely antecedent of eWOM trust in general.

H 5: A person’s trust in eWOM is significantly and positively related to perceived

reviewer altruism.

Various other general perceptions or attitudes towards the reviewer could have an impact on
eWOM trust formation, among them, perceived egoism, which is expected to be negatively
related with the trust concept proposed in this research. Egoism can be described as an
individual’s constant pursuit of his/her own self-interest or selfish goals. The sociologist and
ethnologist Emile Durkheim (1951) characterizes egoism as a form of excessive individualism
and a prepossession of a person that causes him/her to retrieve from the social community and
to pursue personal goals at the expense of collective wealth (Rahn & Transue, 1998). Egoism
seems to share some similarities with egotism or socialized narcissism (Schmalhausen, 1928)
and (ethical) egoism (a moral feeling negatively associated in western societies) is often
contrasted with (ethical) altruism or the felt obligation to help others (a moral feeling positively
associated in western societies). Prior research has demonstrated that the more individuals think
that norms of reciprocity are accepted as well as widespread in the society, the more they tend
to have trust in the others’ willingness to cooperate (Coleman, 2000; Putnam, 1993). Therefore,
the research at hand assumes that if online consumers view eWOM reviewers to be driven by
amoral egoism, purely ego-centric motivations and reckless opportunism by abusing reviews
and recommendations for personal objectives, the generalized trust in this information is likely
to be reduced (Volken, 2002). Thus,

H 6: A person’s trust in eWOM is significantly and negatively related to perceived

reviewer egoism.

144
Tie Strength

It is likely that the perceived social structure in eWOM communications also has an impact on
eWOM trust formation. It is clear that all WOM information flows between the reviewer and
the reader take place within a social relationship (Brown et al., 2007). Earlier research indicates
that investigating the social linkage between the pairs of communication is elementary for
understanding the process of WOM (Brown & Reingen, 1987; Knoke & Kuklinski, 1982).
While there are different approaches to categorizing the different types and qualities of social
relationships, a recognizable number of researchers have devoted their attention to the social
closeness of the relationship between the information recipient and the source mirrored in the
concept of tie strength (e.g., Brown & Reingen, 1987; Burgess et al., 2011; Mittal et al., 2008;
Money et al., 1998). Tie strength refers to “the potency of the bond between members of a
network” (Mittal et al., 2008, p. 196). Similarly, Money et al. (1998, p. 79) define tie strength
as “a multi-dimensional construct that represents the strength of the dyadic interpersonal
relationships in the context of social networks”. The strength of a tie is judged based on
elements such as: closeness, intimacy, support and association (Frenzen & Davis, 1990). At its
heart, tie strength describes the type and quality of the relationship between persons. According
to Granovetter (1973), social ties should be classified as either strong or weak. Strong ties –
family members, friends, spouse etc. – are characterized by a stronger and closer relationship.
Members of this group belong to an individual’s personal network and are able to provide
substantive as well as emotional support (Chu & Kim, 2011; Pigg & Crank, 2004). According
to Walker et al. (1994, p. 57), strong ties are characterized by “(a) a sense that the relationship
is intimate and special, with a voluntary investment in the tie and a desire for companionship
with the partner; (b) an interest in frequent interactions in multiple contexts; and (c) a sense of
mutuality of the relationship, with the partner’s needs known and supported”. Persons who are
in a strong tie relationship are more likely to interact more frequently and also exchange more
information when contrasted to individuals with a weak tie relationship (Brown & Reingen,
1987). On the other hand, weak ties represent less personal and intimate social relationships.
Individuals typically have weak ties with a wide set of acquaintances, colleagues (Chu & Kim,
2011) or with persons who are not personally known by the individual.

Duhan et al. (1997) (among others) have found that perceptions of tie strength affect
information flows and the selection of sources sought for WOM communication. For instance,
it turned out that most WOM comes from those people with whom the consumer typically has
strong ties. In general, stronger ties have a greater influence on the receiver’s behaviour than
weaker ties. Researchers have concluded that the same influence is also evident in purchasing
situations (Granovetter, 1973). This is due to the frequency and perceived importance of the
social contact among strong-tie individuals (Bansal & Voyer, 2000; Browne et al., 2007) as
well as due to the level of trust. Strong ties are said to evaluate products on the basis that the
consumer prefers, due to their understanding of the consumer (Burgess et al., 2011).

145
Nonetheless, in the eWOM context, consumers often choose to interact with unfamiliar entities
or strangers with whom they do not have a meaningful prior relationship (i.e., weak ties). Here,
consumer are forced to make trust inferences based on generalizations and contextual
information or cues (Smith et al., 2005). The consumer’s general attitude towards social
closeness between the sender and the consumer is likely to represent such a generalization. It
can used by the information-seeker to generally judge the trustworthiness of the comments and
reviews in the Internet. Hence, it seems to be reasonable to argue that the consumers’ trust level
can be expected to covariate with the perceived bonds these persons have developed with their
fellow reviewers. A positive link between tie strength and trust is supported by earlier
contributions by Yu and Tang (2010) and Pan and Chiou (2011).

H 7: A person’s trust in eWOM is significantly and positively related to perceived tie

strength between the reader and reviewers.

Homophily

Related but conceptually distinct to tie strength is the concept of homophily (Brown et al., 2007;
Brown & Reingen, 1987). Homophily refers to the extent to which individuals who interact
with one another are congruent or similar in certain attributes (Rogers & Bhowmik, 1970).
Numerous studies agree that friends but also members of a social network tend to be similar in
respect to socio-demographic characteristics, including age, gender, education, as well as other
perceptual attributes like beliefs and attitudes (Chu & Kim, 2011; Festinger, 1957; Gilly et al.,
1998). Accordingly, Schacter (1959) notes that humans tend to affiliate with others who share
similar interests (e.g., hobby) or who are in a similar situation (e.g., product choice). As
individuals tend to socialize with persons with whom they share similar characteristics,
interpersonal communication is more likely to occur between individuals who are alike
(Lazarsfeld & Merton, 1954). It is a law-like situation that most information exchange is
between persons who share some common qualities (Rogers & Bhowmik, 1970). Therefore, it
is likely that the concept of homophily plays a major role in the flow of information when
consumers undertake external searches (Price & Feick, 1984).

The similarity of individuals predisposes them toward a greater level of interpersonal attraction,
understanding, and trust than would be expected among dissimilar individuals (Brown et al.,
2007; Ruef et al., 2003). There seems to exist a profound relationship between homophily and
trust, or as Eccleston and Griseri (2008, p. 575) simply put it: “people trust other people,
especially those `like themselves`”. Homophily represents a reasonable trust-building
mechanism where trust is formed on common characteristics the truster perceives of the trustee
(Lu et al., 2010). To date, a considerable amount of (offline) trust literature provides various
hints concerning a positive relationship between similarity and a person’s willingness to trust
his/her interaction partner. For instance, Crosby et al. (1990) indicate in their research that
similarity between individuals enhances relationship quality within that group. Trust
146
researchers agree that shared attitudes as well as similarity in demographic backgrounds can be
considered critical factors that affect perceptions of trustee’s trustworthiness and stimulate
interpersonal information exchange (Brock, 1965; Feick & Higie, 1992). Here, it is
demonstrated that the influence of a similar source may be greater than that of an expert source
when certain conditions are met (Gilly et al., 1998) – which may be attributable to trust matters.
Simultaneously, some of the best-known literature contributions in the field of marketing
advance that shared values, attitudes and beliefs, behaviours, goals, and policies all impact trust
formation (Dwyer & Oh, 1987; Morgan & Hunt, 1994). Doney and Cannon (1997) conclude
that the buying firm’s trust in a seller’s salesperson is highest when the salesperson and the staff
in the buying firm share common values and interests.

Recent studies have demonstrated that the same rules apply to the online context, as homophily
has been found to be a factor responsible for the formation of trust in reviews. For instance, in
online book stores, consumers are more inclined to adopt recommendations from those
shoppers who appear to be more similar to them (Ziegler & Goldbeck, 2007). Racherla et al.
(2008) also have found evidence that reviews with a high perceived similarity between the
reviewer and the consumer have higher trust scores than reviews with low perceived similarity.
Research has noticed that consumers are inveigled to show identity-granting behaviour when
reviewers on [Link] have similar demographic characteristics. People with similar
characteristics also group in the same online community, such as MySpace (Thelwall, 2009).
Here, these people tend to perceive each other in a favourable manner, which enhances their
trusting beliefs (Lu et al., 2010; McKnight et al., 1998). Perceived similarity between the sender
and the consumer serves as an indicator for the latter that the opinions and concerns conveyed
in the message comply with his/her own needs. This reinforces the extent of trust in the
information. Thus:

H 8: A person’s trust in eWOM is significantly and positively related to perceived

homophily of the reviewer.

Susceptibility to Interpersonal Influence

Interpersonal influence has long been recognized as a critical factor shaping consumer shopping
choices (e.g.g., Ajzen & Fishbein, 1980; Kiel & Layton, 1981). Reference groups have the
potential to exert social influence in different ways. First, they often provide valuable
information in ambiguous, inconvenient situations (i.e., informational influence). Informational
social influence has been conceptualized as the individual’s tendency to accept information
from others as evidence about reality (Deutsch & Gerard, 1955). Here, social influence occurs
when social others provide information that enables another’s understanding of some
phenomenon (Mangleburg et al., 2004) or increases another’s knowledge about some aspects
of the environment (Bearden et al., 1989). Second, they set normative standards and a frame of
reference and/or enhance a person’s self-image (i.e., normative influence). According to
147
Deutsch and Gerard (1955, p. 629), normative social influence can be defined as “an influence
to comply with the positive expectations of another”. Other authors provide similar definitions
(Burnkrant & Cousineau, 1975).

Park and Lessig (1977) state that in the case of informational influence, shoppers may look for
hints from other consumers whom they regard as knowledgeable or they make up their mind
by observing the behaviour of others (Bailey, 2005). Prior research provides evidence that
informational influence has a reasonable impact on consumer behaviour (LaTour & Manrai,
1989; Lord et al., 2001). By drawing from prior research (Kelman, 1961), marketing scholars
have commonly distinguished two types of normative influence, as it has been viewed as either
value-expressive or utilitarian influence (e.g., Bearden & Etzel, 1982; Bearden et al., 1990;
Childers & Rao, 1992). In general, value-expressive influence is present when another person
has relevance for the consumer’s self-concept (Mangleburg et al., 2004). Specifically, value
expressiveness arises from the latent need of an individual to improve his/her self-image and is
characterized by the individual’s want for an association with the reference group in terms of
similarity and feelings towards the referent. Through the same process of identification that
occurs when individuals tend to associate themselves with favourably evaluated groups,
individuals also attempt to distance themselves from unfavourably evaluated groups (Bearden
& Etzel, 1982; Kelman, 1961; Park & Lessig, 1977). The second type of normative influence
is utilitarian influence, which basically reflects a person’s attempts to conform to the
expectations of others in order to gain rewards or to avoid punishments mediated by the others
(Bearden et al., 1990). These three forms of social influence (i.e., informational, value-
expressive, utilitarian) have been regularly applied to consumer research approaches (Ford &
Ellis, 1980; Witt & Bruce, 1972). Additionally, Bearden et al. (1990) developed their consumer
susceptibility to interpersonal influence scale (SUSCEP) on these manifestations in order to
measure the extent to which a person’s consumption choices are influenced by other individuals
(Kropp et al., 2005).

When consumers act in a social environment, consumer science provides profound evidence
that research approaches should consider someone’s base level of social influenceability, which
means how receptive one is to opinions and experiences of others (Aronson, 1985; Valck et al.,
2009). According to Mangleburg et al. (2004), susceptibility to interpersonal influence means
that others’ opinions, claims and evaluations are important. Kropp et al. (2005) argue that this
influence is performed by real or imagined others. In general, consumer susceptibility to
interpersonal influence (CSII) is viewed as a personality trait that is present in different
individuals in varying degrees. Due to Bearden et al. (1990), the construct is drawn from
McGuire’s (1968) contribution on the concept of influenceability and is consistent with the fact
that individuals differ in their response to social influence (Allen, 1965; Cox & Bauer, 1964,
Kelman, 1961). In a nutshell, a consumer’s CSII is likely to determine (to a large extent) how
the individual uses and values online reviewers and recommenders as a reference group (Valck

148
et al., 2009), as well as how trustful the individual perceives the information conveyed by
eWOM. Prior research has shown that WOM is most powerful when consumers are susceptible
to interpersonal influence (Bearden et al. 1989). For instance, Yang and Chou (2000) conclude
that CSII has a positive relationship with a consumer’s affection with WOM. It is documented
that highly susceptible persons are more influenced by others when making purchase decisions
(Schröder, 1996). Another research approach by Boush et al. (1994) finds that CSII is
negatively associated with advertising scepticism, which means that the more susceptible to
interpersonal influence, the less sceptical is the consumer about advertising claims. Similar
patterns are expected to exist online. In fact, consumers who shop online and use online
recommendations and reviews of other consumers find themselves in a social situation. Here,
eWOM provides information about marketing phenomena and/or serves to support positively
valued group norms facilitating direct social comparisons (Mangleburg et al., 2004). Therefore,
it is very likely that eWOM is capable of eliciting informative as well as normative social
influence on the shopper. More specifically, in this thesis it is predicted that consumers who are
more susceptible to informational influence than others place higher trust in eWOM, as they
have a higher need to collect information and guidance from fellow shoppers. These persons
are likely to place more weight on and rely more on this kind of information than consumers
that exhibit limited susceptibility to influence (Park et al., 2011). On the other hand, consumers
who are susceptible to normative influence are more likely to wish to comply with the
expectations of other shoppers and seek social approval in the form of others’ opinions and
recommendations. Consequently, they are more willing to trust any information from
significant others (Chu & Kim, 2011). Hence, it seems to be reasonable to argue that consumer
susceptibility to interpersonal influence will affect their willingness to rely on eWOM in
general:

H 9: A person’s trust in eWOM is significantly and positively related to susceptibility to

interpersonal influence.

Attention towards Social Comparison Information (ATSCI)

A dedicated string of psychological literature deals with the investigation of the conditions
under which individuals feel a need for comparison-based information. As a person’s interest
in comparison with others has been found to be promoted by perceived uncertainty and risk
(Festinger, 1965; Taylor et al., 1990), situations which are associated with novelty, complexity
or change are most likely to stimulate the amount of comparison orientation. Also, most
individuals tend to regard social comparison as important in situations that are characterized by
competition (Ruble & Frey, 1991). Purchasing products online is a situation comparable to
these conditions, as the task of satisfying consumption needs with a specific product is often
new, and deciding on a specific product often leads to changes in life. Additionally, society
tends to value smart consumers. This implies that shoppers are in a continuing competition in

149
finding the best product and demonstrating their capabilities as a skilled consumer. Online
reviews and recommendations can here be a valuable source for comparison information. A
notable number of scholars suggest that certain individuals are more inclined to get involved in
(and rely on) social comparison than others (e.g., Bearden & Rose, 1990; Gibbons & Buunk,
1999; Steil & Hay 1997). Here, the same conditions representing situational reasons for
heightened social comparison are theorized to be interrelated due to an inner predisposition.
According to this literature string, individuals who perceive uncertainty about themselves are
especially induced to consult social comparison information. It is said that persons with specific
personality characteristics such as low self-esteem (Waymnet & Taylor, 1995) or a high
disposition to neuroticism (Lennox & Wolfe, 1984) are generally more interested. The same
individual difference variables have been theorized to be related with the eWOM trust construct
earlier.

Broadly, attention-towards-social-comparison-information (ATSCI) addresses the individual’s

generalized tendency to conform, and it has been demonstrated that it is related with the concern
towards others’ reactions (Lennox & Wolfe, 1984). Individuals who possess a high level of
ATSCI are said to be well aware that their behaviour will elicit reactions in their social
surrounding and they are very concerned about those reactions. As a consequence, these people
constantly look for cues that give them insight into those reactions and furnish them with
information about what conformative behaviour should look like. Formally, in the context of
consumer behaviour, the construct is here defined as the extent to which individuals are
sensitive to social comparison cues relevant to their product choices (Bearden & Rose, 1990).
Persons who score high on ATSCI have the tendency to collect social cues from multiple
sources of market communication (Bearden & Rose, 1990), including statements of the
adequateness of consuming certain products advanced by important referent and aspiration
groups (Miniard & Joel, 1983) or others’ behaviours. Similarly, advertising offers social
comparison information by featuring consumer referents who benefit from positive
reinforcement as a result of product consumption (Nord & Peter, 1980). Or, on the other side,
marketer-driven communication doesn’t shy away from demonstrating social punishment of
these referents as they fail to use the product. eWOM is a comparable informational cue as it
grants consumers the insight into how the community rewards or punishes the usage of market
offerings. Recommendations represent normative demands towards the consumer, which
products to consume and how wise consumers should behave. Persons who are sensitive to such
information not only use it but are also more likely to rely on it than others. Thus,

H 10: A person’s trust in eWOM is significantly and positively related to the attention
towards social comparison information.

150
eWOM Trust Correlates

eWOM Scepticism

Every time consumers read eWOM, they may easily wonder whether the reviewer or
recommender takes an advantage from the company which sources the product or service.
Additionally, consumers are typically uncertain whether the messages originate from a fellow
shopper or are the product of the company itself. In the latter case, they may consider reviews
to be camouflaged company advertising. Therefore, it is very likely that these consumers have
scepticism towards this market information, as they may not know how much they can believe
this information. This thesis has defined eWOM scepticism as the tendency to disbelieve the
information claims of eWOM earlier. This definition is consistent with, and is adopted from, the
work of Obermiller and Spangenberg (1998) on scepticism toward advertising. The authors
show that the construct is related to attitude towards advertising and attitude towards marketing,
but it is also determined by certain personality characteristics, consumer experiences and
socialization processes (Obermiller & Spangenberg, 1998; Obermiller et al., 2005). Hence, it
seems reasonable that consumers can develop scepticism toward a variety of objects, for
instance, toward advertising in general but also toward specific advertisements in particular
(Darley & Smith, 1993; Ford et al., 1990; Mangleburg & Bristol, 1998), or towards marketing
and company practises. For Mohr, Eroglu and Ellen (1998), scepticism is defined as a cognitive
response that varies with the context, situation but also content of the communication. It has
been shown that the presence of consumer scepticism creates negative attitudes toward
advertising and towards the motives of the advertisers (Andrews, 1989; Boush et al., 1994;
Forehand & Grier, 2003; mangleburg & Bristol, 1998). As consumers have been found to be
sceptical about advertisements (Obermiller & Spangenberg, 2000), they may be sceptical about
eWOM, which is generally assumed to represent review-conveyed advertisement. Disbelief in
eWOM claims is likely to arise when eWOM communication is generally not assumed to
represent genuine opinions of fellow shoppers. Once consumers have scepticism toward online
reviews, the persuasive influence as well as the trust in this kind of information is reduced or
diminishes completely. Hence, it is hypothesized that:

H 11: A person’s trust in eWOM is significantly and negatively related to eWOM

scepticism.

Distrust in Reviewers’ Motives

This thesis has discussed trust and distrust as two conceptually distinct and inversely related
constructs. For instance, McKnight et al. (2001, p. 40) state that “distrust is the distinct opposite
of trust”. In contrast to trust, operationalizations as well as empirical insights concerning
eWOM distrust and its relationships with other concepts (such as eWOM trust) are virtually
non-existent. This may be due to conceptual as well as measurement problems of this negative

151
trusting stance of the consumer. However, it seems reasonable that nearly the same rules apply
to consumers’ unfavourable attitudes towards customer reviews. It seems intuitive that
individuals are likely to possess a low level of trust in cases where they distrust the reviewers’
motives underlying eWOM communication. This relationship is already well-supported in the
offline trust literature (see earlier). Therefore, the following hypothesis is proposed:

H 12: A person’s trust in eWOM is significantly and negatively related to eWOM distrust.

eWOM Involvement

This thesis adopts the general view of involvement which focuses on personal relevance
(Greenwald & Leavitt, 1983; Rothschild, 1984; Zaichkowsky, 1985). For example, Hupfer and
Gardner (1971) define involvement as a general level of interest in or concern about a particular
issue. In reference to Zaichkowsky’s (1985, p. 342) seminal article, eWOM involvement is
defined as a person’s perceived relevance of online customer reviews based on his/her inherent
needs, values, and interests. Various authors have attempted to identify the correlates and
determinants of involvement (Betra & Ray, 1985; Goldsmith & Emmert, 1991). For instance,
Zaichkowsky (1986) states that whether an individual is involved or not is mainly determined
by the person’s values, along with the prior experiences the individual has had with the
particular object. Therefore, one can say that personal perceptions of the object are essential
(Soh, 2007). In adoption of this view, this thesis hypothesizes that the earlier perceptions of
eWOM which result in a person’s level of trust have an impact on eWOM involvement. This
notion is supported by prior research, as trust towards the trustee has been found to be positively
correlated with the relevance of this object to the truster (e.g., Luh et al., 2010; Tung et al.,
2001). Similar to the theoretical background of these contributions, it is assumed that
individuals for whom eWOM has personal relevance also place a higher level of trust in this
kind of information. eWOM involvement seems to be closely related to the concept of eWOM
ability. Hence, the following relationship is proposed:

H 13: A person’s trust in eWOM is significantly and positively related to eWOM

involvement.

eWOM Trust Consequences

As discussed earlier, trust is a good predictor of behaviours (Gefen et al., 2003; Menon et al.,
2002; Pavlou & Fygenson, 2006). Consequently, this thesis argues that a person’s general level
of trust in eWOM is likely to determine the level of exposure to this kind of information or,
more generally, multiple forms of eWOM usage and message impact. That is, eWOM trust is
theorized to determine a person’s amount of information seeking, purchase influence, and
information-passing to others. This assumption finds support in prior research that identified a
strong trust-behaviour relationship in the eWOM context. For instance, Chen and Hung (2010)

152
find that interpersonal trust in an online community is positively related with information-
collecting behaviour of the members. Additionally, by borrowing from the media dependency
theory, Park et al. (2011) argue that online shoppers who used online reviews are more often
and more willing to use and adopt online reviews, which can be regarded as a form of trusting
behaviour.

Simultaneously, there is a growing body of literature supporting a negative relationship between

attitude towards market communications and volitional exposure to this kind of information.
That is, a person who is striving to reduce cognitive dissonance and has minor trust in eWOM
will align his/her mental structures accordingly (i.e., hold a negative attitude, motivation
towards OCR) and will also be reluctant to use this information. For these people, eWOM
represents a burden or alienating thing. This is well supported in the advertising context (e.g.,
Abernethy, 1991; Cronin & Menelly, 1992). Here, it is generally agreed that consumers show
ad-avoidance behaviours (e.g., “zipping” TV commercials) when they have a negative attitude
towards advertising in general. Similarly well supported seems to be the inverted relationship
between trusting beliefs and ad avoidance (Soh, 2007). For instance, consumers are more likely
to spend more time looking at advertisements if they have positive perceptions of their
trustworthiness (James & Kover, 1992). The same seems to be true in the opposite direction
(Lee & Lumking, 1992), as it is assumed that the same behaviours apply to the eWOM context.
eWOM avoidance is defined as the consumer’s intentional behaviour to skip or bypass online
review-conveyed information. In accordance with above discussion, the following research
questions are asked:

H 14: A person’s trust in eWOM is significantly and positively related to eWOM

information usage (a: eWOM influence; b: eWOM seeking; c: eWOM passing).

H 15: A person’s trust in eWOM is significantly and negatively related to eWOM

avoidance.

Reliability and Generalizability

In general, reliability describes the consistency of a measure. That is, reliability is concerned
with that portion of measurement that is due to permanent effects that persist from sample to
sample (Netemeyer et al., 2003). The procedures regularly discussed in literature can be
grouped into two general types: (1) test-retest (temporal stability) – which targets the
correlation between the same person’s score on the same set of items at two points of time; and
(2) internal consistency – which reflects the interrelatedness among items or set of items in the
scale. Both issues are addressed in this thesis. While the latter is assessed through multiple
criteria (individual corrected item-to-total correlations, the average inter-item correlation
among scale items, and a number of reliability coefficients like Cronbach’s alpha and AVE)
(Churchill, 1979; Cronbach, 1951; MacKenzie et al., 2011; Nunnally & Berstein, 1994) along

153
the stages of the scale development process, the former is discussed at this point in greater
detail.

As mentioned earlier, test-retest reliability cares about the stability of a respondent’s items
responses over time (Netemeyer et al., 2003). Typically, a test-retest or “stability” coefficient
– which is the correlation between the same measures (and sample) on two different assessment
occasions – is used to assess the scale’s consistency. In cases where the magnitude of the
stability coefficient is low (with no change in the construct over time), then the reliability of the
measure is questionable. Assessing the test-retest reliability is valuable since it is able to provide
confidence that the measure reflects the construct and the scale is generalizable to other
assessment occasions or is stable over time (DeVellis, 1991; Haynes et al., 1999; Nunnally &
Berstein, 1994). Having conceptualized eWOM trust as a stable attitude, this is a critical issue.
Recognizing the importance of such an assessment, the following research question is proposed:

RQ 7: Is there a significant and considerable correlation between the measurements of

eWOM trust on two different assessment occasions?

Generalizability is another critical issue for scale development. Research ventures that strive
for a better understanding of this issue aim to clarify the extent to which a scale can be
generalized to other situations, conditions or slightly other contexts under which it is assumed
to be applicable (Netemeyer et al., 2003). The following three guiding questions could be raised
with respect to the developed eWOM trust measure. First, eWOM trust per se is defined as a
relatively stable global attitude concerning the reliance on customer reviews as an institution.
Is the scale then also applicable to other contexts like measuring generalized trust in specific
eWOM platforms? Second, eWOM trust in general is theorized to influence consumer attitudes
and reactions towards individual review messages to a certain degree. Hence, one may ask
whether such a relationship truly exists or not. Third, the new measure is developed in German-
speaking countries (Germany and Austria); therefore, a valid question is: does the scale
generalize and replicate across other cultures and languages? In the following, these critical
issues will be discussed.

Generalizability to different Online Sources

When doing online pre-purchase information search, consumers were earlier limited to
retrieving this information from online advertising, such as company-sponsored websites.
However, the marketing literature suggests that consumers may be concerned about the
usefulness of seller-created information, since it may be compromised by the vested interest the
seller may have when creating this kind of information (Burgess et al., 2011). The marketer,
for instance, can strive to make his/her product more appealing than other products, instead of
truly informing the consumer about the best alternatives. From the consumer perspective, this
potentially reduces information value and trust in the message. However, in addition to the

154
information available via online advertising, consumers today have access to diverse sources
of eWOM. These platforms include message boards, blogs, chat rooms, discussion forums on
social networking sites (like Facebook or MySpace), and especially reviews on merchant
websites (Brown et al., 2007; Dellarocas, 2003; Jepsen, 2007). But also well-known product
review websites such as Epinions, Consumer Search’s Productopia, [Link], and
[Link] enable shoppers to provide reviews of, and feedback on, a great selection of brands,
products and services available on the marketplace (Bailey, 2005). All these sources nowadays
represent valued forms to share and exchange opinions and experiences regarding companies,
products, and services (Pan & Chiou, 2011). Hence, purchase choices tend to be influenced by
multiple sources, as consumers simultaneously have access to various sources of information
(Horrigan, 2008).

As mentioned, OCR can be found on different platforms. In literature, different approaches

have been proposed to classify review sources (Bae & Lee, 2011; Bailey, 2005; Park e al.,
2012). In general, these classifications are made according to the parties that host eWOM
platforms. For instance, Bailey (2005) assumes that there are two different kinds of review
sources from which consumers can read online customer reviews: (1) company-operated and
(2) consumer-operated review sites. A general characteristic of company-operated review sites
is that consumers are able to read reviews, access accompanying product information, and
efficiently compare products from alternative manufacturers. These sites typically use review
templates, which results in a more or less standardized structure of the content. An example of
such standardization is the evaluation of the product with a star rating that enables consumers
to quickly know whether the product is recommended or not. Seller websites such as
[Link] or [Link] can be regarded as representing typical examples of company-
developed review sites.

Bae and Lee (2011) referred to company-operated review sites as marketer-developed review
sites (MDS) by broadening this class. According to them, marketer-developed review sites do
not only include company-operated websites that encourage their customers to write reviews
about the products which are sold on the same site, but also review sites operated by
independent agents, which essentially direct shoppers to commercial sites in order to purchase
the reviewed product. Hence, marketer-developed review sites also include well-known review
sites such as [Link], [Link], [Link], [Link] and [Link]. The
authors argue that such “independent” websites claim to offer unbiased information. However,
all marketer-developed review sites have come to the market in order to strive for commercial
success derived, for instance, from advertising revenues and generating demand. Hence, the
provided information may tend to mirror vested interest (Bae & Lee, 2011).

The second class of review sites are consumer-operated or, by following the notion of Bae and
Lee (2011), consumer-developed review sites (CDS). Examples are specific online

155
communities and blogs. According to Rheingold (1993) and Ridings et al. (2002), an online
community (e.g., online discussion forum, bulletin board) is made up of consumers who have
similar interests and want to develop social relationships by sharing opinions and ideas with
other shoppers. In contrast, a blog can be defined as “quickly and frequently updated web pages
with a reverse chronological sequence of date entries” (Bae & Lee, 2011; Kumar et al., 2004;
Nardi et al., 2004). Here, the primary motivation of consumers to join such communities is to
form social relationships and bonds with fellow others (Brown et al., 2007; Kozinets, 2002).
Participants tend to develop empathy toward each other and have a latent need to help each
other. Reviews are written by individuals who want to exchange their ideas and opinions with
other often similar-minded persons. Contributing to the stream of reviews available in online
communities and blogs is not driven by rewards or monetary incentives, but by the will to voice
their impressions and consumption experiences through the social network. In contrast,
consumers often feel encouraged to contribute to independent agent sites as they are offered a
monetary compensation or similar reward (Bailey, 2005). Consumers tend to repeatedly visit
online communities and blogs, leading to frequent communication among members and
fostering the development of personal relationships and strong ties (Brown et al., 2007). Such
experiences are a fundamental basis of trust development.

Scholars agree that the forum in which information is presented is quite important (Andreoli &
Worchel, 1978; Burgess et al., 2011; Harkins & Petty, 1987; Kelman, 1961). For instance, Bei
et al. (2004) note that shoppers typically tend to perceive product information quite differently
across information sources. In this research tradition, Senecal and Nantel (2004) conclude that
websites that are seller-independent tend to be preferred by consumers compared to those that
are clearly associated with a business with a vested interest. A review of current research
literature supports the notion that consumer-created information is more trusted and preferred
to seller-created information that may include strong manipulative intent. For instance,
consumers deem information that can be found in the form of eWOM more trustworthy than
information provided by the sellers such as web ads (Cheong & Morrison, 2008). The same
holds true for the OCR platform; recommendations that originate from seller-independent
websites influence online shoppers more than those posted on seller-dependent homepages
(Alba et al., 1997; Bakos, 1997; Lynch & Ariely, 2000). Similarly, consumers perceive eWOM
in online forums that are independent from the sellers as more persuasive than seller-created
product information from corporate websites (Bickart & Schindler, 2001). People may assume
that the independence of CDS allows the published information to be more objective and
credible (Bray & Schetzina, 2006; Gretzel, 2006; Park et al., 2007). Consumers have been
shown to draw differences among the consumer-developed sources as well. For Cheong and
Morrison (2008), consumers are more likely to rely on online communities (i.e., discussion
boards) than on blogs. Hence, there seems to be strong support that CDS are more trusted by
consumers than MDS. However, there are also contradicting findings. A recent Jupiter Research
study showed that only 21% of shoppers trust information provided on social networking sites,
156
while information on corporate websites is evaluated as far more trustworthy (Wassermann,
2006). Bronner and de Hoog (2010) conclude that trust (in terms of reliability) is the same for
both marketer sites and consumer sites, confirming the notion that both types of information
sources are complementary.

This thesis assumes that eWOM trust is generalized across different sources. Nevertheless, it is
also recognized that the source (or platform) on which reviews and recommendations are
readable to the consumer may exert a contextual influence on trust towards reviews or reviews
in general available on such websites – that is, generates specific source-based eWOM trust. In
line with the comparable work of Obermiller and Spangenberg (1998), this thesis proposes that
eWOM trust is a general tendency that blurs perceptions of customer reviews. That is, highly
trustful consumers may be more or less trustful of reviews in company-developed versus
consumer-developed websites, but they should be more trustful of both sources than are their
less-trustful counterparts. Having said that, the above discussion suggests that consumers may
hold different attitudes toward and varying levels of trust in eWOM, depending on the type of
source in which it is presented. Hence, in cases where the new scale is able to validly measure
consumer trust in eWOM across different forms of eWOM sources (MDS vs. CDS), the
developed measure would be generalizable from a measure that assesses eWOM trust in general
to a measure that is also applicable to eWOM trust perceptions in multiple online sources.
Therefore, in analogy with earlier contributions (e.g., Soh, 2007) and in extension to the above
cited literature, the following multi-part research question is asked:

RQ 8: (a) Can the developed measurement of trust in eWOM reliably and validly measure
trust in specific eWOM platforms? (b) Does trust in eWOM vary across different
platforms? (c) On which platform is eWOM most trusted?

If meaningful differences can be identified between eWOM sources, it is reasonable to ask

which factors can be held responsible for such discrepancies. The concept of source (or media)
credibility is presumably one of the most influential correlates to eWOM trust. Credibility
theory proposes that when a source exhibits itself as credible, individuals are more trustful and
more probably persuaded (Anderson, 1971; Hovland et al., 1953; Ohanian, 1991; Petty &
Cacioppo, 1986). A source’s credibility is especially important online, as information on the
source serves as a peripheral cue used in heuristics to judge information trustworthiness
situationally (Dou et al., 2012; Petty & Cacioppo, 1981). Considerable literature has
demonstrated a positive relationship between source credibility and trusting behaviours (e.g.,
message adoption) (Ohanian, 1990; Park et al., 2011), which regularly draws on the
elaboration-likelihood model (ELM) (Petty & Cacioppo, 1986).

Consumers typically exchange eWOM information with a vast and geographically dispersed
group of strangers by means of blogs, online consumer review platforms, retailer websites,
social networking sites, and similar websites enabling UGC (Cheung & Thadani, 2012; Davis
157
& Khazanchi, 2008). This typically raises issues concerning the credibility of the reviews’
sources (Forman et al., 2008). In eWOM, the source’s credibility may be an important surrogate
to infer the credentials of the reviewer and subsequently the trustworthiness of the information.
If the consumer perceives the eWOM source as credible, he/she will place more trust in those
reviews and recommendations on this kind of review site, which brings him/her to use these
reviews and learn from them. In contrast, if the eWOM source is perceived as less credible, the
effect of the review will diminish and scepticism increases. Prior studies have shown that source
credibility tends to vary in differently trusted information sources. Based on these insights, this
thesis assumes that the eWOM source’s credibility relates to eWOM trust as an antecedent and
therefore asks following research question:

RQ 9: Are consumer perceptions of (a) eWOM platforms’ credibility as well as (b) the
credibility of reviewers’ on these platforms associated with trust in eWOM on that
platform?

Generalizability to individual Customer Reviews

One of the claims of advertising research is that a consumer’s attitude towards advertising in
general influences his/her attitude towards specific ads. An early but nevertheless very profound
contribution has here been made by Lutz (1985), who identifies five antecedents of attitude
toward the ad: (1) ad credibility, (2) ad perceptions, (3) attitude toward the advertiser, (4)
attitude toward advertising in general, and (5) mood (Durvasula et al., 1993). This
conceptualization also finds empirical support. For instance, Donthu et al. (1993) suggest that
consumers who have a positive attitude towards advertising are more likely to recall more
advertisements than persons with a more negative general attitude. In a similar manner, James
and Kover (1992) find out that persons who are willing to spend more time looking at print
advertisements are also characterized by a positive general attitude. A similar result is provided
by Mehta (2000). He further advances that the more consumers perceive advertising as truthful
or not manipulative, the more likely they are persuaded by individual ads. Simultaneously, there
exists further empirical evidence for a positive correlation between perceptions of general
advertising credibility and perceptions of the credibility of individual advertisements
(MacKenzie & Lutz, 1989). This thesis has conceptualized attitude towards reviews in general
(RAtt) by borrowing from above-cited literature and it seems reasonable that the same
relationships hold true in the eWOM context. RAtt has been defined as conceptually distinct
from eWOM trust, such as measured review credibility in general (RCred). However, both
concepts represent critical elements of eWOM trust such that one should expect some
relationship between eWOM trust in general and consumer attitudes towards individual
customer reviews. However, it is not assumed that eWOM trust as a generalized attitude will
be able to make perfect predictions about the perceptions of individual reviews. Hence, it is
possible that an eWOM high-truster may not believe every claim made in reviews. Similarly, it

158
is also very unlikely that low-trusters necessarily disbelieve all review claims. Rather, the
concept describes tendencies that affect reactions. While it is recognized that situational factors
(e.g., message characteristics) are likely to play an important role for the acceptance of specific
messages (Obermiller & Spangenberg, 1998; Soh, 2007) or reviews, it is assumed that
consumers with heightened levels of eWOM trust (in general) are (a) more willing to rely on
(trust) specific customer reviews and also (b) exhibit more positive attitudes towards them
compared to individuals scoring low on this personal difference factor. In order to clarify this
relationship, the following research questions are addressed:

RQ 10: Is eWOM trust in general significantly and positively related to eWOM trust in
individual customer reviews?

RQ 11: Is eWOM trust in general significantly and positively related to attitude towards
individual customer reviews?

Generalizability to different Languages and Cultures

Another crucial criterion for evaluating the new eWOM trust measure is its generalizability
across languages and cultures. This study primarily uses samples from Austria and Germany
for developing and validating the original measurement instrument. However, the proposed
measure may be inapplicable to other cultures, unless evidence for cross-cultural
psychometrical properties (i.e., dimensionality and reliability) of the scale is provided
(Durvasula et al., 1993). Hence, this thesis investigates the cross-cultural applicability of the
eWOM trust scale in North America by using an English version of the set of items. By
addressing this aspect, the research approach at hand answers calls from the scientific
community that emphasize the need for cross-cultural validation of new consumer attitude
scales (Netemeyer et al., 2003). These scholars highly recommend validating an instrument
across different populations and languages. The thesis agrees with John and Srivastava (1999)
that if tasks most central to human survival are universal, the most important individual
differences or enduring orientations (e.g., trust), and the terms people use to describe these,
would be universal as well. This argumentation favors the existence of measurement scales that
exhibit factor structures that are stable across cultures. In this respect, eWOM trust is theorized
to have the same meaning across cultures. However, this study also recognizes that culturally
specific dimensions and variations due to social context can potentially exist, as demonstrated
in other research contexts. Therefore, the scale’s congruence shall be further investigated:

RQ 12: Is the eWOM trust scale applicable to different languages and cultures?

159
Measure Application

Making product purchase decisions is a risky endeavor. Therefore, in order to handle

perceptions of deterrent risks, consumers typically search for information (Bettman, 1973;
Srinivasan & Ratchford, 1991). Here, they are not limited to internal information gathered
through prior consumption experiences and mirrored by existing knowledge, but consumers are
eager for external information, especially when the purchasing task is new and not a routine.
While, in earlier times, access to product information was limited to particular sources (e.g.,
newspapers, television), nowadays the rapid development of the Internet grants consumers a
new opportunity to find such information in a previously unexpected depth (Jepsen, 2007).
Actual market reports, as well as earlier research studies, demonstrate that consumers have been
actively using the Internet when looking for purchase-relevant information for more than ten
years (Fallows, 2005; Peterson & Merino, 2003). Moreover, the Internet seems to increase its
importance in the eye of the consumer from year to year, as we are entering a new era of
consumer interaction and cooperation. Research has shown that certain characteristics of the
Internet as an information channel have modified consumers’ information search patterns, as
well as the outcomes of the purchase decision process in general. The medium enables
consumers to fall back on individualized information with minimal effort and cost, whenever
and wherever they want (Peterson, 1997). Further, the opportunity to get access to information
that was previously inaccessible to the shopper community (Kraut et al., 1998) is another critical
aspect of the Internet that helps consumers to increase their decision-making efficiency (Alba
et al., 1997; Widing & Talarzyk, 1993) – and to potentially become “smarter shoppers”.

However, the difficulty of selecting adequate information out of an apparent endless amount of
available online information is an obstacle for improved decision making. Consumers are not
in the possession of systematic filters that help them to access the quality of online information
at face value (Eysenbach & Diepgen, 1999). Rather, they have to evaluate different kinds of
information sources based on generalized attitudes and personal dispositions. Here, a pivotal
concept is the matter of trust in information. It is the tenor of this thesis that trust represents an
important psychological construct that reflects consumers’ dependence on the trusting object
and triggers its personal relevance. Consequently, trust has a critical role in evaluating
information quality, as well as subsequent market information usage (Park et al., 2009).

While information on the Internet is manifold, actual studies demonstrate that consumers
typically source information from two major origins: (1) consumer-to-consumer
communication (i.e., C2C, online customer reviews) (see Chapter 1), and (2) business-to-
consumers communication (i.e., B2C, online advertising). Within recent years, online
advertising has lost its static character. Nowadays simple banner advertising and websites are
still instruments of the online marketing portfolio; however, the development also led to more
innovative advertising models, as well as online channels, including on-demand advertising

160
with ads customized to search engine enquiries and personalized offers via newsletters. Online
advertising has expanded to a diverse field, as can be seen from its definition. In this thesis,
online advertising is defined as any form of corporate market communication via the Internet.
Therefore, it includes corporate websites, banner ads, video clips, but also e-mail messages or
interactive games with corporate sponsorship (Schlosser et al., 1999). All these approaches stem
from the new opportunities offered by the Internet to facilitate interaction between the consumer
and the marketer (Breuer & Brettel, 2012; Rappaport, 2007). The industry’s expectations about
the positive outcomes of online advertising, such as increased customer influenceability, led to
continuous increases in the share of marketing expenditures. While in 2003 Internet advertising
made up $12.6 billion, worldwide total expenditures had already reached $120.4 billion in 2012,
and is further predicted to increase to $132.4 billion by 2015 (ZenithOptimedia, 2012). Hence,
online advertising is expected to account for more than 23% of all global ad expenditures.

In awareness of this global trend, advertising effectiveness as well as accountability are both
topics of continuous interest among marketing researchers and practitioners (Clark, 1999;
McDonald, 2010; Rust et al., 2004). Hence, for various marketers, the analysis of advertising
channel effectiveness and tied consumer behaviour (e.g., ad persuasiveness, impact) are
regarded as the key areas of concern of the professional marketing community (Brettel &
Spilker-Attig, 2010). As a consequence, numerous studies have attempted to assess the
determinants of advertising effectiveness by demonstrating which advertisements or channels
are influential, when, under which circumstances, and for which kind of consumers
(Shamdasani et al., 2001; Tellis et al., 2005). A key insight of this research is that advertising
achieves its highest impact when it is target-group specific and differentiates between various
consumer groups (e.g., in terms of the selected advertising channel) (Iyer et al., 2005; Reutterer
et al., 2006; Zeithaml et al., 2001). It is the guiding tenor of contemporary marketing endeavors
that the application of marketing instruments should depend on the kind of relationship between
the company and its customers (e.g., Bolton et al., 2004; Homburg et al., 2008; Verhoef &
Donkers, 2005). Therefore, it is important to profoundly understand which information sources
specific groups of consumers place their emphasis on/trust in. Or, in other words, as consumers
are likely to trust some information sources more than others, it is crucial to understand which
kinds of trusted sources are most likely to influence online as well as offline purchase decisions
(Cheema & Papatla, 2010). This is also well supported by research, which regards the presence
or absence of trust as key to communication effectiveness (e.g., Cheema, 2008; Gauzente, 2010;
Urban et al., 2000). A deeper investigation into this sphere enables marketers to figure out how
to maximize outcomes of online strategies, for example by better allocating the billions of
advertising dollars, and to anticipate the consequences of consumer information gathering – be
it from marketer- (B2C communication) or consumer-provided information (C2C
communication) sources. The identification of the most responsive/trusting market segments is
hence key.

161
By adopting the definition provided by Soh (2007), trust in online advertising (AdTrust) is defined
as “a consumer’s confidence that [online] advertising is a reliable source of product/service
information and one’s willingness to act on the basis of information conveyed by [online]
advertising” (p. 29). In order to derive a meaningful grouping of online communication
recipients, the present thesis addresses the following question:

RQ 13: (a) Can online consumers be meaningfully segmented according to their trust in
online customer reviews (eWOMTrust) and online advertising (AdTrust)? (b) How can
the segments be profiled?

Answers to this question should fill the still-existing research gap concerning the lack of
knowledge about consumer beliefs and attitudes towards online market information. More
precisely, this research should help to gain a better insight into observable differences in
information-seeking behaviours related to online advertising, as well as online customer
reviews and attitudinal and socio-demographic covariables. Amongst other considerations, this
research sheds light on the question of who accesses the large arsenal of unofficial product
information and who is likely to be influenced by positive or negative information available on
the Internet. But it also provides insight into the characteristics of consumer groupings which
are most responsive to the company’s online advertising arsenal, thus furnishing the company
with the opportunity to control at least some of the product communication sent to the
consumer. Therefore, this research provides answers to several still-unanswered calls of the
research community on the role of different information sources for purchasing decisions (e.g.,
Bucklin et al., 2002; Cheema & Papatla, 2010; Ratchford et al., 2003; Ward & Ostrom. 2003).

In order to evaluate the above-mentioned series of research questions and hypotheses, a multi-
stage/multi-study research approach is proposed. The reader will find the details concerning the
research process and the methods applied in each stage and study in the next chapter, before the
research results are presented in Chapter 5.

162
Chapter 4

Research Methods

In this chapter, a research process is presented which synergizes well-acknowledged procedures

for scale development, as well as a route to apply the new eWOM trust scale (eWT-S) to a
current market problem. A review of relevant literature grants the insight that various scholars
in the field of social sciences emphasize the issue of how to develop a quantitative, multi-item
measure and how to validate it (e.g., Anderson & Gerbing, 1988; Bagozzi, 1998; DeVellis,
1991; Edwards, 2001; Netemeyer et al., 2003; Nunnally & Berstein, 1994; Straub et al., 2004).
For instance, one of the most-cited and influential authors in the field of attitude measurement
is Churchill (1979), who proposes a sequence of eight steps that shall be followed when
developing measures for marketing constructs: (1) specify the domain of the construct; (2)
generate a sample of items; (3) collect empirical data; (4) purify the measure; (5) collect new
data; (6) verify reliability; (7) assess validity; and (8) develop norms. Other scholars have
chosen to advocated slightly different steps and procedures (e.g., DeVellis, 2012; Hildebrandt
& Temme, 2006; MacKenzie et al., 2011). However, the majority of these contributions
typically share a common set of guidelines for the development process and analysis
procedures. Figure 4 shows a diagram of the conducted research process, which begins with the
conceptualization of the construct and culminates with the scale’s application as a segmentation
tool of online information recipients. Due to this research’s various targets, a multi-stage/multi-
study research design has been employed to test the scale’s psychometric properties.
Specifically, the research has been executed in four sequential stages: (1) the identification
stage; (2) the reliability stage; (3) the validity stage; and (4) the application stage (details are
provided bellow). The methods and procedures used closely follow those described in the
current and established psychometric literature (e.g., Bagozzi et al., 1991; Cortina, 1993;
DeVellis, 2012; Robinson et al., 1991; Soh, 2007) and especially mirror the general
recommendations put forward by Netemeyer et al. (2003), as well as the more recent work of
MacKenzie et al. (2011). Both furnish the academic community with valuable frameworks and
suggestions for constructing valid self-report measures of latent social-psychological
163
© Springer Fachmedien Wiesbaden 2017
W. Weitzl, Measuring Electronic Word-of-Mouth Effectiveness,
DOI 10.1007/978-3-658-15889-7_4
constructs. As suggested by these scholars, the developed eWOM trust scale was subject to
rigorous reliability and validity testing using multiple consumer samples from various cultures.

The stages can be outlined as follows. The identification (or conceptualization) stage includes
two fundamental steps. In the first step, the importance of theory in scale development was
recognized by a thorough review of the interdisciplinary trust research literature. This enabled
the proposal of a literature-based theory in which the construct of eWOM trust is embedded.
Here, several key issues have been addressed: (i) the critical role of a clear construct definition
and an explicit specification of the content domain; and (ii) based on these insights, an a-priori
determination of the construct’s structure and dimensionality has been proposed. The findings
of this step were presented in Chapters 2 and 3. The second step in the identification stage
involved generating and judging a pool of observable items, from which scales of the sub-
dimensions were derived. In accordance with Netemeyer et al. (2003), during this step (i) the
theoretical assumptions about the items (e.g., domain sampling) have been considered; (ii) a
large set of potential items, theorized to capture the domain of the construct as indicators of
trust in eWOM, was created based on a review of prior literature, dictionaries and thesauri as
well as interviews with consumers and experts; (iii) the response format was determined; (iv)
marketing experts and members of the relevant population judged the items with a focus on
both content and face validity; and (v) the items for the subsequent stage were finally identified
– all possessing translation validity.

In the following reliability stage, the construct’s composition has been empirically tested with
a reasonable number of respondents. For this purpose a study among university students was
implemented and the received data were split in half for subsequent analysis. Here, exploratory
factor analysis (EFA) was initially applied to the first half of the data in order to investigate the
construct’s structure and to detect a smaller set of items which seem to be appropriate in
psychometric terms to determine the higher-order construct (i.e., scale purification). A series of
internal consistency measures was applied to evaluate the reliability of the identified trust
elements. Afterwards, in order to validate the distinct components of eWOM trust, a
confirmatory factor analysis (CFA) was applied to the same sample. The results were cross-
validated with the data from the holdout sample.

In respect to Churchill’s (1979) and other scholars’ recommendations, new data from a
representative sample of adult online consumers was collected during the validity stage. The
application of CFA and other procedures enabled the confirmation of the construct’s order,
dimensionality and the measure’s reliability. Additionally, various types of validity (e.g.,
concurrent) were assessed by using this study’s data. For instance, to test for convergent and
discriminant validity, the relationships between eWOM trust and review credibility (RCred), as
well as attitude towards reviews in general (RAtt), were investigated. During this stage, various
other forms of validity (e.g. predictive) were evaluated by collecting additional data from new

164
samples in a series of complementary studies. Here, for instance, in order to test for nomological
validity, the earlier theorized relationships between the construct and its antecedents, correlates,
and consequences were empirically investigated. Further, these studies also enabled this
research to cross-validate the findings of the main study by using pooled data as well as the
assessment of the scale’s generalizability to different contexts. The development of norms and
the application of g-theory concluded the validity stage and the scale development process. As
the aim of this research was also to demonstrate the usefulness of the new measure, a
segmentation study was conducted, introducing a new typology of eWOM/Ad trusters by means
of cluster analysis (application stage). In the remainder of this chapter, each of the major stages
of the research process is discussed in greater detail. This discussion includes a presentation of
the studies which were designed to answer the research questions and hypotheses proposed in
Chapter 3.

Identification Stage
Specification of Content Domain, Construct Definition and Dimensionality

According to the majority of scholars, the first step in the scale development process is a sound
theoretical definition of the focal concept, which should entail what is included in the domain
of the construct and what is not (Netemeyer et al., 2003). Further, it shall be clarified to which
entity the construct generally applies, the conceptual theme shall be outlined and a theoretical
justification about the construct’s dimensionality should be provided (e.g., MacKenzie et al.,
2011; Rossiter, 2002). While all these topics have been discussed in greater detail in Chapter 3,
this sub-chapter is dedicated to an issue that may need further explanation. It’s the construct’s
dimensionality or – more specifically – the relationship among the various first-order sub-
dimensions (types of trust), as well as between the dimensions and the focal, second-order
construct (eWOM trust).

In social sciences and specifically applied marketing research, constructs are regularly
conceptualized and operationalized as multi-dimensional entities (Brewer, 2007; Lin et al.,
2005). A construct is theorized to be multi-dimensional “when it consists of a number of
interrelated attributes or dimensions and exists in multi-dimensional domains. In contrast to a
set of interrelated uni-dimensional constructs, the dimensions of a multi-dimensional construct
can be conceptualized under an overall abstraction, and it is theoretically meaningful and
parsimonious to use this overall abstraction as a representation of the dimensions” (Law et al.,
1998, p. 741).

165
 IDENTIFICATION STAGE Conceptualizing the construct

166
Generating an inital item set from literature
Key targets:
Conceptualization of the construct Study 1: Expert interviews
Content validity
Development of a measurement model 8 marketing academics and 8 practitioners
Generating and judging an initial set of items
Study 2: Consumer interviews Face validity
20 adult online consumers

RELIABILITY STAGE Study 3: Online survey

850 students from 2 Austrian universities
Key targets: Split-half sample
Initial empirical assessment of psychometric
Figure 4: Overview of the Research Process

properties (e.g., dimensionality, reliability) First half: Second half:

Estimation sample Holdout sample
425 students 425 students RQ1: Construct dimensionality

VALIDITY STAGE
Study 4: Online survey
Key targets: 526 adult online consumers
Re-assessment of the new measure from Austria and Germany (Online panel)
Scale finalization Quota sampling on age group and gender RQ2: Concurrent validity
Supplemental Evaluations (e.g., diverse RQ3: Convergent validity
validity forms, additional reliability/ RQ4: Discriminant validity
Study 5a-e: Paper-and-pencil surveys
generalizability topics) H1-14: Nomological validity
824 adult online consumers
from Austria and Germany
Quota sampling on age group and gender RQ5: Known-group validity
RQ6: Predictive/postdictive validity
Study 5f-i: Supplemental surveys RQ7: Test-retest reliability
RQ8-9: Generalizability (online sources)
RQ10-11: Generalizability (reviews)

APPLICATION STAGE Study 6: Online survey

517 adult online consumers RQ12: Generalizability (country/language)
Key targets:
from the United States (Online panel) RQ13: Consumer segmentation
Application of the new measure as a
Quota sampling on age group and gender
segmentation basis for online sonsumers
It is also this thesis’ approach, as outlined earlier, to understand eWOM trust to be multi-
dimensional, as a person’s generalized trust is likely to be manifested in multiple aspects. More
specifically, eWOM trust is viewed as a second-order construct that is mirrored by three types
of trust, namely (i) the consumer’s confident beliefs that the information conveyed in customer
reviews is honest, useful, and benevolent (i.e., eWOM trusting beliefs); (ii) the consumer’s
positive attitude towards eWOM in general (i.e., eWOM trusting attitude); and (iii) the
consumer’s generalized willingness to rely on review information and recommendations for
making purchase-related decisions. Further, the construct of eWOM trust is regarded as being
– at the same time – multi-faceted, best described by a set of distinct aspects including cognitive,
emotional and conative elements. This is also captured by the above mentioned sub-dimensions,
where emotions, cognitions and conations trigger different types of trust. While each of these
five facets can be understood as a separate sub-construct, at a more abstract level, they are all
critical elements of a person’s generalized trust in OCR. This conceptualization takes on the
tripartite conceptualization of a person’s mental orientation towards an object, advanced, for
instance, by Bagozzi et al. (1979); but it at the same time parallels MacKenzie et al.’s (2011, p.
301) assertion of multi-dimensional concepts that “if the essential characteristics [i.e.,
dimensions] describe relatively unique aspects of the construct, eliminating any of [its
dimensions] would restrict the conceptual domain of the construct, then the construct is multi-
dimensional from a conceptual perspective”.

Given that eWOM trust is conceptualized to be a higher-order construct, the nature of the
relationships within the construct has to be considered. Or, in other words, a critical issue when
developing a new scale is to determine (i) the nature of the relationship between second-order
latent construct and its sub-dimensions; as well as (ii) the relationship between the latent sub-
constructs and their manifest indicators (e.g., Diamantopoulos et al., 2008; Jarvis et al., 2003).
Latent constructs can be defined as “phenomena of theoretical interest which cannot be directly
observed and have to be assessed by manifest measures which are observable”
(Diamantopoulos et al., 2008, p. 1204). On each abstraction level of the construct, two basic
kinds of relationships are possible. Here, researchers typically make a distinction on the basis
of whether the indicators on the item-level or the sub-dimensions on the construct-level are
influenced by a latent first or respectively the second-order variable – or vice versa (Bollen,
2011). If the direction of the relationship is from the latent variable to its measures, these
measures are regularly called reflective or effect indicators. In contrast, if the direction of the
relationship is reversed – from the measures to the construct – the measures are commonly
denoted as formative or causal indicators (Blalock, 1964; Bollen & Lennox, 1991).

The existence of this first kind of relationship (i.e., the reflective measurement model) has long
been recognized in social sciences. Academic assumptions about this relationship are here
based on classical test theory (Lord & Novick, 1968). Diamantopoulos and Winklhofer (2001,
p. 269), for instance, state that “this perspective reflects the conventional wisdom on

167
measurement […], which is largely based on classical test theory and, in particular, the domain-
sampling model”. This theory asserts that variations in the indicators constitute effects (or
manifestations) of variations in an underlying latent construct (Bollen & Lennox, 1991). Hence,
reflective indicators depend on the latent construct or rather are caused by them
(Diamantopoulos et al., 2008). An important characteristic of this kind of relationship is that
changes in the latent variable lead to changes in all of its measures at the same time.
Additionally, it corresponds to the general notion of the reflective measurement model that all
measures have to be positively correlated (Bollen, 1984); therefore, the measures are more or
less exchangeable.

The second kind of relationship (i.e., the formative measurement model) has only recently
experienced heightened interest in the research community regardless of the fact that the model
was first proposed several decades ago by Curtis and Jackson (1962). These researchers
recognized that, in specific cases, measures of the same construct exhibit no, or even negative,
correlations. Accordingly, it is assumed that measures affect the latent variable or are its causes
and not its effect. Hence, as the indicators represent distinct causes, only their co-occurrence
determines the meaning of the latent construct. It follows that formative measures are not
interchangeable (Jarvis et al., 2003; Rossiter, 2002) and omitting an indicator potentially alters
the nature of the whole construct (Bollen & Lennox, 1991; Diamantopoulos et al., 2008).

Having said this, the measurement model mirrors this research understanding of the focal
construct being best measured with reflective indicators on the item- as well as overall
construct-level. More specifically, this thesis formally models eWOM trust as a second-order
latent construct that has multiple first-order sub-dimensions as reflective indicators and the sub-
dimensions themselves have multiple reflective observable indicators. This implies that it is
assumed that observable indicators reflect the specific characteristics of the latent sub-
constructs. According to Jarvis et al. (2003, p. 202), with such reflective indicators “the
direction of causality flows from the construct [i.e., sub-dimensions or types of eWOM trust]
to the measures [i.e., observable indicators]”. Further, the conceptualization implies the eWOM
trust construct to exist on a higher, more abstract level and that its five dimensions are best
viewed as consequences of the focal construct. Accordingly, the various types of trust – which
represent different aspects of trust and mirror a common orientation towards the trust object –
are considered to reflect a person’s “overall” eWOM trust as their overall latent variable (Bollen
& Lennox, 1991; Diamantopoulos et al., 2006). In other words, the first-order latent constructs
are a manifestation of the same focal construct (Wong et al., 2008). Hence, eWOM trust as a
construct is assumed to exist separately at a deeper and more embedded level than its sub-
dimensions, and further it is assumed that changes in the latent construct (increases/decreases
in the person’s global orientation towards OCR) would lead to simultaneous changes in all of
its sub-dimensions (MacKenzie et al., 2011) or the person’s trusting beliefs, attitudes, and
intentions to rely. It seems appealing to theorize that if individuals adopt their general

168
disposition to trust eWOM, they do this by allowing their general attitude towards the trust
object consistently to influence their thoughts, feelings, and conations. A well-trusting person
is hence unlikely to have a high belief in the OCR’s honesty while not having the opinion that
it is benevolent. In literature, higher-order constructs of this kind have been called totally
disaggregated constructs by Bagozzi and Heatherton (1994), superordinate constructs by
Edwards (2001) and Williams et al. (2009), or type I constructs by Jarvis et al. (2003).

Such an approach is consistent with earlier conceptualizations of trust and closely related
concepts (e.g., Casalo et al., 2011; Corbitt et al., 2003; Flavian & Guinaliu, 2006). However,
this research also acknowledges the rising debate concerning appropriate measurement models
suitable for trust-related constructs. Such a discussion is justified, as various researchers have
misspecified trust-constructs in different contexts. This is because there is more than one type
of trust construct.

When conceptualizing eWOM trust, it was understood that constructs are per se not inherently
reflective or formative (Baxter, 2009; Diamantopoulos, 2011; Wilcox et al., 2008). In fact, it is
particularly the characteristic of the latent variable conceptualized by the researcher and also
the content of the indicators which both determine how the construct and its indicators are
related and whether the use of reflective or formative indicators is appropriate (MacKenzie et
al., 2011, p. 302). According to Hildebrandt and Temme (2006), the underlying research
question dictates whether a reflective or formative measurement approach is appropriate to
operationalize the construct.

In general, attitudinal constructs can be typically operationalized with both approaches (Albers
& Hildebrandt, 2006). While the nature of the observable (reflective) indicators due the items
used may not give rise to controversy, the conceptualized causal relationship between the sub-
dimensions (types) of trust and the overall eWOM trust construct may have to be clarified. The
proposed measurement model is deducible from prominent literature strings in trust research,
as well as personality research. The latter emphasises that trait-like constructs like the
disposition to trust exist, which are stable over time and are closely linked to a person’s identity.
This literature provides a valuable auxiliary theory (Bagozzi, 1982; Blalock, 1969) that enables
this thesis to make a conceptual decision about the linkages between the construct of interest
and its sub-dimensions. The previous chapters (and folk wisdom) tell that individuals are able
to develop trust in various objects, contexts and situations. Trust can arise suddenly due a
calculation process or current feelings about the trust object (which are therefore
environmentally caused) (i.e., situational or initial trust), but trust can also exist on the
personality level of a person (i.e., generalized trust). This makes it unrealistic to argue for the
existence of a single, generalizable trust construct. While the various trust conceptualizations
seem to share some common ingredients, it is the close context but also the specific domain
which both determine the most appropriate measurement model in the eye of the researcher.

169
Since the nature of the trust construct is dependent on various circumstances and therefore
specific to the research object, it is inadequate to argue for a single measurement model that
captures the construct under any circumstances.

For instance, some trust researchers explicitly (or implicitly) advocate the view of what Jarvis
et al. (2003) describe as a type II construct, where the sub-dimensions of trust are considered
causal for the trust construct. Such a conceptualization is in accordance with a distinct literature
stream, understanding trust as consisting of different types or levels of trust (e.g., Boon &
Holmes, 1991; Lander et al., 2004; Panteli & Sockalingam, 2005; Zaher et al., 1998). Earlier
conceptual contributions have confirmed the existence of subsequent trust levels, as well as the
fact that different components of trust become operational at specific trust levels. By adapting
the view of Boon and Holmes (1991), Shapiro et al. (1992), as well as Lewicki and Bunker
(1995), numerous authors (e.g., Ba, 2001; Boon & Holmes, 1991; Lander et al., 2004; Zaheer
et al., 1998) agree that trust is a dynamic process and it possesses a distinct character in different
stages of relationships (Hsu et al., 2007). Trust is here best viewed as a phenomenon that is
time-consuming to engender, hence, trust is developed through repeated interactions over time.
It’s very unlikely that individuals are able to perceive certain degrees of trust without a
minimum of previous experience. However, trust is finally situationally constructed, fragile and
it can be easily destroyed.

The above-cited scholars furthermore advocate the proposition that trust develops in sequential
stages. For example, due to the work of Panteili and Sockalingam (2005), three stages of trust
exist which are linked in a sequential iteration. This hierarchical approach to trust means that
the achievement of trust at one level enables the development of trust in the next level.
Additionally, the trust conceptualization of Sheppard and Sherman (1998), for instance,
identifies four distinct and ordered forms of trust: (i) shallow dependence, (ii) shallow
interdependence, (iii) deep dependence, and (iv) deep interdependence. In a similar vein,
Barney and Hansen (1994) conclude that trust strength can be categorized as weak, semi-strong,
and strong forms of trust. Implicitly, McAllister (1995) has contributed to a better
understanding of trust levels by proposing two different elements of trust: (i) cognitive-based
trust, defined as an individual’s rational use of evidence and analysis to form beliefs and attitude
towards a trusted party, can be easily inferred as representing a low level of trust. In contrast,
(ii) affect-based trust, which aims to measure the emotional bonds among truster and trustee,
signals a high level of trust (Chen & Dhillon, 2003). This view is also supported by Wicks et
al. (1999). Lewicki and Bunker (1995) add that trust can be conceptualized in three sequentially
ordered and linked forms of trust. (i) The lowest level of trust but also weakest form of trust is
calculus-based or cognition-based trust. Here, trust is simply based on the truster’s beliefs
about the trustee’s characteristics and thoughts about the trustee’s future intentions and
behaviour. Calculus-based trust can be typically found in new or first-time relationships. (ii) In
contrast, knowledge-based trust can only be formed over time, where the interaction partners

170
have the opportunity to increase their knowledge about each other. Here, the level of trust is
more distinctive, as the parties form relatively strong bonds, needed to establish an exchange
relationship. (iii) Identification-based trust mirrors the highest and strongest level of trust,
which is least fragile to changing environment. Here, the parties share common interests as well
as values. This view is also consistent with the common argument that trust shall be deemed to
develop over time and repeated interactions (e.g., Mayer et al., 1995; Rempel et al., 1985;
Silence et al., 2006) but can also arise in the first encounter of interaction partners (e.g., Stewart,
2003; Trifts & Häubl, 2003).

Other researchers have supported the view that the individual’s perceptions of the trust types
(i.e., trusting beliefs, attitudes, and intentions) can be used to infer levels of trust. For instance,
Grimes (1978) as well as Wicks et al. (1999) suggest that the attitude of confidence represents
a heightened level of trust. There is additional evidence that trusting intentions, as they involve
a person’s willingness to take a risk or to be vulnerable, implies high trust.

The thesis at hand recognizes and values the contributions of researchers who have discussed
how trust changes with the passage of time. However, having conceptualized eWOM trust as a
relatively stable orientation to rely on online customer reviews and recommendations that has
been developed with prior experiences, this thesis does not see that specific types of trust
become operational at different stages of trust. Rather, in line with trait research, it proposes
that the relationship between the different types of trust (i.e., dimensions of eWOM trust) is not
incremental and additive, but consistent. The general nature of the relationship determines its
consequences and not the kind of relationship between the trust object and the truster.
Accordingly, it is assumed that a person’s high levels of eWOM trust should be mirrored by
adequate levels of trusting intentions, trusting attitudes as well as trusting beliefs at the same
time. Individuals tend to have a uniform orientation and are always able to express this
conformity. Psychological literature is far from answering the question whether cognitions
precede emotions or vice versa and controversies long dominate the scientific discussion (Izard,
1991, 1992; Leventhal & Schwerer, 1987). For instance, Zajonc (1984) argues that affect and
cognition are separate and partially independent constructs and that affect can be generated
without any prior cognitive process. Contrasting with this view, Lazarus (1984) characterizes
emotions to be completely intertwined with cognitive processes. Accordingly, the theoretical
background of above-cited literature may also be compromised, as the problem concerning the
sequence of occurence of the levels or types of trust is also not completely resolved.

In this thesis’ conceptualization of the eWOM generalized trust orientation, the construct is
simultaneously mirrored by discrete, but related types and elements of trust. Such a view
emphasizes trust as a trait-like social concept leading to cross-situational and cross-personal
enduring individual differences (Costa & McCrae, 1995; Murtha et al., 1996). It can be that
some situations induce slight variations in the level of trust, but in general it remains stable over

171
time and context (Hogg & Vaughan, 2011). Concerning its conceptual nature, eWOM trust is
hence more closely related to an individual’s disposition to trust (Bowlby, 1973; McKnight et
al., 1998; Rotter, 1967, 1971) than a situation-dependent construction of trust. Scale
development literature regularly demands from multi-dimensional constructs representing a
general personality trait that the relationship between the construct and its sub-dimensions are
captured by the superordinate construct (MacKenzie et al., 2011; Williams et al., 2009). For
instance, for Fornell and Bookstein (1982, p. 292), “constructs such as “personality” or
“attitude” are typically viewed as underlying factors that give rise to something that is observed.
Their indicators tend to be realized, then as reflective”. This is also acknowledged by Albers
and Hildebrandt (2006), who state that a reflective measurement approach is appropriate for
most attitudinal/behavioural constructs.

According to MacKenzie et al. (2011), a second-order measurement model with multiple first-
order dimensions as reflective indicators is appropriate if (i) the latent construct is stable (i.e.,
an individual difference variable) over time or across situations; and (ii) the focal construct is
measured with randomly selected indicators, each of which is reflective of a focal construct.
The latter implies that the sub-dimensions are different manifestations of the second-order
factor (MacKenzie et al., 2011). A change in any of the first-order factors is therefore assumed
to result in a change in the other underlying indicators (Serva et al., 2005). It follows that first-
order constructs have to be strongly correlated to support a reflective conceptualization.
Numerous studies emphasizing trust and trust-related constructs have revealed that a high
degree of intercorrelations between trust elements (i.e., trusting beliefs) are regularly found
(e.g., Bhattacherjee, 2002; Gefen, 2002b; Gefen et al., 2003; Hsio et al., 2010; Mayer & Davis,
1999; Murphy, 2003; Schoorman et al., 1996). Other studies have identified a reasonable
relationship between trusting beliefs and trusting attitudes, as well as trusting intentions (e.g.,
Benedicktus et al., 2010; Chang & Chen, 2008; Chen & Barnes, 2007). Such positive
intercorrelations imply the existence of an underlying higher-order factor (Bollen, 1989) – or
what is here called eWOM trust.

The above-cited literature provides evidence that the application of a type I construct in this
research context is appropriate, as trusters tend to have uniform perceptions of the different
aspects of the trustee (here, online reviews/reviewers), personal experiences and behavioural
intentions. That is, individuals maintain consistency in their evaluations, especially when these
provide the basis for a constant personality trait (Serva et al., 2005). Such a perspective is
supported by a number of cognitive consistency theories (Abelson et al., 1968), including
Festinger’s (1957) cognitive dissonance theory and Heider’s (1946) balance theory. In their
essence, these theories argue that human beings strive to resolve conflicting perceptions – they
try to maintain internal consistency and agreement among their various cognitions, beliefs, and
feelings (Hogg & Vaughan, 2011). For example, cognitive dissonance theory assumes that if
there are inconsistencies between an individual’s attitudes and his/her behaviour, the person

172
will impose changes in order to resolve the currently existing dissonance. One way that
individuals can reduce dissonance is by altering the dissonant beliefs so that they are no longer
inconsistent (Festinger, 1957). Human beings constantly try to confirm existing perceptions
about other persons or objects. This also holds true in situations where individuals have to
accept beliefs or attitudes that lack reasonable foundation (Kahnemann et al., 1982; Mynatt et
al., 1977), or when they have to reject conflicting information in favor of confirming evidence
(Serva et al., 2005). In order to explain such consistency in individuals’ mental processes, type
I constructs are regularly used in marketing (e.g., Huang et al., 2011; Riefler et al., 2012) as
well as trust research approaches (e.g., Casalo et al., 2011; Wang & Benbasat, 2005). The
process described above is analogous to this research’s second-order conceptualization of
eWOM trust. Consumers are likely to form consistent perceptions of trusting beliefs, attitudes
as well as intentions in order to justify their general level of trust towards and usage of opinions
of the reviewers. A consumer who believes in the integrity of online reviews has at the same
time a strong willingness to rely on such a source of information – and at the same time will
show approach behaviour towards OCR. Attribution theory (Hogg & Vaughan, 2011) or the
constant human process of assigning a cause (that is, having favourable thoughts, feelings, and
intentions towards the trust object) to our behaviour (i.e., regular eWOM usage) may also help
to explain this consistency. This is regarded as a meaningful strategy to avoid dissonance and
express inner, consistent orientation.

The thesis further argues that the content of eWOM trust is determined by the general possibility
of developing a certain position towards the different aspect of trust, as well as the kind of
relationship between the consumer and the trust object. Concerning the former, it has to be
mentioned that eWOM trust assesses a person’s general and relatively stable reliance on a
specific information source. This reliance has been developed over time and has therefore
nothing to do with evaluating a trust object in a specific situation that is maybe new to the
consumer. Rather the individual had the time to develop a certain orientation towards the trust
object and holds a cognitive, emotional and conative position towards the trust object. The
eWOM trust scale hence is intended to capture the true level of trust in situations where the
truster has (at least some) prior experience with the trust object. In contrast, trust in the initial
stage of a relationship or contextual development of trust in specific OCR (i.e., initial and
situational eWOM trust) is likely to be characterized by other parameters. Here, the truster may
not be able to form an attitude or develop emotional bonds yet. Therefore, measuring trust with
affective trust types such as trusting attitude may not be appropriate at this stage (c.f., Kim &
Tadisina, 2007). This approach is also implicitly taken by various scholars who have
investigated initial trust. For instance, the research schools around McKnight and Gefen have
restricted their operationalization of first-time trust on trusting beliefs which are predominantly
cognitive in nature. Similarly, various other trust researchers have concentrated intensively on
the truster’s cognitive beliefs when discussing the components of trust. However, a
recognizable stream of literature emphasizes that emotional orientations as well as behavioural
173
intentions have to exist together with a cognitive component in order to represent a construct
that can be called trust (Johnson & Grayson, 2005; Lewis & Weigert, 1985).

Another proposition of this literature string which is also picked up in this thesis’ construct
conceptualization is that the relative importance of the trust components is dependent on the
kind of (social) relationship between the truster and the trustee (Lewis & Weigert, 1985).
Accordingly, for primary-group relationships (e.g., friendship), trust can be best described in
terms of affective aspects, while cognitive aspects are likely to dominate in the context of
secondary-group relations. The dominant role of cognitive trust has also been found in the
context of social interactions with secondary groups (Johnson-George & Swap, 1982; Rempel
et al., 1985). Basically, when consumers develop trust in online customer reviews and
recommendations over time they form a general orientation towards a social institution or a
kind of secondary-group relationship. This implies that eWOM trust is mainly based on
cognitive trust components.

This research acknowledges that the concept of eWOM trust does not belong to the group of
latent constructs where the causal direction between indicators and their respective constructs
(e.g., Diamantopoulos & Winklhofer, 2001; Podsakoff et al., 2003) can easily be defined. The
true nature of trust is that a generalizable and applicable “trust construct” does not exist. Rather,
trust has to be conceptualized in terms of its specific object, context and scope. The role of the
“tripiad” may be differential in these various trust contexts. Hence, the focal construct has to
be investigated from a larger research context, and the appropriateness of its measurement by
means of several criteria (Wilcox et al., 2008). In such a complex conceptual environment, it
may be sometimes more appropriate to measure trust with an index (Diamantopoulos &
Winklhofer, 2001), while in other cases the development of a trust scale based on conventional
scale development procedures (e.g., Churchill, 1979; DeVellis, 2012) may be more
constructive, as it was considered in this research. This thesis argues that viewing trust as what
Jarvis et al. (2003) called a type II construct may be appropriate in specific trust contexts only.
More specifically, it is assumed that a formative view of trust on the second-order level is
adequate when trust is understood as a situational variable which results from a contextual
judgment process of specific characteristics of the trust object (the reviews just read), such as
trusting beliefs (e.g., situational eWOM trust). Here, the elements of trust are likely to be (more
or less) evaluated independently from each other and together “form” a latent construct which
represents overall trust. In such a research context, trust is typically conceptualized as a
situation-specific variable towards a particular trust object (e.g., Jarvenpaa et al., 2000; Kim et
al., 2008) and consumers are regarded as being able to make independent judgments on the
discrete elements in order to make an objective trust decision. When a consumer faces a specific
review, he/she will decide upon (mainly cognitive) judgments about the characteristics of these
reviews as well as upon personal feelings and general orientations whether or not he/she has

174
trust in the opinions of others. (However, this judgment would be heavily influenced by
generalized eWOM trust, as outlined before.)

Some researcher state that decisions about the nature of the measurement model should not be
taken lightly, as assuming wrongly a reflective measurement model where a formative would
be appropriate leads to expansive errors in model specification – or what some researchers call
a type I error (Diamantopoulos & Siguaw, 2006; Jarvis et al., 2003; Petter et al., 2007). Various
studies attempted to gain insight into the consequences of model misspecification. For instance,
by reviewing an earlier study (Cadogan et al., 1999), Diamantopoulos and Siguaw (2002)
demonstrate that different measurement models lead to different sets of indicators. This
variation finally results in a completely different meaning of the constructs involved. Albers
and Hildebrandt (2006) relax this insight and state that specifying a measurement model
“wrongly” only results in a restricted model and not an erroneous one per se. Differences in
results remain small as long as one refrains from excluding specific items. This said, the eWOM
trust scale developed in this research is based on a literature-based construct conceptualization
deemed to be appropriate for a stable, trait-like construct. Assuming that the five dimensions
are different manifestations of the same construct (Wong et al., 2008), and assuming that
“eliminating any of them would restrict the conceptual domain of the construct” (MacKenzie
et al., 2011, p. 301), this thesis is confident that formally modelling eWOM trust as a (reflective)
second-order construct with five (reflective) first-order dimensions is appropriate.

Generating and Judging Items

Having conceptualized the dimensions of trust being best measured by reflective indicators, the
next step was to generate a set of initial items that captures the domain of each dimension as
specified. This research applied domain sampling and procedures of exploratory research that
have been found productive in this respect (Churchill, 1979; MacKenzie et al., 2011; Netemeyer
et al., 2003). Domain sampling suggests that a measure should be represented by a sample of
items drawn from a large hypothetical domain of items (the sampling universe) which all mirror
(in this case) the meaning of the diverse trust sub-dimensions (Netemeyer et al., 2003). Hence,
in order to identify words and phrases that are associated with trust and associated constructs –
in line with literature recommendations (e.g., Churchill, 1979; DeVellis, 2012; Haynes et al.,
1995) – multiple source have been consulted. First, a systematic review of interdisciplinary
literature in the areas of marketing, management, psychology, sociology, as well as information
systems and communications research was conducted. Here, several items were derived from
previous published sources on trust and trust-related attitudinal constructs (e.g., credibility,
information usefulness/helpfulness, distrust) that were assumed to capture the domain of the
construct, or at least parts of it. Also, this research paid attention to previous theoretical and
empirical research on the focal construct (i.e., trust in online reviews and online reviewers), as
well as other measures of the construct which already exist. Thus, items were generated from

175
general and domain-specific literature. A review of various dictionaries and thesauri completed
the research. Additionally, some items were freely constructed by the scale developer from the
descriptions and definitions of the various dimensions outlined by academic research. Hence,
this study used the often-recommended deductive approach to arrive at a comprehensive set of
items.

This process resulted in an extensive list of 620 words and phrases that have been found to be
frequently used to describe aspects of trust. Given the complex nature of the focal construct,
this quantity of items appears adequate (Netemeyer, 2013). By means of a detailed review and
rigorous item selection process performed by the author and a second marketing researcher, the
extensive list of items was trimmed to a more manageable number, and the initial set of items
for the individual sub-dimensions of eWOM trust was produced. Item elimination was based
on a variety of a priori criteria regularly advocated in literature (e.g., DeVellis, 2012; Peterson,
2000; Podsakoff et al., 2003): Here, the researchers especially paid attention to issues of
contextual fit and word redundancy, but they also looked to include various items with slightly
different shades of meaning, which enabled this study to construct multi-item subscales.
Additionally, the wording had to be simple, short, familiar to the respondent and clear, as well
as precise. The researchers’ aim was also to eliminate items that were ambiguous, impossible
in the context, had questionable relevance or contained obvious social desirability (Nederhof,
1985). There remained only words and phrases that are used in common language. In parallel,
jargon and trendy slang, as well as double-barrel statements, were avoided. The reduced set of
items should also only include items that exhibit meaningful semantic differences (Ablers &
Hildebrandt, 2006).

By means of an iterative discussion process, a total set of 80 items was generated which was
hypothesized to fully capture all the essential aspects of the domain of the focal construct. Note
that none of these items exhibited negative expressions (was reverse coded). Due to the
conceptualization of eWOM trust as a confident or positive orientation best mirrored by positive
judgments, feelings, and motivations towards the eWOM information, this research refrained
from using negatively worded items. Their enclosure in the new scale would automatically lead
to the measurement of eWOM distrust, a related but conceptually different construct (see
Chapter 3). Additionally, in scale development research, items worded in the opposite direction
often perform poorly (DeVellis, 2012). However, the author is aware that some response bias
in the final scale may be attributed to this circumstance.

Literature regularly demands that besides the deduction of items from reviews of literature, also
suggestions by appropriate people – including (i) representatives of the population to which the
construct of eWOM trust is expected to generalize (i.e., online consumers – individuals using
the Internet for product research and/or shopping), as well as (ii) persons from the marketing
environment – should be considered (e.g., DeVellis, 2012; MacKenzie et al., 2011). The

176
research had to ensure that all content areas of the construct were systematically sampled. That
is, they are representative for the various aspects of trust. Hence, all items were judged in terms
of translation validity. Basically, the assessment of translation validity includes the evaluation
of the items in terms of both (1) content and (2) face validity. By referring to Straub et al.
(2004), the term content validity reflects “the degree to which items in an instrument reflect the
content universe to which the instrument will be generalized” (p. 424). Other authors put
forward similar definitions (Kerling, 1973). According to Haynes et al. (1995), content validity
is concerned with “the degree to which elements of an assessment instrument are relevant to
and representative of the targeted construct for a particular assessment purpose” (p. 238). Here,
by “elements”, the authors refer to the diverse items of the scale, the response formats, as well
as to the instructions to the respondents. In line with Netemeyer et al.’s (2003) argumentation,
this thesis understands “representativeness” as the degree to which the items seem to be
proportional to the different aspects of the focal construct and the degree to which the whole
population of the aspects of trust has been sampled. Hence, it has to be clarified whether the
individual items are representative for the distinct sub-dimensions of trust or not, but also the
degree has to be assessed to which the items as a set collectively represent the entire content
domain of the construct (MacKenzie et al., 2011). Literature often suggests that this can be
achieved by a screening of the instrument by judges who have professional expertise in the area
of research (DeVellis, 2012).

While some authors refrain from separating face validity from content validity, others do so.
The same is true for this thesis. In agreement with the work of Netemeyer et al. (2003), this
thesis understands face validity as the mere appearance that a measure, in addition to having
pragmatic or statistical validity, is practical, pertinent and related to the purpose of the
instrument as well (Kaplan & Saccuzzo, 1997; Nevo, 1985). This means that a measure should
not only be valid, but also should “appear valid” in the eyes of the respondents. A face valid
measure ensures that the items, instructions and the response format are clear and
understandable. While face validity is more concerned of the evaluations of the items by the
respondents of the target population, content validity on the other hand is focused on the scale’s
validity from the viewpoint of experts. In order to cope with both forms of translation validity,
as well as to trim and refine the pool of items, interviews with both marketing experts and online
consumers were conducted.

Expert Interviews (Study 1)

In order to assess content validity, the initial set of items was subject to an expert screening by
means of structured interviews. In total, 16 persons were recruited for this purpose. Eight were
marketing faculty members (three persons holding a PhD degree and five were PhD candidates)
from two major Austrian universities, whose primary research area was consumer behaviour
and/or psychology. The second half were experts in the field of Internet marketing (e.g.,

177
consultants). The construct’s content validity was ensured by applying recommended
procedures (e.g., Netemeyer et al., 2003) and adequate methods earlier used in scale
development (e.g., Riefler et al., 2012; Soh, 2007; Walsh & Beatty, 2007). First, the experts
were asked to define trust in eWOM in their own words. This was particularly done to get
insight into the understanding of the main construct, its content, associations and meaning from
the experts’ perspective. An open-ended question appeared most appropriate. Then, the experts
had to identify the construct’s underlying factors that cause heightened trust. Afterwards, this
study’s own general definition of eWOM trust, as well as the definitions of the construct’s sub-
dimensions, were provided. The experts were asked to evaluate their appropriateness. All
experts agreed that the definitions theoretically represent the nature of trust in online customer
reviews and its diverse aspects. Also the diverse sub-dimensions are evaluated as being
legitimate components of eWOM trust and no further dimensions were proposed. Then, by
adopting a technique frequently used in exploratory research on scale development (e.g., Riefler
et al., 2012), the marketers were asked to judge the individual items in respect to their relevance
for capturing the separated trust aspects on a scale ranging as either “applicable” or “not
applicable” (Netemeyer et al., 2003). The experts were also asked to make suggestions for
additional items where the proposed set of items did not comprehensively tap the constructs’
content domain (DeVellis, 2012). Hence, items were also generated inductively – providing the
advantage of gaining a profound insight into the experts’ understanding of the core constructs
and its associations (Albers & Hildebrandt, 2006). Scale items were retained if at least 70 per
cent of the experts consistently rated the items to be “applicable”. This screening process
resulted in the elimination of ten items and no additional items were added. Some suggestions
referred to negatively worded items, but since there was a minor agreement among the
marketing experts that these items truly tap the domain of the focal construct (e.g. Jones &
George, 1998; Sitkin & Roth, 1993), they were not considered relevant for further research.
The pool of items was then pre-tested with 20 consumer interviews in order to make sure that
the remaining items also appear representative, clear and comprehensible to the scale’s target
population.

Consumer Interviews (Study 2)

To further assess the degree of face validity, this research conducted interviews with a
convenience sample of 20 participants. Again, the research’s definitions as well as proposed
items were the subject of evaluation. In line with MacKenzie et al.’s (2011) propositions,
undergraduate business students were recruited to participate in an item-sorting task. All
students participated for extra course credits. The reasons for selecting students are manifold.
For instance, these persons are assumed to possess sufficient intellectual abilities to rate the
correspondence between the theoretical definitions of the trust dimensions and its items.
Further, it is assumed that students are representative of the population to which the new eWOM
measure is supposed to generalize. Recent surveys provide some evidence that Internet

178
shoppers are generally young and well educated compared to the general public (Statistik
Austria, 2012). Therefore, this research assumes that it meets the claim advanced by Anderson
and Gerbing (1991) that the raters should be members of the relevant population – at least to
some degree. In parallel, other authors advocate that college students can be used for such tasks
(Hinkin & Tracey, 1999; Schriesheim et al., 1993b).

The consumer interviews were structured as follows. First, the students were asked to indicate
their overall level of trust in online customer reviews and recommendations. After that, an open-
ended question was posed that targeted the background of why he/she puts trust in eWOM in
general, or not. This question strived for the identification of factors that may impact trust
evaluations. Similarly to the expert items, the students were also asked to find a definition of
eWOM trust in their own words. This research then applied the method advanced by Anderson
and Gerbing (1991), and also by others (e.g., Riefler et al., 2012), for evaluating face validity
of the individual items. For this purpose, the students were furnished with a hypothesized
definition of the focal construction, as well as with descriptions of the five conceptualized trust
dimensions. The students agreed that the definitions of eWOM trust and its sub-dimensions all
cover their intended concepts and that the hypothesized trust components are likely to form a
construct that can be called eWOM trust. Hence, it was assumed that the domain of the eWOM
construct is theoretically broadly covered, as it is regularly regarded as being critical when
specifying reflective indicators (Diamantopoulos & Winklhofer, 2001; Jarvis et al., 2003).

In the course of the interview, the respondents received index cards for the 70 items and were
instructed to read the items attentively, to classify each item as either “relevant” or “irrelevant”
for representing trust in eWOM and to finally assign each card to one of the trust dimensions
which the item best reflects, according to his/her personal position. This sorting task enabled
the calculation of three indices in order to determine the face validity of the individual items.
(1) The degree to which the respondents regarded the item to be relevant for capturing the
meaning of eWOM trust. An item was retained where the majority of students classified the
item as “relevant”. (2) The proportion of substantive agreement (psa), which is defined as “the
proportion of respondents who assign an item to its intended construct” (Anderson & Gerbing,
1991, p. 732). In line with literature, psa was calculated using the following formula:

݊௖
‫݌‬௦௔ ൌ (1)
ܰ
The formula reads as follows: nc is the number of respondents who have assigned an item to its
intended dimension; N, on the other hand, represents the total number of respondents in the
interview. Anderson and Gerbing (1991) argue that larger values of psa are a representation of
greater substantive validity and propose a cut-off value of .5, which also was applied in this
research for eliminating items that potentially lack validity. Data also allowed the calculation
of a third index, to be specific, the substantive validity coefficient (csv). The index is defined as

179
“the extent to which respondents assign an item to its posited construct more than to any other
construct” (Anderson & Gerbing, 1991, p. 734). This research regards an item to be wrongly
classified by the respondent, if the respondent has allocated the item to a sub-dimension which
the item is not intended to measure. The index was computed as follows:

݊௖ െ ݊௢
ܿ௦௩ ൌ (2)
ܰ
Again, N and nc can be described as above, while no is defined as the highest number of
assignments of the items to any other but the intended dimension. All items that did not achieve
the sufficient cut-off value of .3 were candidates for exclusion. Respondents were also asked to
propose additional items that are associated with trust in the present context and which are not
covered by the other items.

This procedure resulted in a list of 53 German adjectives and phrases that are assumed to
comprehensively tap the domains of the eWOM construct. These items were used to create the
initial set of 7-point Likert-format items (see Table 5 for the original items of the eWT-S and
their English translations). For instance, the word “dependable” was perceived to be
representative for eWOM trust. Accordingly, the word was used in the phrase “The information
given in online customer reviews is dependable”, in order to meet the demands of Likert scales.
All items were measured on a 7-point scale ranging from 0 (“I strongly disagree”/”Ich stimme
überhaupt nicht zu”) to 6 (“I strongly agree”/”Ich stimme völlig zu”); with only the extreme
values being labeled. Hence, the used scale included a middle value for respondents who are
neutral or undecided. Research demonstrates that having a middle point increases reliability
and validity of the scale (Saris & Gallhofer, 2007). The selection of the number of response
options was based on recommendations advanced in scaling literature (DeVellis, 2012). While
some researchers suggest that 7-point and 5-point Likert scales produce comparable results
(Dawes, 2008), others state that a 7-point rating scale has slight advantages over a 5-Point scale
(Nunnally, 1978). It is regularly argued that rating scales with more response categories also
provide more insight into the variability within the data and allows greater differentiation of
responses (Alwin, 1992; Finn, 1972). In addition, research suggests that 7-point rating scales
carry a greater amount of information and are less vulnerable to both systematic measurement
error and shared method variance (Alwin, 1997). There is some evidence that more response
options increase validity (Rodgers et al., 1992). This scale also turned out to be more reliable
(Cronbach, 1951), which surely contributed to the fact that 7-point rating scales are also very
common in marketing research (e.g., Brace, 2004; Malhotra & Peterson, 2006). Foddy (1993)
states that, for abstract judgments (which is certainly the case when expressing a general attitude
about eWOM), the number of response scale options should be increased. These arguments
together guided this study to its application in the new scale. Further, this thesis opted to use a
forced-choice scale (e.g., Smith, 2002) throughout the research process. That is, the respondents
had to make a decision on the evaluation of each item.

180
The response format was chosen because of its established status for measuring opinions,
beliefs, and attitudes by means of self-assessment scales (e.g., Churchill & Iacobucci, 2004;
DeVellis, 2012; Netemeyer et al., 2003). Further, a review of empirical literature showed that
it has been frequently used in the context of trust (e.g., Corbitt et al., 2003; Gefen & Heart,
2006; Li et al., 2008; Smith, 2002; Soh et al., 2009) and trust-related constructs. Additionally,
experts as well as respondents from the target group interviewed in the qualitative phase of this
research agreed that Likert-like statements are appropriate. This research refrained from using
a semantic differential or bi-polar adjective pairs, since this response format is theorized to
place an unnecessary higher burden on respondents (Flynn, 1993). Lietz (2010) provides an
excellent review of the methodological shortcomings of using negatively worded items.
Further, negative adjectives are theorized, as mentioned earlier, to unintentionally mix eWOM
trust with separated constructs. The remaining items were subject to quantitative scale
purification methods in the next step of this research. Here, a questionnaire with the 53
statements was put to a sample of college students in order to assess the internal reliability, as
well as the hypothesized order and the dimensionality, of the construct discussed.

Reliability Stage
Having a relatively large set of face- and content-valid items, the next step in the research
process was to obtain quantitative data from an adequately sized sample of respondents which
enables this research to examine the psychometric properties of the proposed scale. What
Netemeyer et al. (2003) call “pilot testing” offers the opportunity to trim the item pool for
further developmental studies on the basis of stringent statistical criteria and to obtain some
initial estimates for the scale’s reliability, as well as validity. In order to refine the scale by
deleting items that do not meet certain psychometric properties, initial data was collected by
means of an online survey (Study 3). As discussed later, the data were randomly split into two
equally sized data sets (Sample 3a and Sample 3b) for further analyses. While sample 3a served
as the estimation (developmental) sample, which enabled the research to delete inadequate
items, sample 3b was used as a holdout sample for cross-validating the construct’s
dimensionality. In order to explore the structures of the construct and to identify appropriate
items, data from the evaluation sample was first subject to an exploratory factor analysis (EFA)
and subsequently to a confirmatory factor analysis (CFA). Similar data analyses were also
applied to the holdout sample in order to compare the scale’s psychometric properties across
the two samples. The details on the applied statistical procedures as well as on the results of the
analyses are reported in Chapter 5. What follows is a discussion of the contents of the used
questionnaire, the data collection routines and processes as well as details about the study’s
respondents.

181
Table 5: Initial Set of eWOM Trust Items

Sub-dimensions Items

The information given in online customer reviews is …

Informationen in Online Kundenrezensionen sind …

Ability versed relevant informative

sachkundig (Ab1) relevant (Ab6) informativ (Ab11)
objective helpful expert
objektiv (Ab2) hilfreich (Ab7) fachkundig (Ab12)
competent instructive experienced
kompetent (Ab3) aufschlussreich erfahren (Ab13)
(Ab8)
factual suitable
sachlich (Ab4) useful geeignet (Ab14)
nützlich (Ab9)
qualified knowledgeable
qualifiziert (Ab5) applicable kenntnisreich (Ab15)
brauchbar (Ab10)

Integrity/Honesty sincere credible serious

aufrichtig (In1) glaubwürdig (In5) seriös (In9)
genuine dependable correct
echt (In2) verlässlich (In6) richtig (In10)
honest authentic
ehrlich (In3) authentisch (In7)
true unambiguous
wahr (In4) widerspruchsfrei
(In8)

Benevolence caring selfless charitable

fürsorglich (Be1) selbstlos (Be4) gemeinnützig (Be7)
social altruistic just
sozial (Be2) uneigennützig (Be5) gerecht (Be8)
kind-hearted benevolent towards fair
gutherzig (Be3) me fair (Be9)
mir gegenüber
wohlwollend (Be6)

Likeability appealing likeable exciting

ansprechend (Li1) sympathisch (Li5) spannend (Li9)
pleasant entertaining nice
erfreulich (Li2) unterhaltsam (Li6) nett (Li10)
enjoyable attractive charming
angenehm (Li3) anziehnd (Li7) reizend (Li11)
positive amusing
positiv (Li4) lustig (Li8)

182
 Table 5 (Cont.): Initial Set of eWOM Trust Items

Sub-dimensions Items

Willingness to rely on When I want to purchase a specific product/service, I am willing to visit online
customer reviews to get relevant information.
Wenn ich ein bestimmtes Produkt/eine bestimmte Dienstleistung kaufen möchte, bin
ich bereit, Online Kundenrezensionen zu besuchen, um relevante Informationen zu
erhalten. (Wi1)
I’m willing to recommend products/services/vendors which have been
recommended in online customer reviews to my friends and family.
Ich bin bereit, Produkte/Dienstleistungen/Händler, die in Online
Kundenrezensionen empfohlen wurden, meinen Freunden oder Familienmitgliedern
weiterzuempfehlen. (Wi2)
I am willing to consider online customer reviews when I make a purchase decision.
Ich bin bereit, Online Kundenrezensionen zu berücksichtigen, wenn ich
Kaufentscheidungen treffe. (Wi3)
If I want to know more about the experiences of other customers with a
product/service/vendor, I am willing to look at online customer reviews.
Im Fall, dass ich mehr über die Erfahrungen anderer Konsumenten mit einem
Produkt/einer Dienstleistung/einem Händler erfahren möchte, bin ich bereit, Online
Kundenrezensionen zu besuchen. (Wi4)
I am prepared to consult online customer reviews in order to find out more about the
attributes of a product/service/vendor.
Ich bin bereit, Online Kundenrezensionen heranzuziehen, wenn ich mehr über die
Eigenschaften eines Produktes/einer Dienstleistung/eines Händlers erfahren
möchte. (Wi5)
I am willing to make relevant purchasing decisions based on online customer
reviews.
Ich bin bereit, wichtige Kaufentscheidungen aufgrund von Online
Kundenrezensionen zu treffen. (Wi6)
When I make purchasing decisions, I am willing to rely on online customer reviews.
Wenn ich Kaufentscheidungen treffe, bin ich bereit, mich auf Online
Kundenrezensionen zu verlassen. (Wi7)
I would use online customer reviews to find out more about a
product/service/vendor.
Ich würde Online Kundenrezensionen verwenden, um mehr über ein
Produkt/eine Dienstleistung/einen Händler herauszufinden. (Wi8)

Research Instrument

An online questionnaire was implemented by using the SoSci Survey

([Link] website. This online platform has been previously used by
various academic studies for data collection. The online survey format was chosen because
filling out a questionnaire on the Internet was deemed appropriate for the sample, that is,
persons who have some earlier experience with online customer reviews as well as using the
Internet for product research and/or online shopping. An online survey hence offers the benefit
of a close contextual match and offers an examination of the focal construct in its “natural
setting”.

183
On the introductory web page of the questionnaire, respondents were informed about the rough
content and intention of the survey. They were told that they would subsequently be asked a
number of questions targeting personal orientations towards online customer reviews in
general. On the same page, a definition of OCR as a market information institution and a typical
example were provided, in order to clarify the study’s context. On the subsequent pages,
respondents were asked to indicate their level of agreement to several Likert-Statements, all
being measured on a scale from 0=”I strongly disagree” to 6=”I strongly agree”. The different
sections of the questionnaire comprised a series of items concerning respondents’ earlier online
experiences, the new measurement instrument, additional eWOM trust proxies, and socio-
demographical variables. Specifically, the questionnaire included (i) four items taken from
Gefen (2000), Jamal and Naser (2002), and Corbitt et al. (2003), measuring respondents’
Internet experience (“I perceive myself pretty experienced in using the Internet.”; “I am familiar
with searching for (products) on the Internet.”; “I know a lot about conducting purchases via
the Internet.”; and “I am experienced in conducting purchases via the Internet”); (ii) four items
for online customer experience taken from Park et al. (2011) (“I always read online reviews
written by other shoppers.”; “I always write down online reviews by myself.”; “I always read
online customer reviews when I was shopping.”; and “I frequently use online customer reviews
to gather information prior making purchases.”); (iii) the 53 trust in online customer reviews
items which have been developed in the identification stage of this research; (iv) five additional
items for measuring eWOM trust in general (“The information given in online customer reviews
is trustworthy.”; “The information given in online customer reviews is reliable.”; “In general, I
trust the information given in online customer reviews.”; “It is safe to trust information given
in online customer reviews.”; “I have confidence in online customer reviews.”) adopted from
diverse literature (e.g., Soh, 2007); and (v) demographic-related questions, including the
respondent’s age, gender, university and study major. Students were also asked to provide their
student ID, as well as family name, in order to be awarded course credits. Here, they were
explicitly told that the data would be collected separately and wouldn’t be used for further
analyses of their person. Since this study’s main interest was to assess the construct’s structure,
accompanied by some initial evaluations of the items’ psychometric properties in terms of
reliability and validity, this research refrained from including additional measures of other
constructs, because of considerations of the questionnaire’s length.

Sample Description and Data Collection Procedure

Before data collection could take place, two critical decisions had to be made: (1) choosing the
adequate sample size and (2) determining the composition of the sample. In respect to the
former, scaling literature provides several, sometimes divergent, recommendations. In this
research, the adequate sample size had to be chosen mainly on the demands of the statistical
procedures that followed. EFA recommendations concerning the minimum sample size range
between 100 and 500 respondents (Comrey & Lee, 2002; Gorsuch, 1983). For instance, while

184
some researchers note that a sample of 100 to 200 persons will suffice (Clark & Watson, 1995),
others suggest that data should be collected in the range of n=300 (DeVellis, 2012). Netemeyer
et al. (2003) here state that for larger pools with more than 20 items and for complex or multi-
dimensional constructs, also larger sample sizes in the 300 range should be aimed for. Hence,
it is also important to consider how many items are in the initial pool. Here, however,
recommendations concerning the minimum ratio of the number or respondents to the number
of items in the scale also often differ and range from 3:1 to 10:1 (Cattell, 1978). In fact, the
sample size appropriate for EFA is dependent on several aspects, including the level of item
communality and factor determination (MacCallum et al., 1999).

In the same way as for EFA, rules of thumb for determining the adequate sample size for CFA
also vary. For example, some scholars suggest a minimum of 10 observations per parameter
estimated (Jöreskog & Sörbom, 1989), while others advocate lower standards of 5 to 10
observations per parameter estimated (Floyd & Widaman, 1995). Alternatively, Hair et al.
(2010) state that CFA should be based on a sample of 200 respondents or more. This said, the
tenor crystallizes that sample size should be large, as some evaluative criteria of CFA are more
likely to be affected by sample size than others (Netemeyer et al., 2003). Additionally, it is
regularly emphasized that if there are too few respondents for factor analysis, the scale
development process may be compromised (DeVellis, 2012). However, there are also some
hints that very large samples are likely to show also trivial differences between the observed
and the implied covariance matrices to be significant (Netemeyer et al., 2003), so an upper
bound is implicitly given. By considering this, the research at hand assumed that a minimum
sample size of 400 respondents for each of the sub-samples (3a and 3b) would fulfil the above-
discussed demands (giving an initial sample-to-item ratio of roughly 7:1).

Another critical issue in choosing a sample is considering how well the sample represents the
population for which the measure is intended (DeVellis, 2012; MacKenzie et al., 2011;
Netemeyer et al., 2003). In this research, drawing a convenience sample of college students was
regarded as sufficient. This decision was based on several reasons. First, as Internet users are
typically younger and well educated (Statistik Austria, 2012), data received from college
students are likely to be generalizable, at least for the most important segments of the Internet
population and a considerable degree of representativeness is hence given. Second, as outlined
earlier, in order to evaluate the distinct aspects of trust in eWOM, respondents had to have some
familiarity with online customer reviews. Several sources have indicated that younger and well-
educated consumers are more likely to use OCR and similar social media content on a regular
basis compared to other Internet users (eMarketer, 2011; Sensis, 2013). Students are now a
typical exemplar of this group, as they are members of the Generation Y or Digital Natives
cohort (Bolton et al., 2013). Third, various scholars point out that using data from college
students for pilot testing is appropriate (Netemeyer et al, 2003) and several studies applied
student sampling for scale development in the past (e.g., Liu, 2003; Obermiller & Spangenberg,

185
1998; Tian et al., 2001). Fourth, as the applied scale development procedures demand a vast
number of survey respondents, it is reasonable to ask college students, as they can be easily
accessed. Fifth, various studies – including investigations into the field of attitude towards
advertising (Yavas, 1994) – indicate that students are effective surrogates for non-students.

The student sample was conveniently drawn with no quota being made on demographics such
as gender and age. Specifically, data were collected from 878 undergraduate students, who were
recruited from two major Austrian universities in exchange for extra course credit. The
participants were told that they would participate in an academic research project that strives to
evaluate consumer perceptions of online customer reviews and that the participation is
voluntarily. The survey was conducted in January 2013 and the questionnaire was available
online for seven days. Data from 28 participants were eliminated from further analyses due to
missing data and/or obvious response bias (e.g., answering heuristics). The characteristics of
the remaining 850 respondents were as follows. The total sample consisted of 64.8% (551)
females and 34.2% (299) males with ages ranging from 18 to 35 years (M = 23.72, SD = 3.03).
This sample comprised two separate sub-samples with each participant being assigned to one
of them by random selection. Estimation sample 3a consisted of 425 undergraduate students
(62% female). The average age in this sample was 23.67 (SD = 2.98). Holdout sample 3b
consisted of the same number of undergraduate students (64% female). Here, the average age
was 23.78 (SD = 3.08). Students in both samples represented a similar variety of different
university backgrounds and college majors (mostly business students). The two groups were
regarded as comparable, as there were no significant differences in terms of gender, age,
Internet usage and attitude (M = 4.08 (SD = 1.16) vs. 4.09 (SD = 1.25)) and customer review
experience (M = 3.21 (SD = 1.33) vs. 3.31 (SD = 1.33)).

Validity Stage
According to pivotal scale development literature, the collection of additional data to enable
the cross-validation of the findings of the reliability stage, to refine the items and to finalize the
scale is an imperative (DeVellis, 2012; MacKenzie et al., 2011; Netemeyer et al., 2003).
Therefore, in the main study (study 4) a new data set was collected using a representative sample
of adult online consumers. This single study, even though its purpose was to assess multiple
aspects of validity, was not enough to confidently infer construct validity. Accordingly, a series
of supplemental studies (studies 5a-i) was conducted which enabled this research to evaluate
multiple applications of the new study over different samples and over time. This research
assumes that this procedure satisfies the demands typically raised in scale development
literature (Hildebrandt & Temme, 2006). By collecting various data sets, this study was also
able to execute extra reliability and validity/generalizability checks. Conducting numerous
studies for deriving and examining a new measure, as has been done in this thesis, is quite
common in scale development research (e.g., Netemeyer et al., 1996). In the remainder of this
186
section, the survey contents and the data collection procedures of the alternative studies are
presented.

Main Validation Study (Study 4)

Research Instrument

The purpose of the main study (study 4) was to reassess the scale’s and the individual items’
psychometric properties with new data gathered by means of an online survey (RQ1). In
addition, the study strived to answer several research questions targeting the scale’s validity.
Hence, various measures for the same and other constructs were included for validation
purposes. More specifically, as part of this data-gathering effort, alternative measures of the
same construct were included in order to test for convergent validity (RQ3). Also, discriminant
validity (RQ4) was assessed by considering measures of constructs similar to eWOM trust that
might be confused with the focal construct. Besides concurrent validity (RQ2), this survey also
accounted for some nomological aspects by including measures of constructs with which
eWOM trust was expected to be linked (i.e., antecedents, consequences, and correlates). Here,
asking respondents to evaluate additional items in the same questionnaire helped to test various
hypotheses about how the new scale relates to measures of other constructs and to profile
eWOM trusters.

The online survey platform SocSci Survey was again deemed adequate for implementing the
inquiry. On the front page, the participants were told that the following survey pertains to
personal orientations towards online customer reviews as an institution. The instructions also
included a definition of customer reviews in order to frame respondents adequately.
Additionally, the respondents were informed about the diverse forms of appearance of customer
reviews and an explicit but fictitious example for a typical online review was provided. It was
made clear that the respondents should not answer the questions in regard to this example, but
rather should evaluate their personal attitudes and general perceptions of OCR in general.
Before being guided to the actual questionnaire, respondents had to answer a screening question
(“Have you been in contact with online customer reviews in the last 6 months?”). It was
communicated that the term “contact” is here defined quite broadly, including any forms of
perceptions – ranging from consciously reading to coincidentally looking at online customer
reviews. Only respondents who answered this question with “yes” were forwarded to further
questions, while the others were excluded. This procedure was to ensure that respondents had
a minimum of knowledge concerning the object of trust and were familiar with the survey
context. Next, the participants were asked to provide some socio-demographic information,
including age, gender, country of living, marital status, education, employment area, and
income. As the survey intended to draw a representative sample of Internet users from two
Central-European, German-speaking countries, some of these variables have been used to
classify respondents: quotas have been imposed on age as well as gender (see below for the

187
sample characteristics); only respondents that had a minimum age of 16 and were qualified in
terms of the applied quotas were able to continue the actual survey.

The core questionnaire was structured in four subsequent parts: (i) Internet usage patterns; (ii)
perceptions of online customer reviews in general; (iii) perceptions of online reviewers; and
(iv) perceptions of others in general. Within these sections, several measures of various
constructs were included for validation purposes. These measures are described with greater
detail below. In general, respondents were forced to provide answers to each of these items.
Within each section, items were randomly mixed. The survey concluded by thanking the
participants and redirecting them to the webpage of the consumer panel provider. The main
questionnaire is available on [Link]. What follows is a description of the
operationalizations of the included constructs.

Measurement of eWOM Trust (eWOMTrust)

The questionnaire included several alternative means to measure consumer’s trust towards
online reviews. First, eWOM trust was measured by a refined set of Likert-like statements
which was subject to development in the course of this thesis (the eWT-S). Here, respondents
were asked to indicate their level of personal agreement for 25 items measuring the diverse
aspects of trust on a 7-point scale ranging from 0=”I strongly disagree” (“Ich stimme überhaupt
nicht zu”) to 6=”I strongly agree” (“Ich stimme völlig zu”). Second, for testing convergence
validity by means of a Multitrait-Multimethod (MTMM) matrix (Campbell & Fiske, 1959;
Churchill, 1979), eWOM trust was also measured using semantic differential scales. In the
absence of an established scale, the semantic differential scale was constructed by adopting the
Likert statements developed in this research. For instance, the original Likert item “The
information given in online customer reviews is credible” was altered to fit a uni-polar semantic
differential scale where the participants had to indicate the presence or absence of a single
attribute. Accordingly, respondents had to evaluate their perception of whether “The
information given in online customer reviews is credible” or “The information given in online
customer reviews is not credible” on a 7-Point rating scale. Third, eWOM trust was also
quantified by using a non-diagnostic single item. Here, the participants had to give an answer
to the question, “On a scale where 0 = “I don’t trust at all” and 6 = “I absolutely trust”, how
much do you trust the information given in online customer reviews?” (“Geben Sie auf bitte auf
einer Skala von 0 (ich habe überhaupt kein Vertrauen) bis 6 (ich habe völliges Vertrauen) an,
in wie weit Sie Online Kundenrezensionen im Allgemeinen vertrauen.”). This scale was adopted
from Boush et al. (1993) and was used in its regular application in the market information
context (e.g., Soh, 2007). Fourth, generalized trust in online customer reviews was finally also
measured by a multi-item reflective measure (see Table 6).

188
Table 6: Measures of Generalized eWOM Trust

Variable Item Source Scale

Label

The information given in online customer reviews is

trustworthy.
Tr1
Information in Online Kundenrezensionen sind
vertrauenswürdig.
The information given in online customer reviews is
reliable.
Tr2
Information in Online Kundenrezensionen sind 7-Point Likert Scale
zuverlässig. (0=”I stongly
adopted from disagree”/”Ich
In general, I trust the information given in online
Soh (2007); stimme überhaupt
customer reviews.
Tr3 new items nicht zu”, 6=”I
Im Allgemeinen vertraue ich Informationen in Online
strongly agree”/”Ich
Kundenrezensionen.
stimme völlig zu”)
It is safe to trust information given in online customer
reviews.
Tr4
Es ist sicher, Informationen in Online
Kundenrezensionen zu vertrauen.
I have confidence in online customer reviews.
Tr5
Ich habe Zutrauen in Online Kundenrezensionen.

Measurement of Attitude towards Reviews in General (RAtt)

The respondent’s general attitude towards eWOM was quantified with two alternative scale-
types, namely a Likert-format scale and a semantic differential scale. This enabled the author
to make an assessment of the construct’s discriminance in various ways, including the usage of
an MTMM matrix. Given the absence of reliable and validated scales in the researched area,
both scales were taken from prior research in the field of advertising, where the original scales
and adaptations found regular application (e.g., Petrovici & Marinov, 2007; Tan & Chia, 2007).
Minor changes had to be made to both scales in order to fit the research context. Table 7 shows
the items being used to measure consumer attitude towards online reviews. The first R Att scale,
originally introduced by Pollay and Mittal (1993), consisted of seven Likert-format scale items
which emphasized consumer considerations in respect to the functional value of reviews, as
well as the respondent’s emotional stance. All participants were asked to rate each of the seven
statements on a seven-part scale ranging from 0 (“I strongly disagree”) to 6 (“I strongly agree”).
The second scale was adapted from Olney et al. (1991) and included twelve semantic
differential items. Persons’ responses were measured on a 7-point rating scale. Both scales are
essentially striving to measure both the affective as well as cognitive aspect of an individual’s
general attitude towards eWOM. While other scales in OCR research typically focus on either
of them, this research assumes that potential similarities are highlighted more clearly with this
approach.

189
Table 7: Measures of Attitude towards Reviews in General

Variable Item Source Scale

Label

Online customer reviews are essential.

AL1
Online Rezensionen sind für mich wesentlich.

Online customer reviews are a valuable source of

information about products.
AL2
Online Rezensionen sind eine wertvolle Quelle für
Produktinformationen.

In general, online customer reviews are misleading.

AL3
Im Allgemeinen sind Online Rezensionen irreführend. (R)
7-Point Likert Scale
Online customer reviews tell me which products have the
(0=”I stongly
features I am looking for. adopted
AL4 disagree”/”Ich
Online Rezensionen sagen mir, welche Produkte das from Pollay
stimme überhaupt
bieten, was ich suche. and Mittal
nicht zu”, 6=”I
(1993)
strongly agree”/”Ich
In general, online customer reviews present a true picture stimme völlig zu”)
of the product discussed.
AL5
Im Allgemeinen zeigen Online Rezensionen ein wahres
Bild des Produkts.

Overall, I consider online customer reviews a good thing.

AL6
Ich halte Online Rezensionen für eine gute Sache.

My general opinion of online customer reviews is

unfavourable.
AL7
Meine allgemeine Einstellung gegenüber Online
Rezensionen ist negativ. (R)

190
Table 7 (Cont.): Measures of Attitude towards Reviews in General

Variable Item Source Scale

Label

Overall, online customer reviews are unpleasant –

pleasant
AS1
Online Rezensionen sind im Allgemeinen
unangenehm - angenehm

fun to read – not fun to read

AS2
lustig zu lessen – nicht lustig zu lessen (R)

not entertaining – entertaining

AS3
nicht unterhaltsam – unterhaltsam

enjoyable – not enjoyable

AS4
amüsant – nicht amüstant (R)

important – not important

AS5
wichtig – unwichtig (R)

informative – uninformative
AS6 adopted
informativ – nicht informativ (R) 7-Point Semantic
from Olney
Differential Scale
et al. (1991)
helpful – not helpful
AS7
hilfreich – nicht hilfreich (R)

useful – not useful

AS8
nützlich – nicht nützlich (R)

makes me curious – don’t make me curious

AS9
macht mich neugierig – macht mich nicht neugierig (R)

not boring – boring

AS10
nicht langweilig – langweilig (R)

interesting – not interesting

AS11
interessant – nicht interessant (R)

keep my attention – don’t keep my attention

AS12 haben meine Aufmerksamkeit – haben nicht meine
Aufmerksamkeit (R)

191
Measurement of Review Credibility (RCred)

Similarly to review attitude, respondents’ opinions concerning the general credibility of eWOM
was measured with two maximally different scales: a Likert-format scale and a semantic
differential scale. Here, participants were asked to answer a series of questions, including five
7-Point Likert format statements derived from the research of Boush et al. (1994) and a set of
seven semantic differential format questions adapted from Beltramini and Evans (1985). Both
scales are well established in advertising research and only needed slight adaptation to fit the
current context. They measure a uni-dimensional construct that may also be called eWOM
believability or honesty. Table 8 includes the items used to quantify generalized review
credibility.

Table 8: Measures of Review Credibility

Variable Item Source Scale

Label

Online customer reviews tell the truth.

CL1
Online Rezensionen sagen die Wahrheit.

I can believe what people write in reviews.

CL2 Ich glaube was Konsumenten in Online Rezensionen
schreiben.

The products recommended in online customer reviews 7-Point Likert Scale

are always the best products to buy. (0=”I stongly
CL3
Die in Online Rezensionen empfohlenen Produkte sind adopted disagree”/”Ich
auch diejenigen, die man kaufen sollte. from Boush stimme überhaupt
et al. (1994) nicht zu”, 6=”I
I can depend on getting the truth from most online strongly agree”/”Ich
customer reviews. stimme völlig zu”)
CL4
Ich kann mich darauf verlassen, dass ich die Wahrheit in
den moisten Rezensionen erfahre.

If a customer review were not true, it would not have been

written.
CL5
Es warden niemals unwahre Behauptungen in
Rezensionen aufgestellt.

192
Table 8 (Cont.): Measures of Review Credibility

Variable Item Source Scale

Label

Overall, online customer reviews are trustworthy –

untrustworthy
CS1
Online Rezensionen sind im Allgemeinen
nicht vertrauenswürdig - vertrauenswürdig

convincing – not convincing

CS2
überzeugend – nicht überzeugend (R)

not credible – credible

CS3
unglaubwürdig - glaubwürdig
adopted from
7-Point Semantic
Beltramini and
reasonable – unreasonable Differential Scale
CS4 Evans (1985)
vernünfigt – unvernünftig (R)

unquestionable – questionable
CS5
unstrittig – fragwürdig (R)

authentic – not authentic

CS6
echt – unecht (R)

dishonest – honest
CS7
unehrlich - ehrlich

193
Measurement of Reviewer Credibility (SCred)

The questionnaire also included the possibility to measure respondents’ attitude towards the
credibility of review creators. Reviewer credibility was measured with nine items adapted from
the established scale proposed by Ohanian (1991) on a 7-Point semantic differential scale.
Ohanian’s scale was previously used to quantify credibility in the context of OCR (e.g., Dou et
al., 2012). The scale assesses the two basic components of source credibility regularly identified
in literature: reviewer expertise and reviewer trustworthiness. Table 9 presents the items
included in the questionnaire.

Table 9: Measures of Reviewer Credibility

Variable Item Source Scale

Label

Consumer writing online reviews are

undependable – dependable
SC1
Konsumenten, die Rezensionen online verfassen
sind nicht verlässlich - verlässlich

dishonest – honest
SC2
unehrlich - ehrlich

unreliable – reliable
SC3
nicht zuverlässig - zuverlässig

sincere – insincere
SC4
aufrichtig – nicht aufrichtig (R)
adopted from 7-Point Semantic
untrustworthy – trustworthy Ohanian (1991) Differential Scale
SC5
nicht vertrauenswürdig - vertrauenswürdig

inexperienced – experienced
SC6
unerfahren - erfahren

knowledgeable – unknowledgeable
SC7
wissend – unwissend (R)

qualified – unqualified
SC8
qualifiziert – unqualifiziert (R)

unskilled – skilled
SC9
nicht sachkundig - sachkundig

194
Measurement of Reviewer Distrust (SDistrust)

Another important antecedent of eWOM trust is distrust in online reviewers in general.

Attempts to operationalize the distrust construct are extremely rare in other trust contexts, and
non-existing in the eWOM context. However, based on the conceptualization earlier advanced
in this thesis, the author is confident that an adopted version of the scale introduced by Boush
et al. (1994) captures the nature of the distrust construct quite reasonably. Therefore, the
respondents were asked to indicate their personal level of distrust towards the motives of
reviewers and recommenders on a set of four 7-Point Likert format statements (see Table 10).
The scale mainly captures the extent to which eWOM recipients think that reviewers typically
intend to manipulate them and their purchasing decisions.

Table 10: Measures of Distrust in Online Reviewers‘ Motives

Variable Item Source Scale

Label

Online reviewers care more about getting you to buy

things than what is good for you.
SD1 Online Rezensenten wollen eher Konsumenten dazu
bringen bestimmte Dinge zu kaufen, als Konsumenten
gut zu informieren.
I often notice tricks that online reviewers play to get
me to buy something. 7-Point Likert
SD2
Ich bemerke oft Tricks von Online Rezensenten, die Scale (0=”I stongly
mich zu einem Kauf animieren wollen. disagree”/”Ich
adopted from
Online customer reviews try to make people buy stimme überhaupt
Boush et al.
things they don’t really need. nicht zu”, 6=”I
(1994)
SD3 Online Rezensionen sollen Konsumenten dazu strongly
veranlassen Dinge zu kaufen, die sie eigentlich nicht agree”/”Ich stimme
brauchen. völlig zu”)

Online customer reviews are different from other

product information on the Internet in the way they try
to influence you.
SD4
Online Rezensionen sind anders als übrige
Produktinformatiionen im Internet da sie versuchen
den Konsumenten stark zu manipulieren.

195
Measurement of Review Usage (RUse)

For measuring the respondents’ different forms of review usage, they were asked to complete
a set of six 7-Point statements. The rating scales were anchored from 0 (“never”) to 6 (“always”)
and included items concerning the dependence on reviews for important purchasing decisions,
actual purchase influence and eWOM approaching behaviour (e.g., adoption, persuasion), but
also behaviours such as the respondents’ own WOM engagement. All measures were adapted
from a variety of established scales, including the work of Obermiller and Spangenberg (1998),
Park and Lee (200), Sharvitt et al. (1998), Chu and Kim (2011), and Bailey (2005). Table 11
presents the items used to quantify general review usage.

Table 11: Measures of Online Review Usage

Variable Item Source Scale

Label

I use information from online customer

reviews for important purchase decisions.
RU1 Ich verwende Informationen von Online
Rezensionen, um wichtige
Kaufentscheidungen zu treffen.
I am influenced by online customer reviews
when I choose a product.
RU2 Ich werde von Online Rezensionen
beeinflusst, wenn ich Produkte im Internet
wähle. adopted from
I rely on online customer reviews when I Obermiller and
purchase products. Spangenberg (1998);
RU3 Park and Lee (2009); 7-Point Likert Scale
Ich verlasse mich auf Online Rezensionen,
wenn ich Produkte im Internet kaufe. Sharvitt et al. (1998); (0=”never”/”niemals”,
Chu and Kim (2011); 6=”always”/”immer”)
Online customer reviews crucially affect Bailey (2005)
my choice of products.
RU4
Online Rezensionen beeinflussen meine
Produktwahl wesentlich.
I like to pass along interesting information
from online customer reviews to others.
RU5
Ich gebe gerne interessante Informationen
von Online Rezensionen an andere weiter.
I visit product review websites prior to
online purchases.
RU6
Ich besuche ein Online Bewertungsportal
bevor ich ein neues Produkt kaufe.

196
Measurement of Review Avoidance (RAvoid)

General review avoidance was assessed by presenting all participants a series of five statements,
inviting the respondent to answer them on a rating scale ranging from 0 (“never”) to 6
(“always”). All items were adapted from the scales advanced by Obermiller and Spangenberg
(1998), Speck and Elliott (1997), and Park et al. (2007). The items include active avoidance
behaviour, as well as mental resistance expressions. Table 12 lists the items used to measure
review avoidance.

Table 12: Measures of Online Review Avoidance

Variable Item Source Scale

Label

I avoid reading online customer reviews.

RA1 Ich vermeide es, Online Rezensionen zu
lesen.
I find online customer reviews irritating as
the information provided is misleading and
unreliable.
RA2
Wenn ich Online Rezensionen betrachte,
finde ich diese irritierend weil sie
irreführend und unzuverlässig sind.
adopted from Obermiller 7-Point Likert Scale
When I buy a product online, reading the and Spangenberg (1998); (0=”never”/
reviews presented on the website impose a Speck and Elliott (1997); ”niemals”,
burden on me. Park et al. (2007) 6=”always”/”immer”)
RA3
Wenn ich online ein Produkt kaufe, stellt
das Lesen von Online Rezensionen für
mich eine Last dar.
When I buy a product online, reading the
reviews presented on the website irritates
me.
RA4
Wenn ich online ein Produkt kaufe, ärgert
es mich, wenn ich Online Rezensionen
lese.

197
Measurement of Review Involvement (RInv)

The review involvement measures were adapted from the established work of Zaichkowsky
(1994). Every participant had to indicate his/her personal involvement with online customer
reviews and recommendations on this 9-item 7-Point semantic differential scale (see Table 13).

Table 13: Measures of Review Involvement

Variable Item Source Scale

Label

Online customer reviews are

important - unimportant
RI1
Für mich sind Online Kundenrezensiionen
wichtig – unwichtig (R)
boring - interesting
RI2
langweilig – interessant
relevant - irrelevant
RI3
relevant – irrelevant (R)
exciting - unexciting
RI4
aufregend – nicht aufregend (R) adopted from
7-Point Semantic
Zaichkowsky
mean nothing – mean a lot to me Differential Scale
RI5 (1994)
bedeuten mir nichts – bedeuten mir viel
appealing - unappealing
RI6
ansprechend – nicht ansprechend (R)
fascinating - mundane
RI7
faszinierend – banal (R)
worthless - valuable
RI8
wertlos – wertvoll
not needed - needed
RI9
von mir gewollt – von mir nicht gewollt (R)

198
Measurement of Disposition to Trust (TDispo)

As this research intended to assess the impact of disposition to trust on the tendency to trust
eWOM, the personality trait construct was operationalized in the questionnaire with a set of
nine 7-Point Likert statements ranging from 0 (“I strongly disagree”) to 6 (“I strongly agree”).
The items were taken from McKnight et al. (2002a), as well as Gefen (2000), and aimed to
quantify the respondents’ general faith in humanity as well as their trusting stance. Table 14
presents these items.

Table 14: Measures of Disposition to Trust

Variable Item Source Scale

Label

The typical person is sincerely concerned about the

problems of others.
DP1
Menschen sind normalerweise ernsthaft um die
Probleme anderer besorgt.
Most of the time, people care enough to try to be
helpful, rather than just looking out for themselves.
DP2
Meistens versuchen Menschen anderen zu helfen,
anstatt sich nur um sich selbst zu kümmern.
In general, most folks keep their promises.
DP3 Im Allgemeinen erfüllen die moisten Menschen ihre
Versprechen.
Most people are honest in their dealings with others.
DP4 Die meisten Menschen sind im Umgang mit anderen McKnight et al.
ehrlich. (2002a) 7-Point Likert Scale
(0=”I stongly
I usually trust people until they give me a reason not disagree”/”Ich
to trust them. stimme überhaupt
DP5
Ich vertraue Menschen normalerweise bis sie mir nicht zu”, 6=”I
Grund geben dies nicht mehr zu tun. strongly agree”/”Ich
stimme völlig zu”)
I generally give people the benefit of the doubt when
I first meet them.
DP6
Wenn ich jemanden zum ersten Mal treffe, glaube ich
ihm/ihr fürs Erste.
My typical approach is to trust new acquaintances
until they prove I should not trust them.
DP7 Es ist meine Einstellung neuen Bekannten zu
vertrauen bis ich einen Grund habe, dies nicht mehr
zu tun.
I generally trust other people.
DP8
Im Allgemeinen vertraue ich anderen Menschen.
Gefen (2000)
I tend to count upon other people.
DP9
Ich neige dazu, mich auf andere zu verlassen.

199
Measurement of Review Experience (RExp) and Internet Experience (IExp)

Table 15 and 16 show the items used to measure participants’ experience with the Internet in
general and particularly with online reviews. The statements include a variety of active and
passive behaviours and were all taken from scales used in prior research.

Table 15: Measures of Review Experience

Variable Item Source Scale

Label

I read reviews written by fellow shoppers.

RE1 Ich lese Online Rezensionen von anderen
Konsumenten.
I search customer reviews on the Internet.
RE2
Ich suche nach Kundenrezensionen im Internet. Park and Lee
(2009)
I refer to online customer reviews whenever I need
7-Point Likert Scale
information on companies or goods.
(0=”never”/
RE3 Ich besuche Online Rezensionen, wenn ich
”niemals”,
Informationen über Produkte oder Unternehmen
6=”always”/”immer”)
benötige.
I write down online reviews by myself.
RE4A
Ich schreibe selbst Online Rezensionen
Park et al.
I read online customer reviews when I was shopping. (2011)
RE5 Ich lese Online Rezensionen bevor ich ein neues
Produkt im Internet kaufe.

Table 16: Measures of Internet Experience

Variable Item Source Scale

Label

I have been using the Internet for … (less than 1 year / between 1 and 2
Corbitt et al.
IE1 Ich verwende das Internet seit … years / between 2 and 3 years/ between
(2003)
3 and 5 years / 5 years or more)
I use the Internet approximately … (Less than 1h per week / between 1
Ich verwende das Internet Corbitt et al. and 3h per week / between 3 and 10h
IE2
durchschnittlich … (2003) per week / between 10 and 20h per
week/more than 20h per week)
How often do you research
products/services on the Internet that
adopted from (never / once a year / once a month /
you are interested in buying?
IE3 Murphy and several times a month / several times a
Wie oft suchen Sie im Internet
Tocher (2011) week / almost every day)
Informationen über Produkte, die Sie
eventuell gerne kaufen möchten?
Total number of online purchase
adopted from
during the last 3 months …
IE4 Lee et al. (0 / 1 / 2-5 / 6-10 / 11 or more)
Wieviele Einkäufe haben Sie in den
(2011)
letzten 3 Monaten online getätigt?
How often do you read online
adopted from (never / once a year / once a month /
customer reviews?
IE5 Murphy and several times a month / several times a
Wie oft lessen Sie Online
Tocher (2011) week / almost every day)
Kundenrezensionen?

200
Questionnaire Pretest

For enhancing the quality of the questionnaire, a pretest was administered to a convenient
sample of 18 undergraduate marketing students enrolled at a major Austrian university. The
student sample was conveniently drawn with no quotes on demographics, such as age or gender.
Respondents received course credits in return for their participation. The pretest was executed
in June 2013, when the online questionnaire was available for one week. As in the later study,
respondents were first prompted by several instructions, including a definition and examples of
eWOM, and proceeded to fill out various questions concerning the various constructs included
in this study. More specifically, all students were asked to go critically over the questionnaire
and to make comments on any apparent mistakes, unclear questions, and ideas of improvement.
Additionally, the questionnaire was also subject to a review by five members of the marketing
faculty. Due to these suggestions, several minor modifications have been made, such as
corrections of spelling mistakes and refinements of wording. Overall, the questionnaire was
deemed appropriate by all respondents.

Sampling Frame and Data Collection

In the main study, a total of 558 completed questionnaires were collected by means of a well-
known online consumer panel, Research Now ([Link] The company
offers a variety of consumer and industrial panels characterized all by consistency,
responsiveness and reliability. Its members are geographically as well as demographically
diverse. Research Now provides access to consumer panels in various countries, including
Austria, Germany, and the United States. The unit of analysis for the main study was the
individual online consumer. Research Now grants access to a consumer sample that is
representative of the Internet population in respect to various characteristics. Earlier studies
have investigated consumers’ behaviour of persons aged 16 and older. Because of this fact, this
study deems people above the age of 16 as proper respondents. Furthermore, in Central Europe,
most Internet users are less than 75 years old (Statistik Austria, 2014; Statistisches Bundesamt,
2014). Therefore, consumers between the ages of 16 to 75 were sampled through Research
Now.

The aim of the main study was to draw a sample that is representative for online consumers
(i.e., persons who use the Internet for shopping purposes such as information seeking and
purchasing) in terms of gender and age. Additionally, this study intended to ensure that
psychometrics of the new scale are not biased by cultural aspects. The fact that Research Now
offers an international online consumer panel, individuals from Austria and Germany were
recruited. In both countries, the majority of individuals use the Internet for shopping (2013:
Austria – 54% (Statistik Austria, 2013); Germany – 65% (Bitkom, 2013)) or product research.
Quota sampling for these two countries was applied (Table 17 presents information on the
Austrian and German Internet population taken from Statistik Austria (2012) and Statistisches

201
Bundesamt (2014). The main study took place at the end of July 2013, when the questionnaire
was available for five days (July 30th, 2013 – August 3rd, 2013). Research Now solicited
respondents based on this research’s demographic requirements. Only qualified panellists
received an invitation to participate and a link to the questionnaire from the panel provider.
After the introductory page, qualified respondents further found several filter questions
regarding their former review usage behaviour, as well as their demographic profile. These
filter questions were used in order to screen out disqualified panellists who did not have recent
contact with the trust object, as well as those who showed improper demographic
characteristics. Additionally, this section allowed the selection of respondents who complied
with the study’s quotas. Hereafter, appropriate respondents were directed to the questionnaire
described above.

Return Rate

Within the assessment period, 1,131 panellists accessed the online survey. On each day,
invitations were sent to a fresh sample of individuals who had the desired characteristics of the
missing quotas. In total, about 1,500 invitations were sent. 369 respondents were screened out
due to low familiarity with online customer reviews, while the access to the survey was denied
to 177 panellists because of already-full quotas. 27 individuals aborted answering the questions
in the course of the survey. No specific exit pattern could be identified. The survey’s overall
response rate was (approximately) 37 %. From the 558 persons who completed the
questionnaire, data of 32 panellists were later excluded from further analysis because of poor
data quality (e.g., abnormalities in time used to answer the questionnaire, applied answering
heuristics) resulting in 269 (Austria) and 257 (Germany) usable questionnaires.

Sample Description

Table 17 shows the composition of the final sample. The Austrian (German) survey participants
were similar to the Austrian (German) Internet population in terms of gender and age groups.
Therefore, representativeness of the sample is assumed. For further analysis, the data obtained
from the two countries was pooled. Table 18 provided further insight into the sample
characteristics.

202
Table 17: Sample Composition of Study 4

Austria Germany
Total Sample
(n=526) Internet Sample Internet Sample
Age (yrs)
Population1 (n=269) Population2 (n=257)
male female male female male female male female male female
16-24 52 44 27 22 25 22
% 9.9% 8.4% 9.0% 8.8% 10.0% 8.2% 10.4% 9.6% 9.7% 8.6%
25-34 48 49 27 31 21 18
% 9.1% 9.3% 10.3% 10.5% 10.0% 11.5% 8.8% 8.1% 8.2% 7.0%
35-44 63 63 26 28 37 35
% 12.0% 12.0% 11.2% 10.9% 9.7% 10.4% 14.2% 13.1% 14.4% 13.6%
45-54 57 55 31 29 26 26
% 10.8% 10.5% 11.6% 10.0% 11.5% 10.8% 9.3% 8.6% 10.1% 10.1%
55-74 54 41 29 19 25 22
% 10.3% 7.8% 10.5% 7.2% 10.8% 7.1% 9.3% 8.6% 9.7% 8.6%
Sum n 274 252 140 129 134 123
Sum % 100% 100% 100% 100% 100%
Sources: 1 = Statistik Austria (2012); 2 = Statistisches Bundesamt (2014)

Table 18: Respondent Characteristics of Study 4

Austrian
Total Sample German Sample
Sample Characteristics Sample
(n= 526) (n= 257)
(n= 269)
Gender
Female 252 (47.9%) 129 (48.0%) 123 (47.9%)
Male 274 (52.1%) 140 (52.0%) 134 (52.1%)
Age
Range 16 – 74 16 – 74 16 – 73
Median 39.0 39.0 39.0
Mean 40.0 40.0 40.0
SD 13.9 14.2 13.6
Education
Compulsory school 29 (5.5%) 23 (8.6%) 6 (2.3%)
Junior high school; 220 (42.0%) 107 (39.8%) 113 (44.0%)
Professional school
High school 141 (26.8%) 86 (32.0%) 55 (21.4%)
University; College 134 (25.5%) 52 (19.3%) 82 (31.9%)
Other 2 (0.4%) 1 (0.4%) 1 (0.4%)

203
Table 18 (Cont.): Respondent Characteristics of Study 4

Austrian
Total Sample German Sample
Sample Characteristics Sample
(n= 526) (n= 257)
(n= 269)
Profession
Student 69 (13.2%) 34 (12.6%) 35 (13.6%)
White-collar worker; 240 (45.6%) 117 (43.5%) 123 (47.9%)
Public servant
Blue-collar worker 46 (8.7%) 17 (6.3%) 29 (11.3%)
Self-employed 59 (11.2%) 33 (12.3%) 26 (10.1%)
Retiree 52 (9.9%) 33 (12.3%) 19 (7.4%)
Unemployed 40 (7.6%) 24 (8.9%) 16 (6.2%)
Other/Rather not say 20 (3.8%) 11 (4.0%) 9 (3.5%)
Net Monthly Income
< 500 EUR 34 (6.5%) 21 (7.8%) 13 (5.1%)
501 – 800 EUR 41 (7.8%) 22 (8.2%) 19 (7.4%)
801 – 1,100 EUR 52 (9.9%) 22 (8.2%) 30 (11.7%)
1,101 – 1,400 EUR 59 (11.2%) 31 (11.5%) 28 (10.9%)
1,401 – 1,700 EUR 69 (13.1%) 36 (13.4%) 33 (12.8%)
> 1,700 EUR 170 (32.3%) 77 (28.6%) 93 (36.2%)
Rather not say 101 (19.2%) 60 (22.3%) 41 (16.0%)
Marital Status
Single 152 (28.9%) 81 (30.1%) 71 (27.6%)
Married; Living in a 305 (58.0%) 151 (56.1%) 154 (59.9%)
relationship
Divorced; Separated; 57 (10.8%) 31 (11.5%) 26 (10.1%)
Widowed
Rather not say 12 (2.3%) 6 (2.3%) 6 (2.3%)
Passive Review
Experience
Median 4.5 4.2 4.3
Mean 4.3 4.3 4.5
SD 1.1 1.2 1.1
Active Review
Experience
Median 2.6 2.6 2.7
Mean 3.0 3.0 3.0
SD 2.0 1.9 2.0

204
Supplemental Reliability, Validity and Generalizability Studies (Study 5)

In the course of this research, additional studies (5a-i) were executed which targeted different
aspects of validity – including known-group validity (RQ5), predictive/postdictive validity
(RQ6) – test-retest reliability (RQ7) and also generalizability issues. The latter emphasized the
assessment of the applicability of the new scale to different forms or platforms of eWOM (RQ8-
9), as well as the investigation into the determinants of situational eWOM trust (RQ10-11). This
set of studies also provided the necessary data to cross-validate the findings of the main study
once again and test the relationship of the focal construct with its causes, correlates, and
consequences. Hence, the majority of the proposed hypotheses were tested with empirical data
obtained from these studies. This helped to validate the construct’s measure in a nomological
network (i.e., nomological validity) – regularly demanded by various researchers (e.g.,
Netemeyer et al., 2003). Most of the supplemental studies were conducted in the form of street-
inverviews using paper-and-pencil questionnaires. The Austrian Internet population was used
as a sampling frame and all participants were selected due to the same quotas on age and gender
as in the main study. The respondent samples were conveniently drawn at various places
throughout the country. The details on the research methodology as well as on the key sample
characteristics are presented in Chapter 5, together with the research results.

Application Stage
The purpose of the sixth study was to cross-validate an English version of the eWOM trust scale
(RQ12) and to introduce a new typology of online consumers considering their trust in
eWOM/Online advertisements (RQ13). It was concluded that the aforementioned objectives
will be best achieved by surveying a representative, adequately sized sample of US online
consumers. Therefore, an online survey among members from the Research Now panel was
conducted. This study shared some reasonable similarities with this research’s main study
concerning its execution and structure. For instance, when accessing the survey, respondents
first faced an introductory page which informed them about the rough content of the following
questionnaire and the privacy policy. On the same page, a similar description of online reviews
and some examples were also provided. Before being able to answer the actual questionnaire,
respondents were prompted with a question concerning their review/online ads habits within
the last 6 months. On the following page, subjects were asked to answer a series of socio-
demographic questions. The following pages were only accessible to persons who had recent
contact with eWOM/Online advertising and who had the desired socio-demographic
characteristics. Similarly to the main study, the US study also enforced quotas on the
respondents’ age and gender which correspond to the demography of the US Internet population
(US Census Bureau, 2011): age group 18-34 yrs (male: 17%, female: 18%), 35-44 yrs (male:
9%, female: 10%), 45-64 yrs (male: 17%, female: 18%), 65 yrs and older (male: 5%, female
6% of total population).

205
The questionnaire was structured as follows: (i) Internet usage behaviour with items taken from
Corbitt et al. (2003), Barnes et al. (2007), and Lee et al. (2011), (ii) perceptions and attitudes
towards online customer reviews in general. This section included measurement items for
quantifying the constructs of attitude toward reviews (with 3 items taken from Park and Kim
(2008)), review usefulness (4 items adapted from Park and Lee (2009)), review avoidance (3
items mainly adapted from Park et al. (2007)), and perceived review usage risk (5 items adapted
from Jarvenpaa and Tractinsky (1999)).

Another part was dedicated to the new eWOM trust scale. As the items were originally
developed in German, they were subject to a rigorous forward-backward translation process.
Four persons who had American English as their mother tongue and were proficient in German
(all academics from two Austrian universities) were asked to translate the items in the target
language. For every item, they were supposed to provide at least one word or phrase that
captures the nature of the item, but they were also able to provide two optional translations.
After collecting the answers, two researchers identified recurring word and phrases. They
agreed on 100% of the items. Finally, two other persons being fluent in English and German
translated the English version again back to German. Again, there was a high agreement on the
meaning of the terms. eWOM trust was measured, as the items mentioned before, on a 7-Point
Likert scale ranging from 0 (“I strongly disagree”) to 6 (“I strongly agree”). Finally, the
questionnaire’s second section also included items measuring respondents’ earlier eWOM
experiences and habits. These items were adapted from the research of Park and Lee (2009),
Obermiller and Spangenberg (1998), and Chu and Kim (2011). Here, a 7-Point rating scale was
used (0 = “never”; 6 = “always”).

The next section (iii) addressed consumers’ attitudes towards the Internet. More specifically,
respondents were asked to answer 4 items taken from Morse et al. (2011) measuring Internet
attitude and 3 items adopted from Barnes et al. (2007) measuring online shopping attitude. All
were measured on a 7-Point Likert scale from 0 (“I strongly disagree”) to 6 (“I strongly agree”).
In addition, measures of active and passive usage behaviours of Web 2.0 applications were
included. Section (iv) targeted perceptions and attitudes towards online advertising in general.
Here, online advertising trust was measured with 9 items (taken from Obermiller and
Spangenberg (1998)) and online advertising attitude with 7 items (taken from Bauer and
Greyser (1968)). (v) Perceptions of business ethics and market offerings in general. The 10-
item scale for consumer alienation was adopted from Burns (2010), consumer attitude towards
product quality was measured with 5 items introduced by Ferdous and Towfique (2008), 2 items
taken from Lyonski et al. (1996) were used to measure product price attitude. The final section
(vi) focused on selected consumer psychographics including disposition to trust (5 items taken
from Gefen (2000)), self esteem (4 items taken from Rosenberg (1965)), self-confidence (12
items taken from Bearden et al. (2001)), and CSII (8 items taken from Bearden et al. (1989).

206
All constructs were measured on a 7-Point Likert scale with 0 (“I strongly disagree”) and 6 (“I
strongly agree”). As in the main study, the order of the items was randomized in each section.

In total, 1220 panellists visited the online survey. Of these, 635 persons were excluded from
the study due to unfamiliarity with the trust objects (eWOM and online advertising) or full
quotas. During the time the questionnaire was available (July 30, 2013 – August 2, 2013), 547
participants completed the survey. Subsequent review of data led to the exclusion of several
respondents and a total number of 517 usable responses suitable for data analysis. The socio-
demographic profile of the total sample was as follows. 53.2 % of the sample were females.
The respondents’ age ranged from 18 to 79 yrs (M = 43.9, SD = 15.0). The majority were
married or lived together with another person (62 %). Singles compromised 24.8 % of the
sample. Almost three-fourths of the sample (72.9 %) had an annual income below USD 75,000.
The majority of the sample were office workers or public servants (36.6 %), followed by retirees
(16.9 %), and blue-collar workers (13.8 %). European Americans represented 81.5 % of the
total sample (African American: 8.2 %; Asians: 5.1 %; Hispanics: 3.9 %; Native Americans:
1.4 %). Almost a third had a 4-year college degree (31.6 %) and 16.4 % a master’s degree. 48.7
% of the sample lived in the suburban area, while the remainder reside in urban (27.6 %) or
rural (23.7 %) areas. Further sample characteristics are provided in Chapter 5.

The next chapter discusses the findings of the surveys outlined above.

207
Chapter 5

Research Results

The following chapter discusses the results of multiple empirical studies which targeted the
development of the new eWOM trust scale (eWT-S), as well as its application. More
specifically, what follows are the details about the diverse assessments of the scale’s
psychometric properties, including its dimensionality and structure, its internal consistency,
validity (e.g., concurrent-, convergent-, discriminant-, nomoglogical validity) and the scale’s
reliability and generalizability across different contexts and situations. Here, amongst other
points, strong evidence is provided that an English version of the new scale is applicable to a
representative sample of US consumers. After providing norms for the eWT-S based on several
representative samples and scale finalization, the results of the segmentation approach are put
forward to describe a typology of four distinct groups of US online consumers. These differ in
respect to their generalized trust in two major sources of online marketing communication; that
is, online customer reviews as well as online advertising.

Assessment of Construct Dimensionality and Consistency

In earlier chapters, eWOM trust has been theorized to be a complex, second-order construct
consisting of several discriminating trust aspects. According to literature and qualitative scale
development procedures, five sub-dimensions which are also assumed to mirror cognitive,
affective as well as conative elements of trust, are most likely to comprehensively capture the
nature of trust in OCR. During the identification stage, 53 items were identified that are
supposed to capture the diverse eWOM sub-dimensions. A critical question is whether the
hypothesized structure of the construct is also evident in empirical data. In order to empirically
verify the adequateness of the advocated operationalization, the following multi-part research
question was proposed:

209
© Springer Fachmedien Wiesbaden 2017
W. Weitzl, Measuring Electronic Word-of-Mouth Effectiveness,
DOI 10.1007/978-3-658-15889-7_5
 RQ 1: (a) Is eWOM trust a complex, higher-order construct best described by several
facets or dimensions? (b) Are cognitive, affective, and conative elements suited to
measure consumers’ generalized tendency to trust customer reviews?

In order to improve reliability in this research’s results, data for answering this question were
gathered from diverse samples during the sequential stages of the research process. Initial data
was collected during the reliability stage by means of an online survey among university
students. This process enabled this research to initially assess the psychometric properties of
the scale but also refine the set of observable indicators to a more manageable but statistically
adequate number of items. To improve the general confidence in the new scale, additional data
which was used to validate and cross-validate the measure were collected during the validation
stage, resulting finally in a valid, reliable and practicable measurement instrument for eWOM
trust.

Results Reliability Stage

As mentioned earlier, in the reliability stage, initial data were gathered by setting an online
questionnaire to 850 students of two major Austrian universities (University of Vienna and
Vienna University of Economics and Business). Given the fact that the sample was sufficiently
large, it was possible to split the data into two sub-samples and to randomly assign each
respondent to one of these groups. As is constantly recommended in literature, one sample
served as the primary development sample (sample 3a; n = 425), which enabled this research to
evaluate the items and to arrive at a preliminary solution for the scale that seems optimal under
consideration of the data. The second sub-sample (sample 3b, n = 425), the holdout sample,
was then used to replicate and cross-check these findings. It is assumed that this process offers
valuable information of the scale’s stability, because it enables the assessment of multiple
aspects of its psychometric properties, such as its external validity and internal consistency
across the split samples. For instance, if Cronbach’s alphas remain quite stable in both groups,
it gives more confidence that these values are not distorted by chance (DeVellis, 2012).

This said, the development as well as the hold-out sample were both subject to the same
statistical analyses. Here, a two-step approach was taken. First, exploratory factor analysis
(EFA) was applied in order to investigate the scale’s dimensionality and its reliability. Together
with additional reliability analyses on the item level, this procedure was used to trim the initial
set of 53 items and only retain the scale’s observable indicators with reasonable statistical
properties. Second, confirmatory factor analysis (CFA) was conducted to confirm the
hypothesized structure of the construct and to further assess the findings of EFA. Besides the
assessment of the scale’s dimensionality, CFA also enabled this research to test for aspects of
convergent validity, discriminant validity among the factors, as well as internal consistency of
the preliminary form of the scale (Anderson & Gerbing, 1988; Fornell & Larcker, 1981).

210
This research’s subsequent application of EFA prior to CFA is the result of both established
recommendations for scale development (e.g., Worthington & Whittaker, 2006), as well as the
regular procedures pursued in literature (e.g., Bearden et al., 2001; Shimp & Sharma, 1987).
For example, Netemeyer et al. (2003) state that prior EFA is valuable, especially under
conditions when item sets are large. Here, it enables researches not only to select the most
appropriate scale items but also to determine the suitable number of factors. Subsequent CFA
increases the researcher’s confidence in the previously identified structure when it can be
successfully replicated. Also Hildebrandt & Temme (2006) acknowledge that the two-step
approach is a common and useful practice. Given this, the same procedure was applied in this
research.

Exploratory Factor Analysis

In its early stages, EFA should generally be part of the scale development process as it here
serves two primary purposes. (1) Assessing the construct’s dimensionality. In general, it’s
possible that scale (reflective) items theorized to represent a scale have one, several or – in the
worst case – no common underlying variable. EFA enables the researcher to empirically
identify potential underlying sub-dimensions (i.e., latent variables, constructs or factors) of the
scale and which underlie the specified item set. Hence, it can be tested whether or not the
empirical findings are consistent with the theoretical a-priori structure of the eWOM construct.
The EFA ensures that the observable indicators used to measure the sub-constructs truly capture
their common character – this has to be done for each dimension (Hattie, 1985). (2) Measure
purification. Another important feature of EFA is to reduce the number of items, so that the
remaining items maximize the new scale’s explained variance in the data, which enhances the
scale’s reliability (Netemeyer et al., 2003).

By using data from the developmental sample, the initial set of items was subject to a principal
components analysis (PCA) with a Promax (oblique) rotation using IBM SPSS 21 for
Windows. For the last decade, an active discussion about the preferred use of PCA over
common factor analysis (e.g., principal-axis analysis, maximum-likelihood factoring) exists,
which has not been resolved yet (Worthington & Whittaker, 2006). While some authors suggest
that common factor analyses should be preferred over PCA (e.g., Netemeyer et al. 2003), other
scholars publishing in the field of scale development do not make an explicit recommendation
(e.g., DeVellis 2012) and examples for applying PCA are far from being far from common (e.g.,
Delgado-Ballester, 2004; Lin & Hsieh, 2011). The discussion is also stimulated by the fact that
it has been shown that PCA and common factor analysis often yield the same results (Velicer
& Jackson, 1990; Velicer et al., 1982). This is especially true when certain conditions are met.
For instance, Netemeyer et al. (2003) note that if the number of investigated items exceeds 30
or communalities for the most items exceed .60 (Hair et al., 2010) (which is the case in this

211
research), results from common factor analysis and principal components analysis tend to be
similar. These conditions are also met in the current study.

This research also opted for an oblique rotation method, which was preferred to alternative
orthogonal rotation methods (e.g., Varimax), due to reasons grounded in theory as well as in
the empirical findings. (1) Earlier literature provides strong theoretical as well as empirical
evidence that the diverse aspects of trust (especially when trust is understood as a stable
psychological construct – a personality trait) are typically strongly intertwined and do not exist
in complete independence from each other (e.g., Delgado-Ballester, 2004; Doney & Cannon,
1997). Under conditions where correlated factors can be assumed, scale development literature
generally suggests the application of oblique rotation methods, as they are theorized to reflect
the structure in the data more adequately and thereby increase interpretability (DeVellis, 2012;
Gerbing & Anderson, 1988; Netemeyer et al., 2003). (2) Even when theory suggests that factors
are uncorrelated, various scholars nevertheless suggest the use of an oblique rotation when
factors are correlated in the data (Worthington & Whittaker, 2006). Hildebrandt & Temme
(2006) also argue for an oblique rotation, as they state that this approach leads to meaningful
results even when factors are uncorrelated. An examination of the sample data revealed notable
factor correlations ranging from .46 to .67 (sample 3a).

Prior to further analysis, Kaiser-Meyer-Olkin (KMO) tests for sampling adequacy (> .80) and
Bartlett tests of sphericity (p < .01) were conducted in line with additional analysis (Backhaus
et al., 2006) to ensure the appropriateness of the data for factor analysis. The initial exploratory
factor analysis on the developmental sample resulted in nine-factor solution with eigenvalues
greater than one. The explained total variance explained by these factors was 62.63%. In order
to further refine the scale in a stepwise procedure, all items were removed which do not
substantially load on the hypothesized factor. In this research, factor loadings in the .40 range
and above were classified as substantial. This respects earlier general recommendation by
various scholars (e.g., Floyd & Widaman, 1995) but also the suggestions by Hair et al. (2010),
who advocate that loadings should also be judged in respect to actual sample size. Netemeyer
et al. (2003) further point out that individual item factor loadings should be evaluated across
multiple data sets (i.e., compared to the hold-out sample) and only items should be retained that
consistently show loadings no less than .40. In order to further reduce the number of items,
indicators with considerable and significant cross-loadings (>.32) on other than their intended
factor were trimmed (Netemeyer et al., 2003; Ohanian, 1990). Also items with communalities
with less than .50 were removed from further analysis, as they were regarded as having limited
explanatory power (Hair et al., 2010). The remaining items were again subject to a new factor
analysis, which was able to identify five factors explaining 66.89% of the total variance in the
data.

212
The final number of factors was determined by consideration of multiple criteria. One of the
most widely used “rules of thumb” or non-statistical (i.e., not based on likelihood criteria)
decision rules is the eigenvalue rule (or Kaiser-Guttman criterion or latent root criterion)
(Kaiser, 1960), which asserts that factors with eigenvalues less than one which contains less
information than the average item should not be retained. However, in literature exist various
hints that the eigenvalue rule may not be a reliable basis for retaining factors (Cliff, 1988). The
scree test (Cattell, 1966) represents another rule of thumb which is also based on the factors’
eigenvalues but considers their relative rather than their absolute values. Here, the adequate
number of factors can be determined by looking at the drop in the amount of information
(represented by the factor’s eigenvalue) across the identified factors. All factors that lie above
a sharp “elbow”, where the factors’ eigenvalues drastically approach zero, should be retained.
The rationale is that deleting factors below this elbow will result in minor losses of explained
variance. However, depending on the analytic context, factors with eigenvalues less than one
may be worth considering (DeVellis, 2012). As the judgment of the elbow is subjective, Horn
(1965) advocated the use of parallel analysis in order to identify the elbow based on
psychometric criteria. Some scholars theorize that examining the amount of variance that is
explained by the extracted factors can serve as a valuable guideline. Here, it is generally
recommended that the number of factors extracted should account for at least 50% of the
variance in the items and that at least 5% of the total variance explained should be attributable
to each factor (Hair et al., 2010; Netemeyer et al., 2003). A final rule of thumb pertains to an a-
priori criteria, that is, to restrict the final factor solution to a prespecified number of factors.
This number must be consistent with a sound theoretical basis.

Finally, 25 items were retained from exploratory factor analysis. The application of parallel
analysis as well as the recognition and evaluation of additional criteria led this research to
assume that a five factor solution was most appropriate. By restricting the number of factors,
the final factor solution obtained explained 66.90% of the total variance in the data of the
estimation sample. Each of the final items loaded on its intended factor with loadings of .50 or
greater. Hair et al. (2010) terms such loadings as “very significant”. Nevertheless, a single item,
In9, showed a slightly lower factor loading of .43. As the loading was still above the .40
threshold, it was regarded as substantial. Table 19 presents the results of the factor analyses for
both reliability samples, including the factor loadings taken from the pattern matrix.

213
Table 19: Results of the Initial EFA (Sample 3a and 3b)

Variance
Sub-
Explained/ Item Factor Loading Communality MSA
dimension
Eigenvalue
Ability 8.60% / 2.15 Ab6 .66 (.58) .54 (.51) .96 (.97)
(39.66% / 9.92) Ab7 .94 (.96) .73 (.78) .96 (.95)
Ab8 .70 (.83) .59 (.63) .98 (.95)
Ab9 .90 (.82) .75 (.73) .96 (.96)
Ab10 .77 (.77) .70 (.63) .96 (.96)
Ab11 .67 (.84) .64 (.73) .96 (.96)
Integrity/ 44.44% / 11.11 In2 .82 (.91) .67 (.65) .96 (.93)
Honesty (10.35% / 2.59) In3 .95 (.96) .67 (.65) .94 (.94)
In4 .87 (.78) .69 (.67) .96 (.94)
In5 .73 (.73) .70 (.65) .97 (.96)
In6 .53 (.42) .66 (.60) .97 (.96)
In7 .55 (.66) .50 (.51) .97 (.96)
In9 .43 (.40) .54 (.53) .97 (.97)
In10 .68 (.48) .62 (.59) .97 (.96)
Benevolence 3.56% / .98 Be1 .83 (.75) .68 (.59) .87 (.88)
(5.28% / 1.32) Be2 .69 (.74) .56 (.56) .96 (.92)
Be3 .66 (.85) .68 (.67) .87 (.81)
Be6 .53 (.74) .43 (.54) .92 (.92)
Willingness to 5.80% / 1.45 Wi1 .58 (.65) .74 (.72) .95 (.95)
rely (6.38% / 1.59) Wi4 .90 (.79) .84 (.75) .93 (.94)
Wi5 .77 (.77) .75 (.77) .95 (.93)
Wi8 .79 (.76) .82 (.77) .94 (.94)
Willingness to 4.50% / 1.12 Wi2 .81 (.67) .63 (.55) .93 (.96)
depend (3.88% / .97) Wi6 .85 (.93) .81 (.81) .83 (.89)
Wi7 .82 (.91) .80 (.79) .91 (.90)
Notes: Results for holdout sample 3b in parentheses; Total variance explained: 66.89% (65.55%); Extraction
method: Principal Component Analysis; Rotation method: Promax with Kaiser Normalization; Rotation
converged in 7 (6) iterations; Kaiser-Meyer-Olkin Measure of Sampling Adequacy (MSA) .95 (.94) and
Bartlett’s test of Sphericity: sig. .001 (.001).

After examining the content of the final items making up the five latent sub-dimensions of
eWOM trust, the following labels were suggested: (1) (eWOM) ability (usefulness), (2)
(eWOM) integrity/honesty, (3) (eWOM) benevolence, (4) willingness to rely (use), and (5)
willingness to depend (risk). The first factor, ability, reflects the respondent’s belief that online
customer reviews, in general, are a useful, insightful, and knowledgeable source to obtain
constructive purchase-relevant information and/or valuable recommendations. Items that load
on this factor were: relevant, helpful, instructive, useful, applicable, and informative. They all
mirror functional attributes of eWOM or attributes of personal relevance. In contrast, the second
factor, integrity/honesty, represents aspects of eWOM trust that consumers associate with
reviews’ reliability. More specifically, this factor combines all consumers’ beliefs that the
information conveyed in online customer reviews is honest, truthful, and verified. Items being
associated with this element of trust included: honest, credible, genuine, true, authentic, serious,
dependable, and correct. Hence, generalized eWOM trust is reflected (as expected) by a
person’s constant belief that this information is believable. The third factor represents eWOM
214
benevolence; that is, the consumer’s beliefs that the reviews mirror their creators’ favourable
social motive, which is shaped by the reviewer’s altruism and positive reciprocation. This
dimension measures the manipulative character of the communication. The items attributed to
this factor are: benevolent towards me, kind-hearted, social, and caring. All these dimensions
accurately capture the factors discussed in the conceptual background section.

However, some modifications had to be made to the original construct concerning the
motivational aspects of trust: More specifically, the conative element was split into two separate
factors, namely the willingness to rely and the willingness to depend. Both factors mirror
respondents’ behavioral intention to use customer reviews for making purchasing decisions.
However, the kind and level of risk associated with these behavioral intentions differs
essentially. It can be concluded that items loading on the former factor capture the consumer’s
willingness to use eWOM under medium purchasing/functional risk conditions (i.e., eWOM as
a research tool), whereas items loading on the latter are more likely to measure behavioral
intentions under heightened personal risk (i.e., eWOM as a shopping initiator). The latter
includes serious economic as well as social risks. The absence of the Likeability factor also
demonstrated that this element cannot be regarded as an impartial element of eWOM trust. All
its theorized observable indicators were highly intertwined with the other sub-dimensions of
trust and hence removed from the final set of items. It is theorized that the kind of social
relationship caused the minor importance of a purely emotional factor.

The above-discussed factor structure was also reproduced by using data from the hold-out
sample (3b) and item characteristics were generally alike to the patterns shown in the
developmental sample. The considerable sample size (n=425 in both samples) increases this
research’s confidence in these findings, as in literature some agreement exists that “large sample
sizes” are likely to result in more stable correlations among the variables and also positively
influence the replicability of EFA results. Worthington and Whittaker (2006), for instance,
argue that a sample size of at least 300 respondents can be regarded as adequate in most cases.
With a final ratio of participants to items (17:1 in both samples), the research at hand also meets
regularly proposed standards (Grosuch, 1983).

This said, items that marginally did not pass required scale thresholds in one sample (e.g., Be6
was not able to reach an acceptable communality level in the developmental sample but in the
hold-out sample) were retained for further analysis. However, they were regarded as candidates
for exclusion.

215
Table 20: Reliability and Item-based Statistics

Average Cronbach’s Corrected

Sub-
Inter-Item Alpha Item Item-to-Total Mean SD
dimension
Correlation Correlation
Ability .59 (.59) .89 (.90) Ab6 .63 (.62) 3.95 (4.02) 1.60 (1.13)
Ab7 .74 (.81) 4.30 (4.42) 1.20 (1.19)
Ab8 .66 (.68) 3.99 (4.00) 1.14 (1.27)
Ab9 .79 (.78) 4.41 (4.41) 1.06 (1.20)
Ab10 .75 (.69) 4.12 (4.18) 1.06 (1.08)
Ab11 .71 (.77) 4.21 (4.29) 1.11 (1.10)
Integrity/ .54 (.50) .90 (.89) In2 .73 (.67) 3.42 (3.29) 1.25 (1.20)
Honesty In3 .67 (.62) 3.67 (3.62) 1.26 (1.18)
In4 .74 (.74) 3.29 (3.24) 1.17 (1.08)
In5 .77 (.72) 3.56 (3.50) 1.17 (1.15)
In6 .70 (.66) 3.23 (3.17) 1.14 (1.19)
In7 .61 (.59) 3.42 (3.40) 1.25 (1.21)
In9 .63 (.61) 2.97 (3.05) 1.15 (1.14)
In10 .70 (.65) 3.08 (3.06) 1.07 (1.08)
Benevolence .42 (.43) .75 (.75) Be1 .60 (.55) 2.34 (2.47) 1.43 (1.30)
Be2 .53 (.51) 3.19 (3.37) 1.36 (1.32)
Be3 .68 (.60) 2.46 (2.53) 1.35 (1.36)
Be6 .45 (.52) 2.84 (2.98) 1.35 (1.30)
Willingness .70 (.70) .90 (.90) Wi1 .75 (.77) 4.59 (4.65) 1.42 (1.29)
to rely Wi4 .81 (.77) 4.63 (4.61) 1.25 (1.30)
Wi5 .75 (.78) 4.52 (4.59) 1.33 (1.29)
Wi8 .81 (.80) 4.54 (4.63) 1.32 (1.23)
Willingness .64 (.61) .84 (.82) Wi2 .58 (.56) 3.42 (3.50) 1.61 (1.57)
to depend Wi6 .81 (.77) 3.35 (3.33) 1.60 (1.63)
Wi7 .74 (.72) 3.27 (3.22) 1.55 (1.54)
Note: Results for sample 3b in parentheses.

As a next step, item- and reliability-based statistics were inspected (see Table 20). This includes
the review of internal consistency estimates, such as Cronbach’s alpha, average inter-item
correlations, corrected item-to-total correlations, as well as item means and variances.
Reliability is a critical issue in reflective measurement models. Here, a high consistency or
correlation between the items is desired, as each item should act as a substitute of the construct
itself. Low correlations indicate the presence of measurement error and insufficient reliability
of the items (Hildebrandt & Temme, 2006). One of the most prominent indicators of a scale’s
quality is Cronbach’s coefficient alpha (Cronbach, 1951), which is regularly reported when
developing a new scale. It serves as an indicator of the proportion of variance in the scale scores
that is attributable to the true score (DeVellis, 2012). All factors were able to achieve what
DeVellis (2012) calls a “respectable” consistency level of .70 or greater. In addition, all
identified sub-dimensions of eWOM trust – except Benevolence – showed what literature
typically considers very good alpha levels, above .80 in both samples (Bearden & Netemeyer,
1998; Clark & Watson, 1995). Specifically the Cronbach alphas for ability was .89 (.90)
(numbers in parentheses indicate results of the holdout sample 3b), integrity/honesty achieved

216
.90 (.89), and the two behavioral intentions dimensions .90 (.90) and .84 (.82) respectively.
However, given that Cronbach’s alpha is also a function of scale length (Netemeyer et al.,
2003), and various authors suggest .70 as an acceptable lower bound of alpha (DeVellis, 2012;
Hair et al., 2010), the slightly lower internal consistency of the benevolence factor is deemed
acceptable. None of the alphas was better than .90, so shortening the scale at this stage was not
an issue.

While virtually all item problems led to a reduction of Cronbach’s alpha, various authors
suggest that the average inter-item correlation is a more meaningful index for internal
consistency (Clark & Watson, 1995). Here, Robinson et al. (1991) propose a threshold of .30
or better. Others advocate correlations as low as .15 to .20 as being desirable (e.g., Bearden et
al., 2001). As the sub-dimensions in this research are closely defined and can be regarded as
exhibiting a narrow focus, a higher average inter-item correlation in the range of .40 to .50 may
be more adequate (Netemeyer et al., 2003). All inter-item correlations were significant and
substantial – ranging from .42 to 70 (holdout sample: .43 to 70). This provides exemplary
evidence for internal consistency.

In order to improve confidence in the internal consistency among a set of highly inter-correlated
scale items, it is desirable that each item correlate considerably with the collection of the
remaining items. The corrected item-to-total correlation provides an insightful index for
evaluating this pattern of the individual items. As a decision rule, items that exhibited item-to-
total correlations below .35 were excluded (Bearden et al, 2001) and items below .50 were
regarded as candidates for later exclusion (Obermiller & Spangenberg, 1998; Tian et al., 2001).
Here, Be6 appeared to have a corrected correlation of .45 in the developmental sample, but
succeeded in achieving an adequate level in the hold-out sample. The remaining items showed
properties well above the set threshold level and were mostly in the .70 and .80 range. Finally,
by following established scale development guidelines advanced by DeVellis (2012) and
Netemeyer et al. (2003), also individual item means, variances and item ranges were inspected:
All items exhibited acceptable patterns (i.e., means were close to the center of the theoretical
range of scores (0-6), the item variances were relatively high, and the observed item ranges
corresponded to the theoretical ranges).

In sum, the initial solution for the new eWOM trust scale was quite promising and only a few
items could be considered candidates for further exclusion. But, as these statistically “weak”
items have both face- as well as content validity, they were preliminarily retained for the next
stage of analysis, which provided further evidence for the scale quality of the individual items.
The importance of ensuring content validity outranking statistical considerations about
reliability issues is emphasized by various authors, including Hildebrandt and Temme (2006).

217
Confirmatory Factor Analysis

Subsequently, a confirmatory factor analysis (CFA) was carried out. This is justified by the
increasing application of structural equation modeling (SEM) techniques, which have become
standard in the assessment of multi-item measures over the years (e.g., Anderson & Gerbing,
1982; Martens, 2005; Martens & Hasse, 2006; Steenkamp & Trijp, 1997). CFA differs from
EFA in that the former can be used to confirm a-priori hypotheses about the relationships on
the diverse levels of the construct (i.e., the measurement model). EFA lacks this ability. Hence,
CFA is a valuable instrument for testing the construct’s dimensionality or structure and can
further be used to assess the reliability as well as validity of the construct (second-order), factor
(first-order), and individual item level. The equally sized (n=425) developmental and the
holdout sample were both processed independently by means of LISREL 8.50 for Windows
(Jöreskog & Sörbom, 1993). LISREL is a software package that is associated with the
covariance structure analysis (CSA). A CSA approach was preferred to partial least squares
(PLS) analysis, as it enables the estimation of measurement models, including error terms, the
empirically comparison of model specifications and the assessment of the overall fit of
(alternative) model structures. In addition, LISREL offers a number of diagnostics which can
be helpful for model respecification (Diamantopoulos & Winklhofer, 2001). All confirmatory
factor models discussed in this thesis have used a covariance matrix as input and the maximum
likelihood estimation method.

Measurement Model and Goodness-of-fit

Figure 5 shows the measurement model and the specified relationships between the construct
and the measures. This model is grounded in the theoretical discussion of the eWOM construct,
but is modified to fit the construct’s factor structure identified via the EFA analysis. The model
represents a higher-order construct with first-order reflective dimensions which are themselves
measured by several reflective manifest items. Specifically, a measurement model with a
second-order construct and five first-order factors (latent variables) was specified and each
observable indicator was prescribed to be loaded on one of these latent variables according to
the factor structure indicated in the exploratory factor analysis. Corresponding to this
conceptualization, error terms can be identified on two different levels: (1) at the level of the
manifest indicators (ε) representing the measurement error, and (2) at the level of the first-order
sub-dimensions (ζ) where it captures the amount of variance in the first-order construct which
the second-order construct does not account for. This type I model (Jarvis et al., 2003) was
appropriately specified for both samples in LISREL, where η1, η2, and η3 represented the belief-
dimensions and η4, and η5 the behavioral aspects of eWOM trust. In this model, 55 parameters
were estimated by the maximum-likelihood (ML) method, resulting in a ratio of participants to
number of parameters of approximately 7:1. This met the requirements of Bentler and Chou
(1987) concerning sample size.

218
 Figure 5: Adopted Measurement Model

Y1 Y5
Y2 Y3 Y4

η1 η2 η3 η4 η5

ζ1 ζ2 ζ3 ζ4 ζ5
λ11 λ21 λ .. λ72 λ82 λ .. λ153 λ163 λ .. λ194 λ204 λ .. λ235 λ245 λ ..
y1 y2 ... y7 y8 ... y15 y19 ... y19 y20 ... y23 y24 y25

ε1 ε2 ε .. ε8 ε8 ε .. ε15 ε16 ε .. ε19 ε20 ε .. ε23 ε24 ε..

Notes: ξ = Second-order construct (eWOM Trust); η = First-order construct (η1 = Ability; η2 =

Integrity/Honesty; η3 = Benevolence; η4 = Willingness to rely; η5 = Willingness to depend); γ = Observable
indicator; ε = error term of the observable indicator.

The estimation of the model provided first evidence for its validity, as the solution was deemed
proper (i.e., the estimation procedure converged and none of the variance estimates was
negative), the relationships theorized were statistically significant (p < .01) and the group of
relationships were consistent within both samples. No offending estimates, such as negative
error variances of items (Heywood cases) and/or completely standardized parameter estimates
greater than one, occurred.

Subsequently, the chi-square statistic (χ2) was assessed to test whether the model adequately
accounts for the samples’ data. For both samples, the χ2 test was significant (χ2dev = 668.12,
dfdev = 270, pdev < .001; χ2hold = 589.12, dfhold = 270, phold < .001). However, various scholars
argue that this statistic is susceptible to numerous sampling fluctuations (e.g., sample size;
model complexity, non-normality), leading to limited insights concerning the model’s quality
(e.g., Hu & Bentler, 1999). Hence, it is reasonable to evaluate the model fit by a number of
alternative indices, which may be classified as (1) stand-alone or absolute, (2) comparative or
incremental, and (3) parsimony-based fit indices. Hu and Bentler (1999) suggest reporting at
least one fit index of each category.

Accordingly, the following stand-alone fit measures were obtained: the goodness-of-fit index
(GFI) (Jöreskog & Sörbom, 1984) and the adjusted goodness-of-fit index (AGFI) (Jöreskog &
Sörbom, 1984), which both can range from 0 to 1, whereas higher values suggest a better fit.
Byrne (1994) proposes that GFI should exceed .90 and Baumgartner and Homburg (1996)
219
advocate the same cut-off value for AGFI. As some researchers argue that both measures are
regularly described as being not “optimal” indicators of model fit due to inconsistencies from
sampling characteristics (Hoyle, 1995), while others vehemently argue against their usage (Hu
& Bentler, 1999), additional indices were also reviewed, among these, the root-mean-square-
error-of-approximation (RMSEA) (Steiger & Lind, 1980). While various cut-off levels exist,
in literature RMSEA estimates of .08 (and less) are generally regarded as being acceptable,
while values at or less than .05 indicate a close model fit. In contrast, values over .10 have been
advocated to be indicative of a bad fit (Browne & Cudeck, 1993; Worthington & Whittaker,
2006). Also root-mean-square-residual (RMR) (Jöreskog & Sörbom, 1981) with values smaller
than .08 being indicative for a good model fit (Browne & Cudeck, 1993) and standardized root-
mean-square residual (SRMR) (Bentler, 1995) – with the same threshold (Hu & Bentler, 1999)
– represent absolute fit indices regularly reported. These fit indices, in general, measure how
well an SEM is able to explain the relationships found in the data (Worthington & Whittaker,
2006).

In contrast, measures belonging to the class of incremental fit indices typically compare the
estimated model to a baseline or “null” (or independence) model, which assumes no
relationships among the data and measures the improvement in a model’s fit (Worthington &
Whittaker, 2006). Most commonly used incremental fit indices are the comparative fit index
(CFI) proposed by Bentler (1990), for which a value of .90 or larger is regarded to indicate
acceptable model fit (Hu & Bentler, 1999), the non-normed fit index or Tucker-Lewis index
(NNFI/TLI) (Tucker & Lewis, 1973) with values over .90 are considered acceptable (Hu &
Bentler, 1999), and the normed fit index (NFI) (Bentler & Bonett, 1980). Again, values should
exceed .90 for NFI (Byrne, 1994; Hair et al., 2010).

Parsimony-based indices typically adjust fit in order to compare models with differing numbers
of estimated parameters (i.e., fit per coefficient) (Netemeyer et al., 2003). Specifically, the
normed chi-square (chi-square/df ratio) for which acceptance levels range from less than 2
(Ullman, 2001) to less than 5 (Schumacker & Lomax, 2004), Akaike’s information criterion
(AIC) (Akaike, 1987) and the parsimonious goodness fit index (PFI) were assessed to ensure
the acceptability of the proposed model from different perspectives. In sum, the initial
estimation achieved a reasonable fit in both samples, such that the conceptual model was
deemed to be an acceptable representation of the hypothesized eWOM trust construct with
RMSEA = .058 (.053), GFI = .90 (.90), AGFI = .87 (.88), NFI = .90 (.90), NNFI = .93 (.94),
CFI = .94 (.94), normed chi-square = 2.47 (2.18), and PFI = .74 (.75).

220
Reliability

Table 21 summarizes the psychometric properties of the scale for the development as well as
holdout sample. In this research, reliability of the items was evaluated by assessing the path
coefficients and examining the squared multiple correlation (R²) (Bollen, 1989). All the items
turned out to load highly on their respective sub-dimension and t-values associated with the
loadings exceeded critical values (2.75) for significance on the .01 level. This verifies the
hypothesized relationships among the observable indicators and constructs. Similar results were
achieved in the holdout sample. In respect to the squared multiple correlation, values better than
.50 (Fornell & Larcker, 1981) are typically preferred. This implies that the majority of the
variance in the indicator is attributable to the latent (sub-)construct. The R² for most items were
well above this threshold. However, in the developmental as well as the holdout sample, seven
items were not able to achieve the required reliability level and had slightly lower values.

Cronbach’s alpha, Fornell and Larcker’s (1981) index of construct (or composite) reliability,
the average inter-item-correlation (discussed earlier), as well as the average variance extracted
(AVE), which provides an index for the amount of variance that is captured by the latent
construct in relation to the amount of variance due to measurement error, were used to assess
the reliability of each latent sub-dimension. In both samples, Cronbach’s alphas range from .75
to .90, while construct reliabilities varied from .75 to .90/.91, which implies that the minimum
threshold of .70 was exceeded in all cases. In respect to the AVE, four of the five sub-constructs
also met the recommended level of 50% (Hair et al., 2010). Only the benevolence factor
performance was inferior to the other components of eWOM trust (mean AVE = .43).

A further investigation demonstrated that all five first-order constructs had positive and
significant loadings on the second-order construct. The squared multiple correlations also
achieved desirable levels above the recommended threshold in all but one relationship
(benevolence) (R² = .33/.26). On the second-order level, the reliability of the first-order
construct was also assessed by the index of construct (or composite) reliability and the AVE
for both samples. The internal consistency measures exceeded the accepted standard of .70
(Nunnally & Berstein, 1994) (.89/.85) and .50 (Fornell & Larcker, 1981) respectively (.63/.54).
In general, the model provides reliable measurement of the eWOM trust construct.

221
 Squared Corrected Average

222
Sub- Cronbach’s Construct Construct
γ Item λ Multiple Item-to-Total Inter-Item AVE AVE
dimension Alpha Reliability Reliability
Correlation Correlation Correlation
Ab6 .68 (.67) .46 (.45) .63 (.62)
Ab7 .78 (.85) .60 (.72) .74 (.81)
Ab8 .71 (.72) .50 (.51) .66 (.68) .59
Ability .92 (.83) .59 (.59) .89 (.90) .89 (.90)
Ab9 .83 (.83) .68 (.69) .79 (.78) (.60)
Ab10 .82 (.74) .60 (.55) .75 (.69)
Ab11 .78 (.82) .61 (.67) .71 (.77)
In2 .76 (.71) .58 (.50) .73 (.67)
In3 .70 (.65) .49 (.43) .67 (.62)
In4 .77 (.79) .60 (.62) .74 (.74)
Integrity/ In5 .82 (.77) .67 (.59) .77 (.72) .56
.85 (.76) .54 (.50) .90 (.89) .91 (.89)
Honesty In6 .76 (.72) .59 (.51) .70 (.66) (.50)
In7 .65 (.63) .42 (.40) .61 (.59)
In9 .69 (.67) .47 (.44) .63 (.61) .63
.89 (.85)
In10 .75 (.70) .56 (.50) .70 (.65) (.54)
Be1 .72 (.65) .52 (.43) .60 (.55)
Be2 .67 (.63) .45 (.40) .53 (.51) .43
Benevolence .57 (.51) .42 (.43) .75 (.75) .75 (.75)
Be3 .67 (.71) .49 (.51) .68 (.60) (.43)
Be6 .55 (.64) .30 (.41) .45 (.52)
Wi1 .82 (.83) .67 (.68) .75 (.77)
Willingness Wi4 .86 (.82) .74 (.66) .81 (.77) .70
.79 (.79) .70 (.70) .90 (.90) .90 (.90)
to rely Wi5 .79 (.84) .63 (.71) .75 (.78) (.70)
Wi8 .87 (.86) .76 (.74) .81 (.80)
Table 21: Psychometric Properties of the eWOM Trust Scale (Samples 3a and 3b)

Wi2 .63 (.63) .40 (.39) .58 (.56)

Willingness .67
.78 (.76) Wi6 .92 (.89) .86 (.80) .81 (.77) .64 (.61) .84 (.82) .86 (.84)
to depend (.64)
Wi7 .88 (.86) .77 (.74) .74 (.72)
Notes: Results for the holdout sample 3b in parentheses; γ = Completely standardized second-order loading; λ = Completely standardized first-order loading;
AVE = Average variance extracted.
Convergent Validity

According to Bagozzi and Yi (1991) and others (Anderson & Gerbing, 1988; Fornell & Larcker,
1981; MacKenzie et al., 2011), weak evidence for convergent validity is given by significant
factor loading on an item under investigation. A stricter assessment of convergent validity is
provided by the evaluation of the squared factor loading. That is, if the squared factor loading
is greater than .50 – meaning that the majority of the total variation in the measure is attributable
to the latent construct – convergent validity can be assumed. In both samples, all factor loadings
(λ) were greater than .60 and showed t-values greater than 2.57 (ranging from 9.15 (10.39) to
20.91 (20.82)), implying a significant loading on their intended sub-dimensions on the .01
significance level. Additionally, 20 (18) of the 25 items had squared factor loadings (λ2) above
the required threshold of .50, whereas two of the items (in both samples) almost reached this
limit. Similarly, the validity of each first-order sub-dimension as indicators of the second-order
construct was tested. All five factor loadings were considerably high and significant (ranges of
t-values: 8.66 (7.96) to 17.82 (15.61)). Four out of five first-order constructs turned out to have
a squared factor loading (γ) greater than .50. However, the loading on the benevolence
dimension was problematic (average γ = .29). Convergent validity can also be evaluated by
assessing the extent to which the first-order constructs are correlated (Bagozzi and Yi 1991).
All correlations in both samples turned out to be considerable (ranging from .45 (.39) to .78
(.65)) and significantly related (p < .05) (see Table 21). This implies that all factors measure a
specific aspect of the same construct; supporting convergent validity.

Discriminant Validity

Another critical issue was to investigate the dimensionality of the identified sub-constructs; that
is, to evaluate whether the single sub-dimensions are distinguishable from one another.
Campbell and Fiske (1959, p. 84) emphasize this issue by noting that “one cannot define
without implying distinctions, and the verification of these distinctions is an important part of
the validation process”. This is called discriminant validity. In the research at hand, the
discriminant validity of the five sub-dimensions of eWOM trust was assessed in three different
ways: first, by testing the confidence interval around the correlation between factors; second,
more rigorous evidence for discriminant validity is obtained by comparing the shared variance
(i.e., squared correlation) between each pair of factors against the AVEs of the two factors
(Fornell & Larcker, 1981); third, by comparing the results of chi square difference tests and fit
statistics of competing measurement models.

In order to argue for discriminant validity, one has to examine the relationship of factors to
ensure that they are not perfectly correlated; that is, the correlations equal to one (Bagozzi &
Yi, 1991; Smith et al., 1996). Hence, evidence for discriminant validity in the first case is given
if the correlations are two or more standard errors below 1.0 or, in other words, the confidence
interval (+/-2 standard errors) around the correlation of two sub-dimensions does not contain
223
the value of 1.0 (Anderson & Gerbing, 1988; Schmitt & Stults, 1996). In the second case,
discriminant validity is given if the shared variance estimate is smaller than the AVE estimates
of the involved constructs (Farrell, 2010; Hair et al., 2010). In the third case, discriminant
validity is assumed if a more constrained model, that is, a model that assumes the existence of
multiple factors, exhibits a significantly better model fit compared to a less constrained model
(Anderson & Gerbing, 1988; Jöreskog, 1971) – that is one that assumes that all indicators load
on a single factor. Literature regularly provides examples for the application of these rules of
thumb (e.g., Netemeyer et al., 1996).

Table 22 reports the AVE of the individual sub-dimensions, the (completely standardized)
correlations and squared correlations of the factor pairings. The table also includes the upper
bound of the 95% confidence interval of the correlations between every possible factor pairing
for the two samples. For instance, the upper bound of the confidence interval was .82 (.68) for
the correlation between the ability and the integrity/honesty sub-dimension, .56 (.46) for the
benevolence and integrity/honesty correlation, and .52 (.47) for the willingness to depend and
benevolence correlation. The highest upper bound of the confidence interval (.82) could be
found between ability and integrity/honesty in the development sample. However, a comparable
result was not achieved in sample 3b. Here, in contrast, the highest upper bound was identified
for the correlation between willingness to rely and ability (.70). As none of the confidence
intervals included 1.0, support for discriminant validity is given.

A slightly different picture is provided by assessing the discriminative power of the factors by
using the second criterion. Here, the squared correlation between ability and integrity/honesty
(.61) was slightly above the AVEs of the integrity/honesty (.56), as well as the ability (.59)
construct. Therefore, discriminant validity might not be given. However, as this pattern cannot
be detected in the second sample (here the squared correlation among the factors was much
lower: .40 compared to the two involved AVE (.50 for integrity/honesty and .60 for ability)),
this threat might be eased. All remaining factors showed higher AVE than the squared
correlation of all possible factor pairings, supporting discriminant validity among the five sub-
dimensions of eWOM trust.

224
Table 22: Correlations among the Sub-dimensions (Samples 3a and 3b)

Integrity/ Willingness Willingness to

Ability Benevolence
Honesty to rely depend
3a 3b 3a 3b 3a 3b 3a 3b 3a 3b
Integrity/ a a
.56 .50
Honesty
b
.78 .63b
Ability .82c .68c .59a .60a
.61d .40d
.49b .39b .53 .42b
Benevolence .56c .46c .59c .49c .43a .43a
.24d .15d .28d .18d
.68b .60b .73b .65b .46b .40b
Willingness to
.73c .66c .77c .70c .53c .48c .70a .70a
rely
.46d .36d .53d .42d .21d .16d
.66b .58b .71b .63b .45b .39b .62b .60b
Willingness to
.71c .64c .76c .69c .52c .47c .67c .66c .67a .64a
depend
.44d .34d .51d .40d .12d .15d .38d .36d
Notes: a = Average variance extracted (AVE); b = Pearson correlation coefficient (r); c = Upper bound of 95%
confidence interval of r; d = Squared correlation (r2); n = 425 (425).

To provide greater confidence in the proposed factor structure and the discriminant validity
among the first-order variables of eWOM trust, alternative measurement models were
estimated. These were compared using chi-square difference tests, as suggested by Anderson
and Gerbing (1988) and a selection of fit indices. This approach follows common practices in
scale research (e.g., Delgado-Ballester, 2004). The number of models as well as the factors that
were combined were determined by the degree of (average) correlation among the different
factors observable in the two samples. Accordingly, one model was proposed that understands
eWOM trust as a second-order construct with three first-order factors (Model 3) and two others
with four first-order factors (Models 4a-4b). For Model 3, the dimensions of ability and
integrity/honesty as well as willingness to rely and willingness to depend were combined in two
constructs because of their considerable interdependence; that is, a correlation of .78 (.63)
between former and of .62 (.60) between the latter constructs. Benevolence was here regarded
as a separate factor. Ability and willingness to rely (r = .73/.65; average correlation (ravg) = .69)
were joined for a four-factor model (Model 4a) and an alternative four-factor model (Model 4b)
originated by the combination of ability and willingness to depend (r = .71/.63; ravg = .67).
Additionally, the hypothesized second-order factor model (Model 5), where responses to each
item are reflective of the five proposed factors, and a one-factor model (Model 1),
conzeptualizing eWOM trust as a uni-dimensional construct where the covariance among the
items being accounted for by a single factor, were specified. Finally, a null model (Model N),
for which the assumption that no systematic relationship among the scale items existed, was
also included.

225
Table 23 and 24 present the results for the competing models. It can be shown that the second-
order model, with its five first-order dimensions, fits the data best. This model had the lowest
chi-square values across the two samples (χ2 = 668.12 (589.12), df = 270)) and also the
additional model fit indices indicate that it’s the most suitable model (RMSEA = .06 (.05), NFI
= .90 (.90), NNFI = .93 (.94), CFI = .94 (.94), RMR = .10 (.11), SRMR = .06 (.07), GFI = .90
(.90), AGFI = .87 (.88)). Additionally, the improvement in chi-square was significant when
comparing this model to all of its less complex counterparts. For instance, the chi-square
difference between the one-factor model and the second-order model with its five first-order
constructs was 1,235.41 (1,536.45) and highly significant (p < .001), and there was also a
significant improvement from the simple three-factor to the five sub-dimensions model (χ2Diff
= 586.09 (784.15), df Diff = 3, p < .001). Chi-square as well as the diverse model fit indices could
also be enhanced when comparing the two four-factor models to the model mirroring the
hypothesized structure of the construct. According to these results, it was concluded that the
five-sub-dimension model is qualified to represent the data inherent in the two samples.
Accordingly, merging any of the proposed dimensions would lead to an impairment of the
measurement model. By presuming that these factors should be treated separately out of
empirical considerations, further evidence for discriminant validity of eWOM trust’s sub-
dimensions was provided.

For establishing discriminant validity, the second-order construct with its five-dimensional
first-order constructs (Model 5) was additionally compared to a five-correlated-factors model
(Model 5b). According to this, conceptualization covariance among the observable indictors is
accounted for by a set of first-order constructs, each construct representing a distinct aspect of
eWOM trust and each item being the effect of only a single factor. All of the Model-5-fit indices
were close to those of this first-order model across the two separate samples (RMSEA = .05
(.05), NFI = .91 (.92), NNFI = .94 (.96), CFI = .95 (.96), RMR = .08 (.08), SRMR = .05 (.05),
GFI = .90 (.91), AGFI = .88 (.89)). These results suggested that a second-order factor structure
of type 1 was acceptable, as it validly captured the construct's true structure embedded within
the sample data.

226
 Chi Square Difference Model Fit Indices

Model Chi Square df p Chi

Competing Chi Square df
Sign. Square / RMSEA NFI NNFI CFI RMR SRMR GFI AGFI
Models Difference Difference
df
Null Model (N) 6,618.07 300 n.a. n.a. n.a. n.a. 22.06 n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a.
One Factor Model (1) 1,903.54 275 .001 6.92 .15 .71 .72 .74 .14 .08 .66 .60

(N-1) 4,714.53 25 ***

Three Factor Model (3) 1,254.22 273 .001 4.59 .11 .81 .83 .85 .12 .07 .75 .71
(A/I; B; WR/WD)
(N-3) 5,363.86 27 ***
(1-3) 649.32 2 ***
Table 23: Competing Models (Sample 3a)

Four Factor Model (4a) 977.05 271 .001 3.61 .09 .85 .88 .89 .11 .06 .82 .78
(A/WR; I; B; WD)
(N-4a) 5,641.02 29 ***
(1-4a) 926.49 4 ***
(3-4a) 277.17 2 ***
Four Factor Model 986.37 271 .001 3.64 .08 .85 .88 .89 .12 .07 .83 .80
(4b) (A/WD; I; B;
WR) (N-4b) 5631.70 29 ***
(1-4b) 917.16 4 ***
(3-4b) 267.84 2 ***
Five Factor Model (5) 668.12 270 .001 2.47 .06 .90 .93 .94 .10 .06 .90 .87

(N-5) 5,949.95 30 ***

(1-5) 1,235.41 10 ***
(3-5) 586.09 3 ***
(4a-5) 308.92 1 ***
(4b-5) 318.25 1 ***
Notes: A = Ability; I = Integrity/Honesty; B = Benevolence; WR = Willingness to rely; WD = Willingness to depend; *** = p <.001.

227
228
Chi Square Difference Model Fit Indices

Model Chi Square df p Chi

Competing Chi Square df
Sign. Square / RMSEA NFI NNFI CFI RMR SRMR GFI AGFI
Models Difference Difference
df
Null Model (N) 5,998.56 300 n.a. n.a. n.a. n.a. 20.00 n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a.

One Factor Model (1) 2,126.16 275 .001 7.73 .17 .65 .65 .68 .16 .16 .61 .53

(N-1) 3,872.40 25 ***

Three Factor Model 1,373.86 273 .001 5.03 .13 .77 .79 .81 .13 .08 .70 .64
(3)
(A/I; B; WR/WD) (N-3) 4,624.70 27 ***
(1-3) 752.30 2 ***
Table 24: Competing Models (Sample 3b)

Four Factor Model 990.81 271 .001 3.66 .10 .84 .86 .87 .11 .07 .80 .76
(4a) (A/WR; I; B; WD)
(N-4a) 5,007.75 29 ***
(1-4a) 1,135.35 4 ***
(3-4a) 383.05 2 ***
Four Factor Model 952.93 271 .001 3.52 .08 .84 .87 .14 .14 .08 .83 .80
(4b) (A/WD; I; B;
WR) (N-4b) 5,045.63 29 ***
(1-4b) 1,173.23 4 ***
(3-4b) 420.93 2 ***
Five Factor Model (5) 589.71 270 .001 2.18 .05 .90 .94 .94 .11 .07 .90 .88

(N-5) 5,408.84 30 ***

(1-5) 1,536.45 10 ***
(3-5) 784.15 3 ***
(4a-5) 401.10 1 ***
(4b-5) 363.22 1 ***

Notes: A = Ability; I = Integrity/Honesty; B = Benevolence; WR = Willingness to rely; WD = Willingness to depend; *** = p <.001.
In order to confidently argue for the construct’s dimensionality, all factors were subjected to
pairwise testing as proposed by Anderson and Gerbing (1988). Here, models for every possible
pairing of the diverse factors were evaluated for their representativeness for the underlying data.
For each pairing, three different models were assessed: (a) a null model which supposed that
among the manifest indicators and the factors, no systematic relationship does exist, (b) a one-
factor model in which all observable items were specified to a single factor; and (c) a two-factor
model (i.e. two correlated first-order factors) representing the hypothesized separate factor
structure, in which the manifest indicators are permitted to load only on their hypothesized
factor. Evidence for discriminant validity among the factors exists if the chi-square fit of the
two-factor model is better than the fit of the one-factor model.

Table 25 and 26 present the fit statistics for the competing models. Throughout the various
factor pairings and samples, fit statistics suggest adequate fit for the two-factor model. To
illustrate, the favourable chi-square measure and diverse fit indices can both be interpreted as
evidence for the discriminability between the dimensions of integrity/honesty and ability. CFA
resulted in a chi-square fit for the one-factor solution of 451.72 (402.65) (df = 77), while the
chi-square for the two-factor model was 167.35 (162.47) (df = 76). The chi-square difference
(χ2Diff = 284.37 (342.71); df Diff = 1) was significant (p < .001), indicating a better model fit of
the two-factor model. Simultaneously, other fit indices of the two-factor model were
considerably enhanced (RMSEA = .06 (.05), NFI = .95 (.95), NNFI = .97 (.97), CFI = .97 (.97),
RMR = .05 (.06), SRMR = .04 (.04), GFI = .94 (.95), AGFI = .92 (.92)), supporting the validity
of the hypothesized model. As in the case of integrity/honesty and ability, all other pairings
showed a significant improvement in chi-square and better fit indices of the two-factor over the
single-factor conceptualizations. Hence, strong empirical arguments for the dimensionality of
the eWOM trust construct and the discriminability of its components lead to the acceptance of
the hypothesized construct structure.

229
 Chi Square Difference Model Fit Indices

230
Sub-
Chi Chi Chi
dimension Model df p Competing df
Square Square Sign. Square / RMSEA NFI NNFI CFI RMR SRMR GFI AGFI
Pairing Models Difference
Difference df
Integrity / Null Model (N) 3,452.40 91 37.94 n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a.
Ability
One Factor Model (1) 451.72 77 .001 (N-1) 3,000.68 14 *** 5.87 .13 .87 .87 .89 .08 .06 .82 .75
Two Factor Model (2) 167.35 76 .001 (N-2) 3,285.05 15 *** 2.20 .06 .95 .97 .97 .05 .04 .94 .92
(1-2) 284.37 1 ***
Integrity / Null Model (N) 2,317.07 66 35.11 n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a.
Benevolence
One Factor Model (1) 320.92 54 .001 (N-1) 1,996.15 12 *** 5.94 .12 .86 .86 .88 .14 .07 .87 .82
Two Factor Model (2) 149.31 53 .001 (N-2) 2,167.76 13 *** 2.82 .07 .94 .95 .96 .08 .05 .94 .91
(1-2) 171.61 1 ***
Integrity / Null Model (N) 2,595.94 55 47.20 n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a.
Table 25: Pairings of Sub-dimensions (Sample 3a)

Willingness
to rely One Factor Model (1) 555.30 44 .001 (N-1) 2,040.64 11 *** 12.62 .18 .79 .75 .80 .16 .10 .78 .68
Two Factor Model (2) 89.47 43 .001 (N-2) 2,506.47 12 *** 2.08 .05 .97 .98 .98 .05 .03 .96 .94

(1-2) 465.83 1 ***

Integrity / Null Model (N) 2,633.58 55 47.88 n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a.
Willingness
to depend One Factor Model (1) 457.63 44 .001 (N-1) 2,175.95 11 *** 10.40 .16 .83 .80 .84 .07 .07 .82 .74
Two Factor Model (2) 128.59 43 .001 (N-2) 2,504.99 12 *** 2.99 .07 .95 .96 .97 .07 .04 .95 .92
(1-2) 329.05 1 ***
Ability / Null Model (N) 1,832.20 45 40.72 n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a.
Benevolence
One Factor Model (1) 320.82 35 .001 (N-1) 1,511.38 10 *** 9.17 .15 .83 .79 .84 .19 .10 .85 .76
Two Factor Model (2) 93.38 34 .001 (N-2) 1,738.82 11 *** 2.75 .07 .95 .96 .97 .09 .06 .96 .93
(1-2) 227.45 1 ***
Ability / Null Model (N) 2703,56 45 60.08 n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a.
Willingness
to rely One Factor Model (1) 399,33 35 .001 (N-1) 2304,23 10 *** 11.41 .19 .85 .82 .86 .10 .07 .78 .66
Two Factor Model (2) 85,32 34 .001 (N-2) 2618,24 11 *** 2.51 .06 .97 .97 .98 .05 .03 .96 .94
(1-2) 314,01 1 ***
(Continued on next page.)
 Chi Square Difference Model Fit Indices
Sub-dimension Chi Chi
Model df p Competing df Chi
Pairing Square Square Sign. RMSEA NFI NNFI CFI RMR SRMR GFI AGFI
Models Difference Square / df
Difference
Ability / Null Model (N) 2,182.27 36 60.62 n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a.
Willingness to
depend One Factor Model (1) 402.74 27 .001 (N-1) 1779,53 9 *** 14.92 .18 .82 .77 .83 .18 .08 .83 .71
Two Factor Model (2) 60.03 26 .001 (N-2) 2122,24 10 *** 2.31 .05 .97 .98 .98 .05 .03 .97 .95
(1-2) 342,71 1 ***
Benevolence / Null Model (N) 1,488.92 28 53.18 n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a.
Willingness to
rely One Factor Model (1) 341.79 20 .001 (N-1) 1147,13 8 *** 17.09 .22 .77 .69 .78 .28 .15 .80 .63
Two Factor Model (2) 43.93 19 .001 (N-2) 1444,99 9 *** 2.31 .06 .97 .98 .98 .10 .05 .97 .95
(1-2) 297,86 1 ***
Benevolence / Null Model (N) 1,104.34 21 52.59 n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a.
Willingness to
depend One Factor Model (1) 295.73 14 .001 (N-1) 808,61 7 *** 21.12 .25 .73 .61 .74 .28 .15 .80 .60
Two Factor Model (2) 42.21 13 .001 (N-2) 1062,13 8 *** 3.25 .07 .96 .96 .97 .12 .06 .97 .94
Table 25 (Cont.): Pairings of Sub-dimensions (Sample 3a)

(1-2) 253,52 1 ***

Willingness to Null Model (N) 1,966.32 21 93.63 n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a.
rely /
Willingness to One Factor Model (1) 371.99 14 .001 (N-1) 1594,33 7 *** 26.57 .26 .81 .72 .82 .20 .09 .78 .57
depend
Two Factor Model (2) 75.80 13 .001 (N-2) 1890,52 8 *** 5.83 .10 .96 .95 .97 .10 .05 .96 .91
(1-2) 296,19 1 ***
Notes: *** = p <.001; n.a. = not available.

231
 Chi Square Difference Model Fit Indices

232
Sub-
Chi Chi Chi
dimension Model df p Competing df
Square Square Sign. Square RMSEA NFI NNFI CFI RMR SRMR GFI AGFI
Pairing Models Difference
Difference / df
Integrity / Null Model (N) 3,122.84 91 34.32 n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a.
Ability
One Factor Model (1) 685.94 77 .001 (N-1) 2,436.90 14 *** 8.91 .19 .78 .76 .80 .12 .09 .71 .60
Two Factor Model (2) 162.47 76 .001 (N-1) 2,960.37 15 *** 2.14 .05 .95 .97 .97 .06 .04 .95 .92
(1-2) 523.47 1 ***
Integrity / Null Model (N) 2,009.13 66 30.44 n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a.
Benevolence
One Factor Model (1) 320.18 54 .001 (N-1) 1,688.95 12 *** 5.93 .12 .84 .83 .86 .13 .08 .87 .81
Two Factor Model (2) 123.52 53 .001 (N-1) 1,885.61 13 *** 2.33 .06 .94 .95 .96 .06 .04 .95 .93
(1-2) 196.66 1 ***
Integrity / Null Model (N) 2,245.10 55 40.82 n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a.
Willingness
Table 26: Pairings of Sub-dimensions (Sample 3b)

to rely One Factor Model (1) 602.75 44 .001 (N-1) 1.642.36 11 *** 13.70 .18 .73 .68 .75 .18 .11 .78 .67
Two Factor Model (2) 82.80 43 .001 (N-1) 2,162.31 12 *** 1.93 .05 .96 .98 .98 .04 .03 .96 .94
(1-2) 519.95 1 ***
Integrity / Null Model (N) 2,187.78 55 39.78 n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a.
Willingness
to depend One Factor Model (1) 413.29 44 .001 (N-1) 1,774.49 11 *** 9.39 .14 .81 .78 .83 .18 .08 .85 .77
Two Factor Model (2) 99.80 43 .001 (N-1) 2,087.98 12 *** 2.32 .06 .95 .97 .97 .07 .04 .96 .94
(1-2) 313.49 1 ***
Ability / Null Model (N) 1,864.92 45 41.44 n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a.
Benevolence
One Factor Model (1) 381.59 35 .001 (N-1) 1,483.33 10 *** 10.90 .18 .80 .76 .81 .23 .13 .81 .71
Two Factor Model (2) 87.94 34 .001 (N-1) 1,776.98 11 *** 2.59 .06 .95 .96 .97 .09 .06 .96 .94
(1-2) 293,65 1 ***
Ability / Null Model (N) 2,681.00 45 59.58 n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a.
Willingness
to rely One Factor Model (1) 468.05 35 .001 (N-1) 2,212.95 10 *** 13.37 .22 .83 .79 .84 .12 .08 .74 .59
Two Factor Model (2) 38.21 34 .284 (N-1) 2,642.79 11 *** 1.12 .02 .99 1.00 1.00 .03 .02 .98 .97
(1-2) 429.84 1 ***
(Continued on next page.)
 Chi Square Difference Model Fit Indices
Sub-dimension Chi Chi Chi
Model df p Competing df
Pairing Square Square Sign. Square RMSEA NFI NNFI CFI RMR SRMR GFI AGFI
Models Difference
Difference / df
Ability / Null Model (N) 2,100.59 36 58.35 n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a.
Willingness to
depend One Factor Model (1) 402.65 27 .001 (N-1) 1,697.94 9 *** 14.91 .18 .81 .76 .82 .24 .10 .83 .71
Two Factor Model (2) 60.13 26 .001 (N-1) 2,040.46 10 *** 2.31 .05 .97 .98 .98 .09 .05 .97 .95
(1-2) 342.52 1 ***
Benevolence / Null Model (N) 1,481.34 28 52.90 n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a.
Willingness to
rely One Factor Model (1) 365.79 20 .001 (N-1) 1,115.54 8 *** 18.29 .23 .75 .67 .76 .29 .16 .79 .62
Two Factor Model (2) 49.27 19 .001 (N-1) 1,432.06 9 *** 2.59 .06 .97 .97 .98 .08 .05 .97 .95

(1-2) 316.52 1 ***

Benevolence / Null Model (N) 981.21 21 46.72 n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a.
Willingness to
depend One Factor Model (1) 290.44 14 .001 (N-1) 690.77 7 *** 20.75 .25 .70 .57 .71 .27 .15 .80 .59
Table 26 (Cont.): Pairings of Sub-dimensions (Sample 3b)

Two Factor Model (2) 15.99 13 .001 (N-1) 965.22 8 *** 1.23 .02 .98 1.00 1.00 .07 .03 .99 .98
(1-2) 274.45 1 ***
Willingness to Null Model (N) 1,813.34 21 86.35 n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a.
rely /
Willingness to One Factor Model (1) 309.39 14 .001 (N-1) 1,503.95 7 *** 22.10 .23 .83 .75 .84 .22 .09 .83 .65
depend
Two Factor Model (2) 44.66 13 .001 (N-1) 1,768.68 8 *** 3.44 .07 .98 .97 .98 .09 .04 .97 .94
(1-2) 264.73 1 ***
Notes: *** = p <.001; n.a. = not available.

233
Results Validity Stage

For the validity stage, data was gathered from a new sample deemed to be representative of the
scale’s target population, in order to evaluate the consistency of the results between studies and
to re-examine scale properties. During reliability stage, a split sample was used for scale
development. While one half of the sample served as the primary development sample, results
were cross-checked with the holdout sample which was used to replicate the findings. As
discussed, one can infer with a certain degree of comfort that the scale properties are not
distorted by chance, as the results of this stage were quite promising. However, additional new
data had to be collected to avoid shortcomings due to scale development out of a single sample
and to comply with scientific standards (DeVellis, 2012; MacKenzie et al., 2011).

Insights from samples 3a and 3b could be distorted due to a number of reasons. First, the two
sub-samples which were divided randomly from the entire reliability sample very likely
represent a homogeneous but very specific population (i.e., Austrian university students).
Whereas this population has some very appealing characteristics (e.g., high familiarity with the
trust object), it is desirable to compare the scale properties to a different sample which is more
representative for the online consumer population in the two German speaking countries.
Second, the data collection period for the two sub-samples was not separated by time and it can
be assumed that any special condition under which data was collected in the reliability stage
may equally apply for both samples. In order to rule out such circumstances restricting scale
generalizability and internal validity (e.g., history effects), the collection of new data sets was
deemed reasonable.

Particularly, two different samples were used for scale finalization. The first one (sample 4; n
= 526) was taken from an online panel and provided data for conduct EFA as well as additional
item analyses as precursors to confirmatory factor analysis, and to re-estimate the model by
using CFA to finalize and confirm the hypothesized structure of the higher-order construct over
multiple data sets. A second sample (samples 5(a-e); n = 824) consisting of pooled data from a
series of interviews among (Austrian) online consumers served cross-validation purposes.
Together with additional data, this sample enabled this research to exercise supplemental
validity and reliability testings, establish norms, and to apply generalizability theory.

Exploratory Factor Analysis

For item reduction and testing the dimensionality of the construct, a principal component factor
analysis (PCA) was again carried out on the 25 manifest indicators after the factorability of the
data was verified. EFA resulted in a definite factor structure similar to this research’s earlier
findings. All items – except Ab6 (“relevant”) – loaded considerably on their intended
components. During the reliability stage, this item showed a clear relationship with the ability
dimension. However, according to data from the new sample, the item is strongly associated

234
with the integrity/honesty dimension. Hence, this item was considered ambiguous and was
deleted.

Subsequently, by evaluating the data by the various decision rules described earlier, the number
of factors was restricted to five and the model was re-estimated. This resulted in the explanation
of 73.70% of the total variance. All remaining 24 items exhibited significant loadings on their
appropriate component of the sub-dimensions of eWOM trust (see Table 27). Due to sample
size, factor loadings of .40 and higher were regarded as significant (Hair et al., 2010). Two
items (In7, Be6) showed significant, but relatively weak loadings (.43 and .44) and also low
communalities; however, the rest of the items passed the .60 benchmark. Considerable cross-
loadings (>.32) were not detected.

Table 27: Results of the EFA (Sample 4)

Variance
Sub-
Explained/ Item Factor Loading Communality MSA
dimension
Eigenvalue
Ability 4.80% / 1.15 Ab7 .61 .78 .96
Ab8 .57 .71 .98
Ab9 .92 .80 .96
Ab10 .61 .76 .98
Ab11 .87 .76 .98
Integrity/ 51.51% / 12.36 In2 .78 .74 .98
Honesty In3 .87 .73 .97
In4 .98 .80 .97
In5 .73 .75 .97
In6 .81 .79 .97
In7 .43 .59 .99
In9 .77 .69 .98
In10 .92 .64 .98
Benevolence 3.87% / .93 Be1 .90 .77 .90
Be2 .76 .70 .94
Be3 .89 .75 .91
Be6 .44 .50 .97
Willingness to 10.44% / 2.51 Wi1 .94 .83 .93
rely Wi4 .90 .79 .95
Wi5 .89 .79 .95
Wi8 .84 .80 .96
Willingness to 3.07% / .74 Wi2 .86 .74 .96
depend Wi6 .70 .73 .97
Wi7 .77 .76 .96
Notes: Total variance explained: 73.70%; Extraction method: Principal Component Analysis; Rotation method:
Promax with Kaiser Normalization; Rotation converged in 7 iterations; Kaiser-Meyer-Olkin Measure of
Sampling Adequacy (MSA) .96 and Bartlett’s test of Sphericity: sig. .001; Effective sample size: n=526; Item
to respondent ratio: approx. 1:21.

235
In addition to the explorative factor analysis, the individual item properties were examined. All
items demonstrated desirable qualities advanced by established scale development guidelines
(DeVellis, 2012), including means that are close to the center of the theoretical range, relatively
high item variances, and the correspondence of the observed item ranges to the theoretical
ranges.

Confirmatory Factor Analysis

Model Goodness-of-fit

Consequently, the new data were subject to a CFA in order to validate the hypothesized
structure of the eWOM trust construct. The adopted model was specified so that each of the 24
items exclusively load on one of the five sub-dimensions which were identified in the earlier
factor analysis. While the chi-square test was highly significant (χ2 = 773.84, df = 247, p <
.001), the absolute, incremental, and parsimonious fit indices suggest that the hypothesized
construct is acceptable (absolute fit indices: GFI = .89, AGFI = .86, RMSEA = .065, RMR =
.10, SRMR = .06; incremental fit indices: CFI = .95, NNFI = .94, NFI = .92; parsimonious fit
indices: normed chi-square: 3.1, AIC = 898.40, PIF = .73).

Reliability

The reliability statistics are provided in Table 28. All observable indicators were found to be
significant (t-values ranging from 13.36 to 26.71 and, therefore, passing the critical value for
.01 significance) and considerably related to their specified constructs. This is indicative of the
existence of the posited relationship between the sub-dimensions and their indicators. Further
evidence for the indicator’s reliability was gained by reviewing the items’ squared multiple
correlations, which turned out to be mainly over or near to the limit of .50. Only one item (Be6)
was found to possess a slightly lower R2 of .40, which corresponded to earlier findings that had
indicated a problematic low explanation of the item’s variance by the benevolence construct.
The squared multiple correlation of the rest of the individual items ranged from .52 to .77. All
sub-dimensions exhibited a high level of internal consistency: The Cronbach’s alphas varied
from .81 to .94, the average inter-item correlation from .52 to .74, and the construct reliabilities
ranged from .82 to .94. Hence, minimum internal consistency levels (i.e., .70 for Cronbach’s
alpha and construct reliability; .50 for inter-item correlation) were achieved. Also the AVE met
the recommended .50 threshold (Diamantopoulos & Siguaw, 2000; Hair et al., 2010) for all
trust components. Here the Benevolence dimension had with .52 the lowest AVE value.
Nevertheless, in the validation samplem the majority of its variance is still explained by the
higher-order construct. The corrected item-to-total correlations of the individual items were all
above .50.

236
Evidence for reliability of the first-order constructs as indicators of the second-order construct
was given by positive and significant loadings on the higher-order construct. Completely
standardized loadings (γ) ranged from .64 to .93. Similarly to the reliability samples, the squared
multiple correlation of the benevolence dimension was slightly below the proposed cut-off
value of .50 (R2 = .43), while the remaining constructs demonstrated adequate levels with .53
or better. An AVE of .91 and a construct reliability of .91 both supported the internal
consistency of the second-order construct.

Convergent Validity

As given in Table 28, the factor loadings of all items were greater than .6 and all were
statistically significant at the .01 significance level, giving weak support for convergent
validity. In the validity sample, factor loadings of all observable indicators were equal to or
greater than .63. Evidence for strong support of the item’s convergent validity can be obtained
by reviewing their squared factor loadings, which are desired to achieve the cut-off level of .5
or higher. Squared factor loadings for all but one item (Be6, λ2 = .40) turned out to accomplish
this threshold, as the majority of items demonstrated squared loadings in the .80s.

The validity of each first-order sub-dimension as an indicator of the higher-order construct was
evaluated similarly. All five-factor loadings were considerably high, varying from .64 to .93
and significant (range of t-values: 12.43 to 20.94). Four out of five first-order constructs turned
out to have a squared factor loading (γ) equal to or greater than .53. In line with the findings of
the reliability stage, the benevolence dimension turned out to represent a somewhat limited
effect of eWOM trust from this perspective. Additional evidence for convergent validity is
shown in Table 29, which includes the correlations between the components of eWOM trust
which are all significant, varying from .47 to .83. Therefore, one is able to conclude that the
five sub-dimensions all measure distinct aspects of the same construct.

Discriminant Validity

Similar to the earlier research, stage three alternative tests were undertaken in order to assess
discriminant validity of the five sub-dimensions: first, a test for the confidence interval of the
factors’ correlations; second, a test that compares the factor’s shared variance with other factors
in the model relative to the AVE by each factor (Fornell & Larcker, 1981); the third test for
discriminant validity involved the evaluation of competing factor models by means of chi-
square difference tests and the evaluation of fit indices.

237
238
Squared Corrected Average
Sub- Cronbach’s Construct Construct
γ Item λ Multiple Item-to-Total Inter-Item AVE AVE
dimension Alpha Reliability Reliability
Correlation Correlation Correlation
Ab7 .85 .72 .80
Ab8 .81 .66 .76
Ability .93 Ab9 .82 .71 .81 .70 .92 92 .70
Ab10 .86 .73 .80
Ab11 .82 .67 .78
In2 .82 .67 .80
In3 .83 .68 .80
In4 .86 .73 .84
Integrity/ In5 .85 .72 .81
.90 .65 .94 .94 .66
Honesty In6 .87 .76 .84
In7 .73 .53 .69
In9 .81 .65 .78
In10 .72 .52 .70 .91 .67
Be1 .78 .60 .70
Be2 .75 .56 .63
Benevolence .64 .52 .81 .82 .53
Be3 .75 .57 .67
Be6 .63 .40 .53
Wi1 .88 .77 .83
Willingness Wi4 .85 .73 .81
Table 28: Psychometric Properties of the eWOM Trust Scale (Sample 4)

.73 .74 .92 .92 .74

to rely Wi5 .85 .72 .81
Wi8 .86 .75 .82
Wi2 .71 .50 .64
Willingness
.86 Wi6 .83 .68 .70 .62 .83 .84 .63
to depend
Wi7 .84 .70 .73
Notes: γ = Completely standardized second-order loading; λ = Completely standardized first-order loading; AVE = Average variance extracted.
Table 29 presents the correlations, the squared correlations, as well as the upper bound of the
correlations’ 95 percent confidence intervals between the factors. To illustrate, the upper bound
of the confidence interval was .63 for the benevolence and integrity/honesty correlation, .72 for
the willingness to rely and ability correlation, and .61 for the willingness to depend and
benevolence correlation. The highest upper bound of the confidence interval was achieved for
the correlation between the ability and the integrity/honesty dimensions with .86. The two upper
bounds of the correlation interval for willingness to depend and integrity/honesty (.81) and
willingness to depend and ability (.83) were only slightly lower. However, no factor pairing
was found to be perfectly correlated, as none of the identified confidence intervals included 1.0.
This result supported the discriminability of the five factors. To further evaluate discriminant
validity for each possible factor pairing, the factors’ shared variance was compared with the
AVE of the involved factors. It has been suggested that if the square of the parameter estimates
between two constructs is less than the average variance extracted of the two constructs,
discriminant validity is supported (Fornell & Larcker, 1981). In nine out of ten possible factor
pairings, this test led to desirable results, as the squared correlation between those factors was
judged lower than the respective AVEs. However, the second test revealed some problems with
discriminant validity, as the shared variance between ability and integrity/honesty (.69) was
equal to the AVE of the integrity/honesty dimension. A similar pattern was also detected in one
sample during the reliability stage. Therefore, while the dimensionality of the remaining factors
was without question, the discriminability of these two factors was potentially at risk.

Table 29: Correlations among the Sub-dimensions (Sample 4)

Integrity/ Willingness to Willingness to

Ability Benevolence
Honesty rely depend
Integrity/
0.69a
Honesty
0.83b
Ability 0.86c 0.70a
0.69d
0.58b 0.60b
Benevolence 0.63c 0.65c 0.53a
0.33d 0.36d
0.65b 0.67b 0.47b
Willingness to
0.70c 0.72c 0.53c 0.74a
rely
0.43d 0.45d 0.22d
0.77b 0.80b 0.55b 0.63b
Willingness to
0.81c 0.83c 0.61c 0.68c 0.63a
depend
0.60d 0.64d 0.31d 0.39d
Notes: a = Average variance extracted (AVE); b = Pearson correlation coefficient (r); c = Upper bound of 95%
confidence interval of r; d = Squared correlation (r2); n = 526.

239
Evidence for discriminant validity of the five sub-dimensions was provided by the assessment
of the chi-square difference statistics and fit indices of several competing models. Here, ten
alternative measurement models were proposed with variable factor combinations. Due to
considerable correlations among the ability, integrity/honesty, and willingness to depend
dimensions (.83 between Ability and integrity/honesty; .80 between ability and willingness to
depend; and .77 between integrity/honesty and willingness to depend), they were combined in
a single factor which is hypothesized to coexist with the benevolence and the willingness to rely
as first-order factors in a second-order model (Model 3). Additionally, five four-factor models
(Model 4a-4e) were estimated, where two first-order constructs were combined due to their
considerable intercorrelation while treating the remaining factors separately: Model 4a: ability
and integrity/honesty (r = .83); Model 4b: ability and willingness to depend (r = .80); Model 4c:
integrity/honesty and willingness to depend (r = .77); Model 4d: integrity/honesty and
willingness to rely (r = .65); and Model 4e: willingness to rely and willingness to depend (r =
.63). Also a one-factor model (Model 1) where eWOM trust was hypothesized as a
unidimensional construct and a null model (Model N) that assumed no systematic relationship
within the data were specified, along with the advocated second-order construct consisting of
five separate sub-dimensions (Model 5). As shown in Table 30, Model 5 outperformed the other
conceptualizations by having the lowest chi-square statistic (χ2 = 774.84, df = 247, p < .001)
and also met levels of satisfactory model fit (RMSEA = .06, NFI = .92, NNFI = .94, CFI = .95,
RMR = .11, SRMR = .06, GFI = .89, AGFI = .86). This research was also able to demonstrate
that, when comparing the five factor model to alternative models, chi-square difference
statistics indicate a significantly better fit.

A final assessment of the discriminant validity among the five sub-dimensions of eWOM trust
was applied using confirmatory factor analysis for the comparison of alternative factor pairings.
For each possible pair, a two-factor model of which the assumption that the two factors are
separate from each other formed the basis, was compared to a one-factor model that handled
the two factors as a unidimensional construct. Another frame of reference was the null model,
for which it was hypothesized that the items belonging to the two factors did not have any
statistically relevant relationship.

As shown in Table 31, the two-factor models generally were superior to the single factor
models. For instance, the two-factor model for integrity/honesty and ability appeared to have a
chi-square statistic (χ2 = 184.43; df = 64) which was considerably lower compared to the one-
factor alternative (χ2 = 627.37; df = 65), considering a common underlying construct for both
dimensions. Accordingly, the chi-square difference test was significant (χ2Diff = 442.94; df Diff
= 1; p < .001), which guided this research together with evident improvements in the fit indices
to the acceptance that data could be best described by the more complex two-factor structure.
As such a pattern was consistent throughout all possible pairings, considerable support in favor
of the hypothesized dimensionality of the first-order constructs was gained.

240
 Chi Square Difference Model Fit Indices
Chi
Model df p Competing Chi Square df Chi Square /
Square Sign. RMSEA NFI NNFI CFI RMR SRMR GFI AGFI
Models Difference Difference df
Null Model (N) 9,934.83 276 36.00 n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a.
One Factor Model (1) 2,598.86 252 .001 10.31 .18 .74 .73 .76 .15 .09 .59 .51

(N-1) 7,335.97 24 ***

Three Factor Model 1,362.67 249 .001 5.47 .11 .86 .87 .89 .11 .07 .77 .72
(3) (A/I/W2; B; W1)
(N-3) 8,572.16 27 ***
(1-3) 1,236.19 3 ***
Table 30: Competing Models (Sample 4)

Four Factor Model 1,230.70 248 .001 4.96 .11 .88 .89 .90 .11 .07 .78 .73
(4a) (A/I; B; W1; W2)
(N-4a) 8,704.13 28 ***
(1-4a) 1,368.16 4 ***
(3-4a) 131.97 1 ***
Four Factor Model 981.33 248 .001 3.96 .08 .90 .92 .92 .13 .08 .86 .83
(4b) (A/W2; I; B; W1)
(N-4b) 8,953.50 28 ***
(1-4b) 1,617.53 4 ***
(3-4b) 381.34 1 ***
Four Factor Model 990.32 248 .001 3.99 .08 .90 .91 .92 .13 .08 .85 .82
(4c) (I/W2; A; B; W1)
(N-4c) 8,944.51 28 ***
(1-4c) 1,608.54 4 ***
(3-4c) 372.35 1 ***
(Continued on next page.)

241
242
Chi Square Difference Model Fit Indices
Chi
Model df p Competing Chi Square df Chi Square /
Square Sign. RMSEA NFI NNFI CFI RMR SRMR GFI AGFI
Models Difference Difference df
Four Factor Model (4d) 1,849.14 248 .001 7.46 .14 .81 .82 .83 .14 .09 .69 .63
(I/W1; I; B; W2)
(N-4d) 8,085.69 28 ***
(1-4d) 749.72 4 ***
(3-4d) -486.48 1 n.s.
Four Factor Model (4e) 1,191.77 248 .001 4.81 .10 .88 .89 .90 .17 .10 .81 .76
(I; A;B; W1/W2)
(N-4e) 8,743.06 28 ***
(1-4e) 1,407.09 4 ***
(3-4e) 170.90 1 ***
Five Factor Model (5) 774.84 247 .001 3.14 .06 .92 .94 .95 .11 .07 .89 .86
Table 30 (Cont.): Competing Models (Sample 4)

(N-5) 9,159.99 29 ***

(1-5) 1,824.02 5 ***
(3-5) 587.83 2 ***
(4a-5) 455.86 1 ***
(4b-5) 206.49 1 ***
(4c-5) 215.48 1 ***
(4d-5) 1,074.31 1 ***
(4e-5) 416.93 1 ***
Notes: A = Ability; I = Integrity/Honesty; B = Benevolence; WR = Willingness to rely; WD = Willingness to depend; *** = p <.001; n.s. = not significant; n.a. = not available.
 Chi Square Difference Model Fit Indices
Sub-dimension Chi
Model df p Competing Chi Square df Chi Square
Pairing Square Sign. RMSEA NNFI CFI RMR SRMR AGFI
Models Difference Difference / df
Integrity / Null Model (N) 5,582.59 78 71.57 n.a. n.a. n.a. n.a. n.a. n.a.
Ability
One Factor Model (1) 627.37 65 .001 (N-1) 4,955.23 13 *** 9.65 .16 .88 .90 .08 .06 .70
Two Factor Model (2) 184.43 64 .001 (N-2) 5,398.17 14 *** 2.88 .06 .97 .98 .04 .03 .92
(1-2) 442.94 1 ***
Integrity / Null Model (N) 4,264.64 66 64.62 n.a. n.a. n.a. n.a. n.a. n.a.
Benevolence
One Factor Model (1) 452.41 54 .001 (N-1) 3,812.23 12 *** 8.38 .13 .88 .91 .13 .07 .80
Two Factor Model (2) 190.69 53 .001 (N-2) 4,073.95 13 *** 3.60 .07 .96 .97 .08 .05 .92
(1-2) 261.72 1 ***
Integrity / Null Model (N) 4,934.08 66 74.76 n.a. n.a. n.a. n.a. n.a. n.a.
Willingness to rely
Table 31: Pairings of Sub-dimensions (Sample 4)

One Factor Model (1) 1,205.73 54 .001 (N-1) 3,728.35 12 *** 22.33 .24 .71 .76 .19 .13 .51
Two Factor Model (2) 122.89 53 .001 (N-2) 4,811.19 13 *** 2.32 .05 .98 .99 .05 .03 .94
(1-2) 1,082.84 1 ***
Integrity / Null Model (N) 4,204.70 55 76.45 n.a. n.a. n.a. n.a. n.a. n.a.
Willingness to
depend One Factor Model (1) 306.68 44 .001 (N-1) 3,898.02 11 *** 6.97 .11 .92 .94 .10 .05 .84
Two Factor Model (2) 106.79 43 .001 (N-2) 4,097.92 12 *** 2.48 .05 .98 .99 .04 .02 .97
(1-2) 199.90 1 ***
Ability / Null Model (N) 2,769.95 36 76.94 n.a. n.a. n.a. n.a. n.a. n.a.
Benevolence
One Factor Model (1) 538.88 27 .001 (N-1) 2,231.07 9 *** 19.96 .21 .75 .81 .23 .13 .63
Two Factor Model (2) 132.84 26 .001 (N-2) 2,637.11 10 *** 5.11 .09 .95 .96 .10 .06 .91
(1-2) 406.04 1 ***
Ability / Null Model (N) 3,794.88 36 105.41 n.a. n.a. n.a. n.a. n.a. n.a.
Willingness to rely
One Factor Model (1) 622.38 27 .001 (N-1) 3,172.50 9 *** 23.05 .26 .79 .84 .11 .08 .52
Two Factor Model (2) 101.95 26 .001 (N-2) 3,692.93 10 *** 3.92 .08 .97 .98 .03 .02 .93
(1-2) 520.43 1 ***
(Continued on next page.)

243
 Chi Square Difference Model Fit Indices

244
Sub-dimension Chi Chi
Model df p Competing df Chi Square
Pairing Square Square Sign. RMSEA NNFI CFI RMR SRMR AGFI
Models Difference / df
Difference
Ability / Null Model (N) 2,798.41 28 99.94 n.a. n.a. n.a. n.a. n.a. n.a.
Willingness to
depend One Factor Model (1) 274.29 20 .001 (N-1) 2,524.13 8 *** 13.71 .17 .87 .91 .07 .07 .77
Two Factor Model (2) 53.89 19 .001 (N-2) 2,744.52 9 *** 2.84 .06 .98 .99 .03 .02 .95
(1-2) 220.40 1 ***
Benevolence / Null Model (N) 2,345.50 28 83.77 n.a. n.a. n.a. n.a. n.a. n.a.
Willingness to rely
One Factor Model (1) 703.47 20 .001 (N-1) 1,642.03 8 *** 35.17 .28 .59 .71 .35 .19 .48
Two Factor Model (2) 85.03 19 .001 (N-2) 2,260.47 9 *** 4.48 .08 .96 .97 .10 .06 .93
(1-2) 618.44 1 ***
Benevolence / Null Model (N) 1,522.09 21 72.48 n.a. n.a. n.a. n.a. n.a. n.a.
Willingness to
depend One Factor Model (1) 379.86 14 .001 (N-1) 1,142.23 7 *** 27.13 .26 .63 .76 .21 .10 .57
Table 31 (Cont.): Pairings of Sub-dimensions (Sample 4)

Two Factor Model (2) 60.21 13 .001 (N-2) 1,461.87 8 *** 4.63 .08 .95 .97 .10 .05 .93
(1-2) 319.65 1 ***
Willingness to rely Null Model (N) 2,415.06 21 115.00 n.a. n.a. n.a. n.a. n.a. n.a.
/ Willingness to
depend One Factor Model (1) 347.93 14 .001 (N-1) 2,067.13 7 *** 24.85 .23 .79 .86 .21 .10 .65
Two Factor Model (2) 37.52 13 .001 (N-2) 2,377.54 8 *** 2.89 .06 .98 .99 .03 .02 .96
(1-2) 310.41 1 ***
Notes: *** = p <.001; n.a. = not available.
Measure Purification and Model Modification

In order to purify the new scale, two items were removed in a stepwise procedure. First, Be6
was excluded from the scale due to a weak relationship with its hypothesized sub-dimension,
manifested by a squared multiple correlation of less than .50. This pattern constantly appeared
in all samples across the two stages of the scale development process. Consequently, the new
model was re-estimated and its fit was assessed. The inspection of the modification indices and
the expected parameter changes revealed that In7 showed large and significant cross-loadings
on several non-hypothesized sub-dimensions and correlated measurement errors with other
items. Such patterns violate the basic assumptions of classical test theory. Correlated
measurement errors are especially critical when a substantial number of items is affected,
because this may threaten the dimensionality of the construct (Netemeyer et al., 2003). As In7
turned out to be a problematic indicator across all samples, the item was removed. This resulted
in a purified second-order measurement model which comprised a total of 22 items. The
remaining item pool still exhibited content validity; therefore, the removal of the items
exhibiting minor statistical quality was deemed acceptable (Hildebrandt & Temme, 2006).
Three of the five sub-constructs are measured with a minimum of three items. This accounts
for the insight that entropy increases only slightly when additional items are included (Albers
& Hildebrandt, 2006). The two remaining sub-constructs certainly have some internal
consistency to spare, but as they are key concepts of the eWOM trust and a conservative
approach was preferred, this research refrained from further item elimination, as the overall
scale number is still practicable.

Results Validity Stage (Modified Model)

For testing the modified model, it was specified in LISREL with data taken from sample 4.
Several scholars recommend that if model modifications were made during the scale
development and refinement process, the final model should be cross-validated by using a new
sample (e.g., MacKenzie et al., 2011; Netemeyer et al., 2003; Worthington & Whittaker, 2006).
While this rarely is done in scale research, the study at hand collected new data (sample 5(a-e);
n = 824) from Austrian online consumers in order to confirm scale reliability and convergent
validity. This data was gathered with a series of street-intercept interviews, throughout using a
paper-and-pencil questionnaire. It was expected that the construct of eWOM trust is likely to
percist in alternative population.

Model Goodness-of-fit

Model validity was supported, as the solution was proper, all hypothesized relationships were
statistically significant, none of the items showed negative error variances, and the completely
standardized parameter estimates were all greater than one. The purified model showed a
satisfactory fit. While the chi-square statistic was significant (χ2 = 627.59 (822.45), df = 204, p

245
< .001) the absolute, incremental and parsimonious fit measures suggested that the proposed
second-order model, with its five first-order constructs measured by 22 observable indicators
adequately represented the structure within the sample data (GFI = .90 (.89), AGFI = .88 (.85),
RMSEA = .057 (.064), RMR = .09 (.11), SRMR = .06 (.07); CFI = .95 (.92), NNFI = .95 (.91),
NFI = .93 (.91); normed chi-square: 3.1 (4.0), PIF = .73 (.71)).

Reliability

The measures to assess the construct’s reliability are shown in Table 32. All loadings of the
items turned out to be high (.73 to .88) but did not show extreme values over .90. All t-values
associated with each of the loadings exceeded the critical values for .01 significance, implying
that the included items are reliable indicators of their hypothesized constructs. This is also
supported by data from the validation sample, as all items also achieved similar loadings
(ranging from .66 to .90) and significance levels. The items also demonstrated notable corrected
item-to-total correlations, varying from .62 (.56) to .84 (.85). Further evidence for the reliability
on the item level was provided by the items’ squared multiple correlation (R2), which were all
equal to or greater than the recommended cut-off value of .50 (Fornell & Larcker, 1981).

To assess reliability of the constructs on the first-order level Cronbach’s alpha, construct
reliability, the average inter-item-correlation and the average variance extracted were
evaluated. All first-order constructs evidenced remarkable alpha levels: ability: .92 (.91),
integrity/honesty: .94 (.91), benevolence: .81 (.79), willingness to rely: .92 (.93), and willingness
to depend: .83 (.82), meaning that the internal consistency of almost all constructs (except
benevolence) was above the more restrictive alpha level of .80 proposed for newly developed
constructs (Netemeyer et al., 2003). Construct reliabilities ranging from .81 (.81) to .93 (.94),
as well as average item-to-total correlations varying from .59 (.55) to .74 (.77), provided
additional evidence for the reliability of the five constructs. The AVE were all equal to or
greater than .59 (.56), hence the majority of the sub-dimensions variance was explained by the
higher-order construct. Every first-order construct was reasonably and significantly related with
the second-order construct. Here, factor loadings fluctuated between .70 (.67) and .93 (.87).
The squared multiple correlations were also mainly the limit of .50. Only the benevolence
dimension showed a lower level in the development sample (.34), but only a slightly lower level
in the validation sample (.45). On the second-order level, the reliability of the first-order
indicators was also assessed by the index of construct reliability and the AVE. The internal
consistency measures exceeded the .80 threshold (Netemeyer et al., 2003) with values of .90
(.88) and also the .50 cut-off value (Fornell & Larcker, 1981) (.65/.61). This implied that the
first-order constructs are commonly reliable measures of the higher-order construct.

246
Convergent Validity

As presented in Table 32, factor loadings of all items turned out be greater than .71 (.66), and
every one was statistically significant at the .01 significance level. A similar pattern was
observable for the relationship between the first-order and the second-order constructs,
supporting on both levels evidence for convergent validity (Bagozzi & Yi, 1991). More
stringent evidence for convergent validity is provided where the squared factor loading between
the item and its hypothesized sub-construct and/or the loadings between the sub-construct and
the higher-order construct is .5 or better. On the item level, the lowest squared loading (.50)
was observed for Wi2. The remaining items, however, showed loadings typically in the .70
range. In the validation sample, the same item (besides In3) also achieved the lowest loading
but which was slightly above the one in the development sample (.53) encouraging the item’s
validity. On the first-order level, almost all constructs turned out to have a squared factor
loading equal to or greater than .50. However, benevolence showed a slightly lower loading
(.45) only in the validation sample. Second, validity on the first-factor level was supported by
significant correlations among the five constructs (see Table 32), implying that they all measure
the same construct.

247
 Squared Corrected Average

248
Cronbach’s Construct Construct
Sub-dimension γ Item λ Multiple Item-to-Total Inter-Item AVE AVE
Alpha Reliability Reliability
Correlation Correlation Correlation
Ab7 .85 (.83) .72 (.69) .80 (.78)
Ab8 .81 (.82) .66 (.68) .76 (.78)
Ability .93 (.87) Ab9 .84 (.83) .71 (.69) .81 (.80) .70 (.68) .92 (.91) .92 (.92) .70 (.68)
Ab10 .85 (.85) .73 (.72) .80 (.80)
Ab11 .82 (.80) .67 (.64) .78 (.76)
In2 .82 (.77) .67 (.59) .79 (.73)
In3 .83 (.73) .69 (.53) .80 (.70)
In4 .86 (.81) .74 (.66) .83 (.78)
Integrity/
.88 (.83) In5 .85 (.82) .72 (.68) .81 (.77) .68 (.60) .94 (.91) .94 (.91) .68 (.60)
Honesty
In6 .87 (.77) .76 (.59) .84 (.73)
In9 .81 (.76) .65 (.57) .77 (.72)
In10 .73 (.77) .53 (.59) .71 (.72) .65
.90 (.88)
(.61)
Be1 .78 (.80) .61 (.63) .69 (.66)
Benevolence .70 (.67) Be2 .76 (.66) .57 (.51) .62 (.56) .59 (.55) .81 (.79) .81 (.81) .59 (.56)
Be3 .77 (.80) .59 (.61) .68 (.66)
Wi1 .88 (.89) .77 (.79) .83 (.85)
Willingness to Wi4 .85 (.90) .73 (.81) .81 (.85)
.74 (.70) .74 (.77) .92 (.93) .92 (.93) .74 (.77)
rely Wi5 .85 (.87) .72 (.76) .81 (.71)
Wi8 .87 (.85) .75 (.73) .82 (.70)
Wi2 .71 (.73) .50 (.54) .64 (.64)
Willingness to
.86 (.82) Wi6 .83 (.79) .69 (.63) .70 (.69) .62 (.61) .83 (.82) .84 (.82) .63 (.61)
depend
Wi7 .84 (.81) .70 (.66) .73 (.70)
Table 32: Psychometric Properties of the eWT-S (Modified Scale, Samples 4 and 5)

Notes: Results for cross-validation sample 5 in parentheses; γ = Completely standardized second-order loading; λ = Completely standardized first-order loading;
AVE = Average variance extracted.
Discriminant Validity

Initial evidence for the dimensionality of the first-order constructs was provided by the inter-
factor correlations which turned out to be significantly different from one (p < .05). As reported
in Table 33, the highest upper bound of the confidence interval for the correlation between
integrity/honesty and ability was .85 followed by .83, representing the upper bound of the
willingness to depend – ability confidence interval. Hence, none of the confidence intervals
around the correlations at the 95% level included “1.0”.

A more stringent cut-off value is provided by MacKenzie et al. (2005), who suggest testing
whether the correlation between two constructs is less than .71. Values less than the cut-off
value imply that the constructs have significantly less than half of their variance in common.
All but three construct pairings had correlations below .71. The three pairing were: ability –
integrity/honesty: .82; integrity/honesty – willingness to depend: .76; and ability – willingness
to depend: .80. When comparing the average variance extracted by each factor relative to that
factor’s shared variance with other factors in a model, this research was able to demonstrate
that the AVE was greater than the squared correlation in all but two circumstances. First, due
to the equally sized shared variance between integrity/honesty and ability and the minimum
AVE (AVEIntegrity = .65), one might question the discriminability of these two constructs. Here,
a similar pattern was also observed in samples 3a and 3b. Another questionable relationship
existed between the willingness to depend and the ability dimension, where the squared
correlation between these two constructs (.64) was slightly above the AVE of the former
construct (AVE = .63). For the remaining constructs, discriminant validity is supported due to
the AVE-criterion.

Table 33: Correlations among the Sub-dimensions (Modified Scale, Sample 4)

Integrity/ Willingness to Willingness to

Ability Benevolence
Honesty rely depend
Integrity/
0.65a
Honesty
0.82b
Ability 0.85c 0.70a
0.68d
0.52b 0.55b
Benevolence 0.58c 0.60c 0.59a
0.27d 0.30d
0.65b 0.69b 0.43b
Willingness to
0.70c 0.73c 0.50c 0.74a
rely
0.42d 0.47d 0.19d
0.76b 0.80b 0.50b 0.63b
Willingness to
0.79c 0.83c 0.56c 0.68c 0.63a
depend
0.57d 0.64d 0.25d 0.40d
Notes: a = Average variance extracted (AVE); b = Correlation coefficient (r); c = Upper bound of 95%
confidence interval of r; d = Squared correlation (r2).

249
In order to further assess the discriminant validity among the sub-dimensions, ten measurement
models with alternative factor combinations of varying model abstractness were compared by
means of chi-square difference tests and additional model fit indices.

First, as a basic reference model, the null model (Model N) was specified, for which it was
theorized that no systematic relationship among the observable indicators exists. In contrast,
eWOM trust was conceptualized as a uni-dimensional construct in Model 1. A series of second-
order constructs followed that met the concerns that various sub-dimensions showed
considerable interrelations. For instance, for ability, integrity/honesty, and willingness to
depend, correlations were relatively high (.82 between ability and integrity/honesty; .80
between ability and willingness to depend; and .76 between integrity/honesty and willingness
to depend). Hence, they were integrated in one general first-order factor which is hypothesized
to coexist with the Benevolence and the willingness to rely in a second-order factor model
(Model 3). Likewise, five four-factor models (Model 4a-4e) were estimated, where in each
model two first-order constructs were combined, while treating the remaining factors
separately. These conceptualizations were compared to the hypothesized second-order
construct of eWOM trust, consisting of five separate first-order factors (Model 5).

Table 34 demonstrates that the proposed eWOM trust model performed best in terms of the chi-
square statistic and also achieved the most satisfactory model fit. Additionally, all chi-square
difference tests comparing the five-factor model to alternative models were significant,
indicating a better fit.

Similarly to the earlier research process, the purified scale was also subject to a pairwise
comparison of its sub-dimensions by means of CFA. For every possible pairing, two alternative
models were evaluated. More specifically, the one-factor model integrating both factors was
compared with the hypothesized two-factor model. If the chi-square fit of the two-factor model
is better than the fit of the one-factor model, evidence for discriminant validity among those
factors exists (Anderson & Gerbing, 1988). Table 35 provides the results of this analysis. It can
be shown that, across all pairings, the two-factor models more adequately represented the data.
This is mirrored by significant improvements in chi-square from the single factor solution to
the two-factor solution but also by better fit indices. For instance, the fit of the one-factor model
for ability-integrity/honesty, χ2 = 616.03 (df = 54, p < .001) was worse than the fit of the two-
factor model, χ2 = 158.64 (df = 53, p < .001). This results in a chi-square difference of 457.39
(df = 1), which is significant on the .001 level. The same was also true for the factor model of
benevolence and willingness to depend as for the rest of the pairings. These findings imply that
the five first-order constructs, while commonly measuring trust in online customer reviews,
discriminate from each other.

250
Implications for Construct’s Structure and Content

In reference to the first research question and due to the results so far, one can conclude that
eWOM trust is best described as a second-order construct with five separate first-order sub-
dimensions: (1) ability (usefulness), (2) integrity/honesty, (3) benevolence, (4) willingness to
rely (use), and (5) willingess to depend (risk). These factors capture cognitive, emotional, as
well as conative trust aspects. In line with empirical results, this thesis formally defines eWOM
trust as a consumer’s general confidence that the information conveyed in online customer
reviews is reliable. More specifically, eWOM trust is formally defined as a five-dimensional,
second-order construct capturing the extent to which a consumer has the belief that information
given in online customer reviews is (a) useful, (b) honest, and (c) benevolent; (d) is willing to
rely on (use) online customer reviews in general; and (e) is willing to depend on this kind of
market information. According to exploratory as well as confirmatory factor analyses, the
eWOM trust scale (eWT-S) with its 22 observable indicators adequately measures a consumer’s
general tendency to trust claims of fellow shoppers made in online reviews. Each of the five
sub-dimensions is measured with an appropriate number of items fulfilling the statistical and
theoretical requirements of measurement theory (e.g., Albers & Hildebrandt, 2006).

251
 Chi Square Difference Model Fit Indices

252
Model Chi Square df p Chi Chi
Competing df
Square Sign. Square / RMSEA NFI NNFI CFI RMR SRMR GFI AGFI
Models Difference
Difference df
Null Model (N) 9,218.44 231 39.91 n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a.
One Factor Model (1) 2,461.36 209 .001 11.78 .19 .73 .72 .75 .16 .10 .58 .50

(N-1) 6,757.08 22 ***

Three Factor Model (3) 1,246.37 206 .001 6.05 .12 .87 .87 .88 .11 .07 .76 .71
(A/I/W2; B; W1)
(N-3) 7,972.08 25 ***
(1-3) 1,215.00 3 ***
Four Factor Model (4a) 1,105.33 205 .001 5.39 .12 .88 .89 .90 .11 .07 .77 .72
(A/I; B; W1; W2)
(N-4a) 8,113.11 26 ***
(1-4a) 1,356.03 4 ***
(3-4a) 141.04 1 ***
Table 34: Competing Models (Modified Scale, Sample 4)

Four Factor Model (4b) 822.28 205 .001 4.01 .08 .91 .92 .93 .12 .07 .86 .83
(A/W2; I; B; W1)
(N-4b) 8,396.17 26 ***
(1-4b) 1,639.09 4 ***
(3-4b) 424.09 1 ***
Four Factor Model (4c) 843.11 205 .001 4.11 .08 .91 .92 .93 .12 .07 .86 .83
(I/W2; A; B; W1)
(N-4c) 8,375.33 26 ***
(1-4c) 1,618.25 4 ***
(3-4c) 403.26 1 ***
(Continued on next page.)
 Chi Square Difference Model Fit Indices

Model Chi Square df p Chi Chi

Competing df
Square Sign. Square / RMSEA NFI NNFI CFI RMR SRMR GFI AGFI
Models Difference
Difference df
Four Factor Model (4d) 1,728.46 205 .001 8.43 .15 .81 .81 .83 .14 .09 .68 .61
(I/W1; I; B; W2)
(N-4d) 7,489.98 26 ***
(1-4d) 732.90 4 ***
(3-4d) -482.10 1 n.s.
Four Factor Model (4e) 1,034.97 205 .001 5.05 .10 .89 .90 .91 .16 .09 .81 .77
(I; A;B; W1/W2)
(N-4e) 8,183.48 26 ***
(1-4e) 1,426.40 4 ***
(3-4e) 211.40 1 ***
Five Factor Model (5) 627.59 204 .001 3.08 .06 .93 .95 .95 .10 .06 .90 .88

(N-5) 8,590.85 27 ***

(1-5) 1,833.77 5 ***
(3-5) 618.77 2 ***
(4a-5) 477.74 1 ***
Table 34 (Cont.): Competing Models (Modified Scale, Sample 4)

(4b-5) 194.68 1 ***

(4c-5) 215.52 1 ***
(4d-5) 1,100.87 1 ***
(4e-5) 407.37 1 ***
Notes: A = Ability; I = Integrity/Honesty; B = Benevolence; WR = Willingness to rely; WD = Willingness to depend; *** = p <.001; n.s. = not significant; n.a. = not available.

253
254
Chi Square Difference Model Fit Indices
Sub-dimension Chi Chi Chi
Model df p Competing df
Pairing Square Square Sign. Square / RMSEA NNFI CFI RMR SRMR AGFI
Models Difference
Difference df
Integrity / Null Model (N) 5,199.06 66 78.77 n.a. n.a. n.a. n.a. n.a. n.a.
Ability
One Factor Model (1) 616.03 54 .001 (N-1) 4,583.03 12 *** 11.41 .18 .87 .89 .08 .06 .67
Two Factor Model (2) 158.64 53 .001 (N-2) 5,040.42 13 *** 2.99 .06 .97 .98 .04 .03 .93
(1-2) 457.39 1 ***
Integrity / Null Model (N) 3,606.23 45 80.14 n.a. n.a. n.a. n.a. n.a. n.a.
Benevolence
One Factor Model (1) 369.50 35 .001 (N-1) 3,236.73 10 *** 10.56 .14 .88 .91 .14 .08 .79
Two Factor Model (2) 94.51 34 .001 (N-2) 3,511.72 11 *** 2.78 .06 .98 .98 .05 .03 .94
(1-2) 275.00 1 ***
Integrity / Null Model (N) 4,552.47 55 82.77 n.a. n.a. n.a. n.a. n.a. n.a.
Willingness to rely
One Factor Model (1) 1,189.06 44 .001 (N-1) 3,363.41 11 *** 27.02 .26 .68 .75 .02 .14 .47
Two Factor Model (2) 91.35 43 .001 (N-2) 4,461.12 12 *** 2.12 .05 .99 .99 .03 .02 .95
(1-2) 1,097.71 1 ***
Integrity / Null Model (N) 3,835.03 45 85.22 n.a. n.a. n.a. n.a. n.a. n.a.
Table 35: Pairings of Sub-dimensions (Modified Scale, Sample 4)

Willingness to depend
One Factor Model (1) 292.07 35 .001 (N-1) 3,542.96 10 *** 8.34 .13 .91 .93 .11 .05 .83
Two Factor Model (2) 87.60 34 .001 (N-2) 3,747.43 11 *** 2.58 .06 .98 .99 .04 .02 .95
(1-2) 204.47 1 ***
Ability / Null Model (N) 2,541.11 28 90.75 n.a. n.a. n.a. n.a. n.a. n.a.
Benevolence
One Factor Model (1) 460.46 20 .001 (N-1) 2,080.65 8 *** 23.02 .21 .76 .83 .23 .12 .66
Two Factor Model (2) 97.97 19 .001 (N-2) 2,443.15 9 *** 5.16 .09 .95 .97 .08 .06 .91
(1-2) 362.49 1 ***
Ability / Null Model (N) 3,794.88 36 105.41 n.a. n.a. n.a. n.a. n.a. n.a.
Willingness to rely
One Factor Model (1) 622.38 27 .001 (N-1) 3,172.50 9 *** 23.05 .26 .79 .84 .11 .08 .52
Two Factor Model (2) 101.37 26 .001 (N-2) 3,693.52 10 *** 3.90 .08 .97 .98 .03 .02 .93
(1-2) 521.01 1 ***
(Continued on next page.)
 Chi Square Difference Model Fit Indices
Sub-dimension Chi Chi Chi
Model df p Competing df
Pairing Square Square Sign. Square / RMSEA NFI NNFI CFI RMR SRMR GFI AGFI
Models Difference
Difference df
Ability / Null Model (N) 2,798.42 28 99.94 n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a.
Willingness to depend
One Factor Model (1) 274.29 20 .001 (N-1) 2,524.13 8 *** 13.71 .17 .90 .87 .91 .14 .07 .87 .77
Two Factor Model (2) 53.89 19 .001 (N-2) 2,744.53 9 *** 2.84 .06 .98 .98 .99 .03 .02 .97 .95
(1-2) 220.40 1 ***
Benevolence / Null Model (N) 2,142.71 21 102.03 n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a.
Willingness to rely
One Factor Model (1) 549.49 14 .001 (N-1) 1,593.22 7 *** 39.25 .27 .74 .62 .75 .34 .18 .76 .53
Two Factor Model (2) 62.08 13 .001 (N-2) 2,080.63 8 *** 4.78 .09 .97 .96 .98 .08 .05 .97 .93
(1-2) 487.41 1 ***
Benevolence / Null Model (N) 1,295.71 15 86.38 n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a.
Willingness to depend
One Factor Model (1) 350.70 9 .001 (N-1) 945.00 6 *** 38.97 .30 .73 .56 .73 .24 .13 .79 .50
Two Factor Model (2) 27.07 8 .001 (N-2) 1,268.64 7 *** 3.38 .07 .98 .97 .99 .07 .03 .98 .95
(1-2) 323.64 1 ***
Willingness to rely / Null Model (N) 2,415.06 21 115.00 n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a.
Willingness to depend
One Factor Model (1) 347.93 14 .001 (N-1) 2,067.13 7 *** 24.85 .23 .86 .79 .86 .21 .10 .82 .65
Table 35 (Cont.): Pairings of Sub-dimensions (Modified Scale, Sample 4)

Two Factor Model (2) 37.52 13 .001 (N-2) 2,377.54 8 *** 2.89 .06 .98 .98 .99 .03 .02 .98 .96
(1-2) 310.41 1 ***
Notes: *** = p <.001; n.a. = not available.

255
Supplemental Validity Assessments
Concurrent Validity

For assessing concurrent validity, the following research question was asked:

RQ 2: Does the developed measure of eWOM show the ability to distinguish groups of
consumers that have indicated varying levels of eWOM trust on a single-item scale?

In the course of the online questionnaire of the main survey, respondents from sample 4 were
asked to indicate their overall level of trust in reviews on a single-item scale. This earlier-
validated scale ranged from 0 (“I don’t trust online customer reviews”) to 6 (“I totally trust
online customer reviews”); thus, higher values indicated heightened trust in eWOM. The mean
score of the 526 respondents on this scale was 3.80 (SD = .05) and the median was 4.
Consequently, the sample was separated into two groups: the first one represented low-trusters
(n = 177) who have indicated values ranging from 0 to 3; the second group (high-trusters, n =
131) stated 5 or 6 on the scale. Respondents having shallow trust were excluded in order to
classify respondents clearly as high or low trusters. For both groups, the mean value for trust in
customer reviews was calculated by using the new eWOM trust scale. An independent-samples
t-test indicated that trust scores were significantly higher for high-trusters (M = 4.58, SD = .58)
than for low-trusters (M = 2.94, SD = .77), t(306) = 21.39, p < .001. These results suggested
the new scale’s ability to distinguish between groups with varying trust levels, supporting its
concurrent validity.

Convergent Validity and Discriminant Validity

As a scale’s ability to correspond with other, “maximally dissimilar” (Netemeyer et al., 2003)
measures of the same construct is an important scale property, the following research question
was asked:

RQ 3: Does the developed measure of eWOM trust correlate significantly and

considerably with other methods to measure eWOM trust?

In order to prove convergent validity, the eWOM trust scale was correlated with four different
measures of trust in online customer reviews: (1) a non-diagnostic single item measure
(previously discussed), (2) a Likert-format, multi-item measure of overall eWOM trust, (3) a
semantic differential, multi-item measure of overall eWOM trust, and (4) a qualitative measure
of overall eWOM trust. Sample 4 (n = 526) provided the necessary data for the first three
approaches. Table 36 reports the results of the correlations between the new eWOM trust scale
(measured by a composite value obtained by averaging the respondents’ scores across the five
sub-dimensions, which were themselves calculated by averaging the dedicated observable
items) and the alternative measures. The correlations were all strong (ranging from .78 to .89)

256
and significant on the .001 level. Additionally, the strong correlations among the alternative
measures (apart from the new measure) suggested that all approaches assess the same construct.

For the evaluation of the association between the new scale and a qualitative measure of eWOM
trust, a separate survey was conducted. By using a paper-and-pencil questionnaire, 117
respondents were asked to answer the following open-ended question: “Why do you
trust/distrust information given in online customer reviews?”. 54% of the business and non-
business students were females and the average age was 24.1 (ranging from 19 to 30 years). In
the course of the survey, participants had also to indicate their personal level of eWOM trust
by answering the 22-item eWT-S. Two independent judges, unfamiliar with the research
purpose, content-coded the responses of the open-ended question on a 7-Point scale from -3 (no
trust), to 0 (neutral), to +3 (high trust). Subsequently, for each respondent the scores of the
raters were averaged. The correlation between the eWOM Trust scale and this average value
was r = .90, p < .001. Taken together, the correspondence between the eWOM trust scale and
the responses to the five alternative measures strongly supported the scale’s convergent validity.

Table 36: Convergent Validity

Alternative Measures of eWOM Trust

Developed
measure of Non- Likert Semantic
Sample Mean SD diagnostic multi- differential Qualitative
eWOM
Trust single item multi-item measure
measure measure measure
Developed
measure of eWOM 3.73 .89 (.95)1
Trust
Non-diagnostic
3.80 1.03 .78*** (n.a.)
single measure 4
Likert multi-item (n=526)
3.42 1.09 .89*** .78*** (.92)1
measure
Semantic
differential multi- 3.91 .92 .86*** .76*** .78*** (.96)1
item measure
Qualitative 5
3.06 1.60 .90*** n.a. n.a. n.a. n.a.
measure (n=117)
1
Notes: = Internal consistency estimate (Cronbach’s alpha); *** p < .001; n.a. = not available.

Cook and Campbell (Cook & Campbell, 1979) claim that for assuring construct validity, two
different assessments have to be made by the researcher: first, testing the new scale’s
convergence with alternative measures of the same construct, and second, ascertain that it is
discriminative from other measures that are supposed to assess related but conceptually
different concepts. While research question three targeted the first assessment, the following
addresses the psychometric properties of the eWOM scales that assure its discriminant validity
from two well-established constructs:
257
 RQ 4: Is the developed measure of eWOM trust significantly different from the
measurements of (a) review credibility and (b) attitude towards reviews in general?

The following discussion is conducted on two different levels of complexity: first, this thesis
investigates the divergence on the overall construct level of the three concepts; after that, a more
restrictive investigation targets the discriminant validity of the five sub-dimensions of eWOM
trust in relation to review credibility (RCred), as well as attitude (RAtt).

One of the most frequently used methods to test discriminant but also convergent validity on
the construct level is the multitrait multimethod matrix (MTMM matrix) introduced by
Campbell and Fiske (1959). At its head, this procedure demands that in addition to the trait in
focus, at least a second trait has to be measured with two alternative methods. In this study,
three conceptually different traits were included: (i) trust in online reviews (eWOM trust), (ii)
eWOM credibility (RCred), and (iii) eWOM attitude in general (RAtt). These three traits were
assessed with two alternative measurement methods, namely a multi-item Likert scale and a
multi-item semantic differential scale. The latter was used because it represents a measurement
method frequently used in social research that is maximally different from Likert scales at the
same time. The author is aware that the further inclusion of a third trait and a third measurement
method would have enabled this study to further investigate possible systematic influences
(e.g., method effects) by means of CFA (Hildebrandt & Temme, 2006). However, the author at
the same time refraind from doing so as the burden on the survey respondent was regarded as
too high.

Consumer trust in online reviews and recommendations was assessed by the application of the
new Likert-formatted eWOM trust scale, as well as an adopted version of the scale which
enabled the measurement of the same construct with a semantic differential. In both cases, the
respondents’ answers to the 22 scale items were averaged in order to derive the Likert (TL) and
the semantic differential (TS) score. The Likert score for eWOM credibility (C L) was obtained
by averaging the five scale items adapted from Boush, Friestad and Rose (1994) and the
construct’s semantic differential score (CS) by averaging the responses to the scale proposed by
Beltramini and Evans (1985). In a similar manner, the measurement scores for consumers’
eWOM attitude was produced. Here, seven items taken from Pollay and Mittal (1993) formed
the Likert score (AL), while the average over participants’ responses to the scale introduced by
Olney, Holbrook and Batra (1991) formed the semantic differential score (AS). Before
averaging, the scales’ uni-dimensionality as well as reliability was verified. The MTMM matrix
(see Table 37) represents the (unstandardized) correlation coefficients between the scores of all
measures.

258
Table 37: Multi-Trait Multi-Method Matrix

Method I - Likert Scale Method II - Semantic Differential

eWOM eWOM eWOM eWOM eWOM eWOM

Trust Credibil Attitude Trust Credibility Attitude
(TL) ity (CL) (AL) (TS) (CS) (AS)

eWOM Trust (TL) (.95)1

Method I -
Likert eWOM Credibility (CL) .722 (.81)1
Scale
eWOM Attitude (AL) .792 .592 (.86)1

Method II eWOM Trust (TS) .863 .604 .764 (.96)1

-
Semantic eWOM Credibility (CS) .734 .753 .704 .812 (.87)1
Differenti
al eWOM Attitude (AS) .564 .264 .763 .682 .702 (.88)1
1 2
Notes: = Internal consistency reliability estimates (Cronbach’s alpha); = Heterotrait-monomethod
correlations; 3 = Monotrait-heteromethod correlations; 4 = Heterotrait-heteromethod correlations.

As mentioned, the MTMM matrix is also capable of providing additional insights concerning a
scale’s convergent validity. Campbell and Fiske (1959) state that evidence for convergent
validity is provided where the coefficients of the reliability diagonal are consistently the highest
in the matrix. Table 37 demonstrates that these coefficients meet this standard and surpassed a
desirable threshold. The two scales applied to measure eWOM trust turned out to perform
especially well, as the Likert format scale (TL) and the semantic differential scale (TS) exhibited
the highest reliabilities: α = .95 and .96 respectively. Internal consistency of the remaining four
measurement approaches ranged from .81 to .88 and, hence, also achieved respectable
Cronbach’s alphas. Campbell and Friske (1959) also demand that the monotrait-heteromethod
coefficients have to be statistically significant and sufficiently large in order to support scale
convergence. Similarly, Bagozzi and Yi (1988) argue that only where this condition is fulfilled
one can plausibly argue that all alternative instruments are likely to measure the same
underlying construct. The research at hand was able to show that also this condition was
fulfilled, as all monotrait-heteromethod correlations were significant (p < .001) and large.
Specifically, the correlation between TL and TS was .86, between CL and CS was .75, and
between AL and AS was .76. While some other correlations also turned out to be large relative
to other correlations in the matrix (e.g., the correlation between TS and AL was .76), this thesis’
author is convinced that – together with the earlier-discussed findings – strong evidence for the
new scale’s convergent validity exists.

Concerning discriminant validity, Campbell and Fiske (1959) propose that an MTMM matrix
should be reviewed in respect to three criteria. First, the entries in the validity diagonal should
be higher than the entries in the heteromethod block that share the same row and column.
Empirical data shows that the monotrait-heteromethod coefficient of eWOM trust (TL – TS) is

259
bigger than the corresponding heterotrait-heteromethod coefficients. That is, r = .60 (for TS -
CL), r = .76 (for TS - AL), r = .76 (CL - TL), and r = .56 (AS - TL). The same held true for the
other two constructs. One exception was the monotrait-heteromethod coefficient of eWOM
attitude (r = .76), which was the same as the correlation between TS and AL.

The second criterion demands that the correlations between different measures of a trait should
be higher than the correlations among traits which have methods in common (Campbell &
Fiske, 1959). The MTMM matrix only partially passed this second discriminant validity
standard. While the monotrait-heteromethod coefficient of eWOM trust (r = .86) was higher
than all heterotrait-monomethod coefficients associated with this trait (.72 for CL and TL; .79
for AL and TL; .81 for CS and TS; and .68 for AS and Ts), the coefficient was sometimes lower
for the other two constructs. Here, the monotrait-heteromethod coefficient of eWOM
Credibility (r = .75) was less than the correlation between CS and TS (r = .81). Similarly, the
correlation coefficient for eWOM Attitude (r = .76) was smaller than the .79 which represented
the correlation between AL and TL.

According to Campbell and Fiske’s (1959) third criterion, the pattern of the correlations should
be the same in all the monomethod and heteromethod triangles. This was hardly achieved,
suggesting that some methods influence existed. While it was refrained in doing so in this
research, Hildebrandt and Temme (2006) propose several models that strive to uncover various
methods effects by means of confirmatory analysis. Some of these may also be applicable to
this research.

This thesis also considered additional assessments to evaluate the scale’s discriminant validity
on the construct level (by using a single-method approach). While the MTMM matrix showed
the correlations between the new Likert-formatted eWOM trust scale (TL) and the two related
constructs to range from .56 to .79 – depending on the measurement instrument – and are
therefore desirably low, it is advisable to provide more restrictive evidence for the scale’s
discriminant validity. To ensure that the constructs are less than perfectly related, Anderson and
Gerbing (1988) recommend that the confidence interval (±2 standard errors) around the
constructs’ correlations does not contain a value of 1. The upper bound of the 95% confidence
interval of the correlation between eWOM trust (TL) and eWOM credibility was .76 when the
construct was measured with a Likert scale and .77 when it measured with a semantic
differential. For the correlation of eWOM trust with eWOM attitude, it was .82 (TL – AL) and
.62 (TL – AS). MacKenzie et al. (2011), on the other hand, propose that the test whether the
constructs’ intercorrelation is less than .71 represents a more stringent method to assess
discriminant validity. Here, the (unstandardized) correlations were regularly slightly about this
threshold. That is, the correlation between eWOM trust and eWOM credibility was .72 [.73]
and with eWOM attitude .79 [.56]. Taken together, the above findings demonstrate that the
three constructs are separate and the new scale possesses discriminant validity on the construct

260
level. However, the disadvantage of using composite scores for the involved constructs (as was
done here, until now) is that this ignores measurement error. Scholars agree that this approach
is obviously problematic, since the resulting attenuation makes it more likely that the constructs
are distinct. Hence, further analyses were necessary to identify potential threats of discriminant
validity on the sub-dimensional level.

To assess discriminant validity on the sub-construct level, three different tests using CFA were
performed. First, for each possible pairing of constructs, a one-factor model was compared with
a hypothesized two-factor model which separates the individual eWOM sub-dimension from
eWOM credibility and eWOM attitude respectively. For comparison a chi-square difference
test between measures allowing phi (Φ) to vary (i.e., two factor model) and then constraining
phi correlation to unity (i.e., one factor model) (Anderson & Gerbing, 1988). This criterion was
met by all competing models. The chi-square fit of the unconstrained models was always
significantly lower (p < .001) than the fit of the one factor models, providing first evidence for
discriminant validity on the sub-dimensional level.

Next, it was tested whether the cross-construct correlations for each pairing is significantly less
than one (Bagozzi & Hearherton, 1994). This was done by examining the confidence interval
(CI) of the (completely standardized) correlation estimates. Support for discriminant validity
was provided, as none of the confidence intervals for the pairwise correlation estimates (±2
standard errors) included the value of one. The highest upper bond was found for the correlation
between integrity/honesty and eWOM credibility (.90), followed by the correlation between
ability and eWOM attitude (.89). Both relationships can be explained due to conceptual overlap,
explained earlier, and were therefore expected. Finally, Fornell and Larcker (1981) suggest that
if the squared phi correlation between two constructs is less than the average variance extracted
(AVE) of each involved construct, discriminant validity is supported. Accordingly, the AVEs
and the squared correlations between the individual eWOM dimensions and the two related
constructs were compared. Here, only three of ten possible pairings passed this standard.
Benevolence exclusively turned out to possess discriminative validity towards both related
constructs. Here, the highest phi square correlation (.32) was smaller than the AVEs (ranging
from .50 to .59.). The rest of the sub-dimensions seemed to be more intertwined. Apart from
the results for the most restrictive criterion, the tests provided sufficient evidence to assume
that all five eWOM trust sub-dimensions discriminate from both eWOM credibility (RCred) and
eWOM attitude (RAtt).

It may be reasonable to question whether generalized trust in online customer reviews differs
from dispositional trust. Hence, the following research question was proposed:

RQ 4: (c) Is the developed measure of eWOM trust significantly different from the
measurement of dispositional trust? (d) Do consumers develop trust that is specific
to online customer reviews?
261
Empirical evidence for the important distinction between a consumer’s disposition to trust
(TDispo) and his/her general tendency to trust eWOM was provided by data sourced in this
research’s main study (sample 4). In total, 526 respondents – from which 48% were females
(average age: 40 years; age ranging from 16 to 74 years) – had to answer an online questionnaire
which included, besides the new eWOM trust scale and additional items measuring other
constructs, also nine items adopted from McKnight et al. (2004) and Gefen (2000), intended to
measure dispositional trust. These items were measured on a 7-Point Likert scale with the
maxima 0 (I strongly disagree) and 6 (I strongly agree). Higher values on this scale indicated
heightened disposition to trust. The Cronbach alpha of this construct was .90. Whether the two
constructs were distinguishable or not was assessed by using a similar approach to that
described above.

The first hint suggesting that the two constructs are distinct was obtained by an investigation of
the means (composite values), which were 3.73 (SD = .89) for eWOM trust and 3.35 (SD =
1.04) for dispositional trust. These were statistically significantly different, t(525) = 7.85, p <
.001, with less trust in generalized others and more trust in information given in online customer
reviews. The mean values suggest that this sample discriminated in their targets for trust.
Further evidence was obtained by the construction of a confidence interval (±2 standard errors)
around the correlation coefficient (r = .36). (The significant correlation between the two
constructs (p < .001) finds support in this thesis’ nomological framework, which proposes
disposition to trust as a determinant of eWOM trust. Hence, H1 is supported.) The upper bound
of the 95 percent confidence interval of this correlation was .43 and, hence, did not include the
value of 1. Although a correlation less than 1 is a necessary condition for discriminant validity
(Anderson & Gerbing, 1988), the test whether the construct intercorrelation is less than .71 is
regarded as a more stringent method to ensure constructs’ distinction (MacKenzie et al., 2011).
While the correlation between the constructs was significant, it was, however, far below this
recommended threshold. Additional evidence for discriminant validity was gathered by means
of confirmatory factor analysis (CFA). Here, a one-factor model, where all items were assumed
to load on a single factor, was compared with the hypothesized two-factor model, which
separates eWOM trust from dispositional trust. Discriminant validity exists if the chi-square
value of the two-factor model is significantly lower than the chi-square value of the one-factor
model (Anderson & Gerbing, 1988). Table 38 presents the indicators of model performance for
the alternative models. In general, the two-factor model showed satisfactory fit (χ2 = 1,075.77
(df = 426, p < .001), absolute fit indices: GFI = .88, AGFI = .85, RMSEA = .06, RMR = .10,
SRMR = .06; incremental fit indices: CFI = .94, NNFI = .94, NFI = .91; parsimonious fit
indices: normed chi-square: 2.53). Additionally, the fit of the two-factor model was
significantly better than the fit of the one-factor model (χ2Diff = 3,819.16; df Diff = 8; p < .001).

262
 Chi Square Difference Model Fit Indices
Chi Chi
Model df p Competing df Chi Square /
Square Square Sign. RMSEA NFI NNFI CFI RMR SRMR GFI AGFI
Models Difference df
Difference
Null Model (N) 12,050.83 465 25.92 n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a.
One Factor Model (1) 4,894.93 434 .001 11.28 .19 .59 .59 .62 .25 .13 .49 .41

(N-1) 7,155.90 31 ***

Two Factor Model (2) 1,075.77 426 .001 2.53 .06 .91 .94 .94 .10 .06 .88 .85

(N-2) 10,975.06 39 ***

(1-2) 3,819.16 8 ***
Notes: *** = p <.001; n.a. = not availabe.
Table 38: Model Comparison (eWOM Trust – Disposition to Trust)

263
Table 39: Results of the EFA (Sample 4)

Factor Loading Communality MSA

Item
Factor 1 Factor 2 Factor 3 Factor 4 Factor 5
Dp1 . . . .82 . .66 .91
Dp2 . . . .79 . .69 .91
Dp3 . . . .61 . .63 .93
Dp4 . . . .56 . .59 .93
Dp5 . . .86 . . .73 .90
Dp6 . . .85 . . .72 .92
Dp7 . . .88 . . .73 .89
Dp8 . . .73 . . .73 .93
Dp9 . . .51 . . .51 .95
Ab7 .54 . . . . .69 .96
Ab8 .51 . . . . .68 .97
Ab9 .49 . . . . .68 .96
Ab10 .64 . . . . .72 .97
Ab11 .46 . . . . .64 .97
In2 .84 . . . . .71 .98
In3 .90 . . . . .71 .97
In4 .91 . . . . .75 .97
In5 .88 . . . . .75 .97
In6 .91 . . . . .77 .97
In9 .76 . . . . .67 .98
In10 .96 . . . . .62 .97
Be1 . . . . .91 .77 .89
Be2 . . . . .76 .67 .94
Be3 . . . . .87 .73 .90
Wi1 . .97 . . . .78 .93
Wi4 . .94 . . . .75 .95
Wi5 . .98 . . . .76 .95
Wi8 . .94 . . . .78 .95
Wi2 . .44 . . . .51 .95
Wi6 . .52 . . . .60 .97
Wi7 . .45 . . . .58 .95
Eigenvalue 12.44 4.07 2.50 1.15 1.10
% of 40.14% 13.13% 8.08% 3.72% 3.55%
Variance
Notes: Total variance explained: 68.61%; Factor loadings below .30 not shown. Extraction method: Principal
Component Analysis; Rotation method: Promax with Kaiser Normalization; Rotation converged in 7 iterations;
Kaiser-Meyer-Olkin Measure of Sampling Adequacy (MSA) .95 and Bartlett’s test of Sphericity: sig. .001.

By applying exploratory factor analysis (EFA), one is able to get additional insight into the
relationship between the two constructs. This approach can be regarded as an even more
demanding approach, because no a-priori assignment of the items to their hypothesized latent
constructs takes place, but the individual items were grouped according to their relationship
inherent in the data. Hence, principal components analysis (PCA) with Promax rotation was
applied to the data. An oblique rotation was used due to theorized linear relationships between
eWOM trust and disposition to trust, which is assumed to be one of its antecedents (Mayer et

264
al., 1995; McKnight & Chervany, 2002, 2006; Rotter, 1971). The PCA resulted in a five factors
with eigenvalues greater than 1 explaining 68.61% of total variance (see Table 39). Factors 1,
2, and 5 represented the cognitive, behavioral and emotional aspects of eWOM trust and
included solely items theorized to belong to the eWOM construct. That is, ability and
integrity/honesty items all loaded substantially and significantly on the first factor (loadings
ranging from .46 to .96). All willingness to rely as well as willingness to depend items showed
a significant relationship, with the second factor representing the behavioral aspects of eWOM
trust. Wi2 was the item with the weakest – but nevertheless a significant – loading (.44). The
emotional/social aspect of trust is mirrored in the fifth factor, including the three benevolence
items with strong loadings (≥ .76). Disposition to trust is represented by the third and the fourth
factor. According to literature, dispositional trust consists of two separate dimensions: (1) Faith
in humanity and (2) Trusting stance (Gefen, 2000; McKnight et al., 2002b; McKnight et al.,
2004). The PCA showed that the items intended to measure faith in humanity (Dp1-Dp4) all
loaded on a single factor (Factor 4), while the remaining items hypothesized to measure trusting
stance (Dp5-Dp9) loaded significantly on the third factor (item loadings ranging from .51 to
.88). None of the items showed notable cross-loadings. Taken together, the above results
provided strong evidence in favour of a two-construct-perspective.

In order to address the second part of the research question (RQ4d), this research compared the
new scale assessing eWOM trust (eWOMTrust) with scales targeting the measurement of trust in
other objects. For this purpose, the new scale was modified by substituting for the word online
customer reviews two other targets – friends and family members (i.e., word-of-mouth/WOM)
(WOMTrust) as well as salespersons (SPTrust). Besides these two alternative forms of market
communication, trust in online advertising (OADTrust) was measured by an adopted 19-item
version of the scale advanced by Soh (2007). All items were measured on a 7-Point Likert scale,
ranging from 0 (“I strongly disagree”) to 6 (“I strongly agree”). Consequently, three different
questionnaires containing these scales were set to a convenience sample. The survey was
conducted online in spring 2013. A total of 133 usable responses were collected. 73.7% of the
respondents were females and the average age of the sample was 22.6 years (ranging from 18
to 29 years). In total, 45 usable responses were collected for the eWOMTrust – WOMTrust
questionnaire, 42 for eWOMTrust – SPTrust, and 46 responses for the eWOMTrust – OADTrust
instrument. For analysis, a composite measure for each construct was calculated. All scales
achieved desired levels of internal consistency as Cronbach alphas’ surpassed the recommended
.70 threshold: .95 for trust in WOM, .98 for trust in salespersons, and .96 for trust in online
advertising. The means of the various scales (with higher scores indicating higher trust), were
3.47 (SD = .78) for eWOMTrust, 4.63 (SD = .71) for WOMTrust, 3.76 (SD = 1.16) for SPTrust, and
1.89 (SD = 1.11) for OADTrust. The differences in means between eWOM trust and the other
measures were all significant, with consumers having the highest trust in offline word-of-mouth
and the lowest trust in online advertising. These differences suggest that the sample
discriminated in their willingness to trust different objects. Additional evidence of
265
discrimination was given by a review of the correlations between the eWOM trust scale and the
other measures. The only correlation that was significantly different from zero was the moderate
correlation between eWOM trust and WOM trust, r = .33, p < .05. This may be due to
consumers’ perceptions that these two kinds of information have a similar basic nature.
However, in sum, results suggest that consumers who are willing to trust online reviews are not
necessarily more confident in other sources of market communication. Therefore, from this
perspective, the eWOM trust scale does not reflect a general tendency to trust per se, but a
consumer’s tendency to trust a specific kind of market information; that is, online customer
reviews and recommendations.

Known-group Validity

In scale research, it is a common practice to regard the theoretical identification of groups that
are hypothesized to score low/high on the focal construct and subsequently compare the scores
of these groups on the measure as additional evidence for construct validity (e.g., Heimberg &
Holaway, 2007; Netemeyer et al., 1995; Tian et al., 2001). In this research, it is expected that
two convenient populations, namely university students and faculty members, are likely to
differ on their tendency to trust consumer reviews. More precisely, it is assumed that faculty
members are more likely to show lower levels of eWOM trust due to several reasons which
correspond to the proposed nomological network. For instance, faculty members are older, are
more skeptical, have higher self-esteem and do have more marketplace experiences. Hence,
members of this group are more likely to be aware of potential manipulations by online reviews
and perceive heightened skepticism for marketing communication in general. Also they have
been subject to a different socialization process, providing them with cultural folk wisdom.
Further, faculty typically exhibits different Internet and computer usage patterns as, for
instance, its members less regularly participate in social networks, like Facebook, where they
could socialize with others. Accordingly, the following research question was proposed:

RQ 5: Do students and faculty members differ significantly in their trust in customer

reviews?

Significant difference between these two groups would provide evidence for known-group
validity. However, also differences within these two groups were checked. Therefore faculty
members were categorized by disciplines (natural sciences, business sciences, humanities, law,
social sciences, others) and academic qualifications (PhD, professorship). On the other hand,
the (business) students were grouped into undergraduates and graduates.

Data for the student sample were collected from two Austrian universities during the reliability
stage (samples 3a, 3b). In total, 850 usable questionnaires were obtained (574 undergraduates;
276 graduates). For the faculty sample, 1,182 invitations were sent to qualified respondents in
academic institutions in Austria, Germany, and Switzerland. The emails sent in several waves

266
included a link to an online questionnaire which was accessible for three weeks in August 2013.
The final convenience sample comprised 205 academics, resulting in a response rate of 17.34%.
Of these, 112 persons (54.63%) held at least a PhD title and 93 a professorship (45.37%). The
sample included 47 persons from natural sciences, 62 from business sciences, 29 from
humanities, 20 from law, 32 from social sciences, and 15 from other disciplines.

As always, all respondents were requested to indicate their personal level of generalized eWOM
trust with the 22 items on a 7-Point Likert scale ranging from 0 to 6. Here, higher numbers
indicated heightened trust. Overall, the mean trust in online reviews for students was 3.82 (SD
= 1.84), with undergraduate students averaging at 3.82 (SD = 1.86) and graduates at 3.84 (SD
= 1.78). The mean eWOM trust for faculty members was 3.12 (SD = 1.11), with a break-down
of 2.85 (SD = 1.18) for natural sciences, 3.31 (SD = 1.04) for business sciences, 2.87 (SD =
1.24) for humanities, 3.26 (SD = .96) for law, 3.03 (SD = 1.15) for social sciences, and 3.66
(SD = .80) for faculty belonging other disciplines. The difference between students and faculty
was statistically significant, t(1,053) = 5.24, p < .001, with students having more trust in online
reviews than members of the faculty, as expected. Thus, known-group validity is supported.

The comparison eWOM trust for undergraduate and graduate students revealed no significant
differences between these groups (t(848) = .14, n. s.) and there was also no significant
difference in means between the students from the University of Vienna (M = 3.88, SD = .11)
and the Vienna University of Economics and Business (M = 3.76, SD = .04); t(554) = 1.03, p
> .05. However, there was a significant difference in the trust level for professors (M = 2.84,
SD = 1.17) and for academics, who possessed only a PhD title (M = 3.35, SD = 1.01); t(203) =
3.37, p < .05, implying that professors are more skeptical towards online customer reviews. The
difference across the academic qualifications was not predicted, but it is consistent with the
assumption that eWOM trust decreases with a number of personal characteristics, such as
advanced age, widespread market place knowledge and awareness of manipulation strategies
that the consumer becomes aware of over time. By applying one-way ANOVA, the mean scores
on the eWOM trust scale of the respondents belonging to different academic disciplines were
also compared. This revealed no significant effect on the level of trust caused by the
membership of one of the six disciplines (F(5,199) = 2.05, p > .05). Similarly, there was no
significant difference in the means of the faculty members belonging to universities (M = 3.05,
SD = 1.13, n = 173), universities of applied sciences (M = 3.49, SD = 1.04, n = 23), or other
academic institutions (M = 3.59, SD = .75, n = 9); (F(2,202) = 2.51, p > .05).

Predictive and Postdictive Validity

Predictive and postdictive validity concern the relationship of the scale under development to
some external criterion separated in time and context (Netemeyer et al., 2003). More
specifically, it is the extent to which a scale on a given construct is related to the scores on a
criterion measure administrated (a) at a future point of time (i.e., predictive validity), or (b) at
267
a previous point of time (i.e., postidictive validity) (Cronbach & Meehl, 1955). Having said
that, the following research question was posed:

RQ 6: Does the eWOM trust measure significantly predict (postdict) eWOM-related

consumer behaviors (e.g., opinion seeking, adoption, giving, passing)?

To assess predictive validity, a sample of undergraduate business students from an Austrian

university was invited to participate in a short survey three months after the participants had
handed in a questionnaire with the eWOM trust scale. A fraction of the participants were
contacted in class and were asked to fill out a paper-based questionnaire, while to the second
half an email was sent including a link to an online survey. In total, 89 usable questionnaires
were obtained by this process. 64% of the sample were females and the average age was 22.5
years (range: 19 to 37 years). In the research instrument, participants were asked to rate a series
of statements concerning their online review behavior within the last two months. Specifically,
seven questions for measuring review purchase influence (Cheung et al., 2008; Lee & Koo,
2012; Park & Lee, 2009), six items for review seeking (Cheung & Hung, 2010; Chu & Kim,
2011; Park & Lee, 2009), and three items for review giving behavior (Bailey, 2005; Chu &
Kim, 2011) were included. Respondents were asked to answer the questions on a 7-Point Likert
scale ranging from 0 (strongly disagree) to 6 (strongly agree). The internal consistency
measures and correlations are reported in Table 40. The eWOM trust measure and the scores of
the measures of consumer review usage were strongly correlated, being .75 for purchase
influence, .74 for opinion seeking, and .42 for opinion giving (all p < .001). These correlations
represent a high predictive power of the eWT-S that is well above the regularly observed
correlation between attitude scales and behaviours. According to Wicker (1969), correlations
are seldom as high as .30 (cited in Hogg & Vaughan, 2011). In addition, considerable
correlations among the three behavioral constructs (ranging from .44 to .84) – all significant on
the .001 level – were revealed. These results indicated the ability of the new scale in helping to
explain and predict future consumer behaviors.

268
Table 40: Predictive Validity

Future eWOM
Developed Future Future eWOM
opinion
Mean SD measure of purchase opinion giving
seeking
eWOM Trust influence behavior
behavior
Developed measure of
3.57 .86 (.95)1
eWOM Trust
Future purchase influence 3.05 1.49 .75*** (.95)1
Future eWOM opinion
3.35 1.64 .74*** .84*** (.95)1
seeking behavior
Future eWOM opinion
1.74 1.65 .42*** .44*** .59*** (.83)1
giving behavior
Notes: Sample: 5 (n=89); 1= Internal consistency estimates (Cronbach’s alpha); Pearson correlation, 2-tailed;
*** = p < .001.

Another sample of university students was used to evaluate the scale’s postdictve validity. Here,
a similar data collection procedure was applied. While participating in a marketing course,
students were asked to fill out a questionnaire containing a series of 7-Point Likert items
emphasizing their behavioral patterns concerning online customer reviews within the last two
months. Specifically, the brief questionnaire included six items to measure the influence of
online reviews on past purchases (Cheung et al., 2008; Gilly et al., 1998; obermiller &
Spangenberg, 1998; Park & Lee, 2009), six items measuring past opinion seeking behavior
(Bansal & Voyer, 2000; Chu & Kim, 2011), and two items targeting past opinion giving
behavior (Chu & Kim, 2011). All items were measured with a 7-Point Likert scale ranging from
0 (stongly disagree) to 6 (stongly agree). Additionally, demographic variables were also
included. After three months, all respondents received an email with a link to the online follow-
up questionnaire containing the eWOM trust scale. On the introductory page, a definition of
online customer reviews and some examples were provided. Then, the students were asked to
complete the questionnaire for course credit. In total, 50 completed and assignable
questionnaires were used for analysis. 70% of the sample were females and the average age
was 23.9 years (range: 21 to 33 years). Table 41 reports the observed Pearson correlation
coefficients among the items. Internal consistency of the constructs was assumed, as the
Cronbach alphas were above the proposed threshold level. Exploratory factor analysis ensured
unidimensionality of the constructs in advance. Again, all correlations showed expected
patterns, as they were significant, considerable and positive ranging from .58 (for past purchase
influence and opinion seeking behavior; both p < .001) to .33 (for past opinion giving behavior,
p < .05). The analysis further showed that - as with the predictability sample - all behavioral
constructs were significantly related. Taken together, pre- as well as postdictve validity were
both supported by empirical data, indicating a high trust-behaviour consistency.

269
Table 41: Postdictive Validity

Past eWOM
Developed Past Past eWOM
opinion
Mean SD measure of purchase opinion giving
seeking
eWOM Trust influence behavior
behavior
Developed measure of
3.50 .73 (.93)1
eWOM Trust
Past purchase influence 2.82 1.37 .58*** (.96)1
Past eWOM opinion seeking
3.43 1.31 .58*** .80*** (.92)1
behavior
Past eWOM opinion giving
1.95 1.35 .33* .57*** .37** (.79)1
behavior
Notes: Sample: 5 (n=50); 1= Internal consistency estimates (Cronbach’s alpha); Pearson correlation, 2-tailed;
*** = p < .001; ** = p < .01; * = p < .05.

Nomological Validity

Responses from the main study were used to partly assess the measure’s nomological validity.
Table 42 reports the Pearson correlations between eWOM trust and measures dealing with
variables representing eWOM trust antecedents, correlates, and consequences. Positive
correlations were predicted between customer trust in online reviews and reviewer credibility
(H4), as well as distrust in reviewers’ motives (H12) as antecedents, between eWOM Trust and
eWOM involvement (H13) as a correlate, and between eWOM trust and eWOM usage, (H14)
as well as eWOM avoidance (H15) as consequences. By using data from study 4, all of these
correlations were proven to be significant and in the hypothesized directions. Specifically, the
correlation coefficient between scores on the eWT-S and the scale assessing perceived reviewer
credibility in general was .70 (p < .001) and positive. In line with these research predictions,
the correlation with perceived reviewer distrust was negative, substantial (r = -.33) and
significant (p < .001). According to the data, respondents who have heightened trust in online
customer reviews perceive this kind of market communication as more relevant. This fact was
manifested by a strong positive correlation between the new measure of the focal construct and
eWOM Involvement (r = .69, p < .001). The new scale also showed its ability to predict
hypothesized outcomes such as general eWOM usage (r = .76, p < .001) as well as general
eWOM avoidance (r = -.21, p < .001). Thus, all these hypotheses were supported. The
relationship between the individual constructs also behaved in a manner reasonable to the
rationale – supporting, as a consequence, the validity of the scale’s validity.

270
Table 42: Nomological Validity (Sample 4)

eWOM eWOM eWOM eWOM Reviewer Reviewer

Mean SD
Trust Usage Involvement Avoidance Distrust Credibility
eWOM Trust 3.73 .89 (.95)1
eWOM Usage 3.39 1.19 .76*** (.87)1
eWOM
3.86 .94 .69*** .64*** (.91)1
Involvement
eWOM Avoidance 1.56 1.23 -.21*** -.87* -.42*** (.86)1
Reviewer Distrust 2.43 1.18 -.33*** -.19*** -.41*** .53*** (.77)1
Reviewer
3.63 .80 .70*** .54*** .70*** -.27*** -.37*** (.90)1
Credibility
Notes: n=526; 1= Internal consistency estimates (Cronbach’s alpha); Pearson correlation, 2-tailed;
*** = p < .001; * = p < .05.

Further evidence for the nomological validity of the scale was provided by a series of street-
intercept interviews among Austrian Internet users. These samples were later later pooled to
represent sample 5. The separate sub-samples were used to evaluate the relationships between
the scale of the focal construct and measures of different constructs. In general, most of the
proposed relationships were supported by sample data. Specifically, with data from a sub-
sample (n = 268) the correlation between eWOM trust and two hypothesized antecedents,
namely perceived tie strength (measured with 10 items (De Bruyn & Lilien, 2008; Frenzen &
Davis, 1990; Money et al., 1998; 7-Point Likert scale) and perceived (behavioral, attitudinal)
homophily with online reviewers (measured with 22 items adopted from McCroskey et al.
(2006) and De Bryn and Lilien (2008); 7-Point Likert scale). Results showed that in line with
the proposed hypotheses (H7 and H8), the relationship among the measures of the constructs
was significant. That is, the correlations coefficient between eWOM trust and perceived tie
strength was .77 (p < .001), and .68 (p < .001) between eWOM trust and perceived reviewer
homophily.

A positive correlation was predicted between perceived reviewer altruism (H5) (construct
measured with 11 items adopted from Cheung and Lee (2012), Fang and Chiu (2010),
Podsakoff et al. (1990), and Price et al. (1995); 7-Point Likert scale) and a negative correlation
between perceived reviewer egoism (H6) (construct measured with 3 items adopted from Power
et al. (2008); 7-Point Likert scale) and eWOM trust in general. While a significant relationship
could be identified in the first case (r = .76, p < .001) by using data from another sub-sample (n
= 274), H6 did not find empirical evidence (r = -.12, p > .05). However, the results found a
negative but insignificant relationship.

According to literature, a negative relationship between the eWOM trust scale and a scale
assessing the different dimensions of consumer self-confidence (measured with 18 items taken
mainly from Bearden et al. (2001) and Rosenberg (1965); 7-Point Likert scale) was expected
271
(H2). By using data from an alternative sample (n = 80), however, a significant but positive
interrelation was discovered for information acquisition or the consumer’s confidence in his/her
ability to obtain needed marketplace information) (r = .29, p < .05), persuasion knowledge (i.e.,
the individual’s confidence in his/her knowledge regarding the tactics used by marketers in
order to persuade consumers) (r = .24, p < .05), and consideration set formation, which is the
defined as the consumer’s confidence in one’s ability to identify acceptable choice alternatives
(r = .28, p < .05) (Bearden et al., 2001). Bearden et al. conceptualize self-confidence as a
construct consisting of self-efficacy and self-esteem. Self-confidence and esteem are hence
highly intertwined. Radecki and Jaccard (1995) argue that persons with a low level of self-
esteem typically hesitate to reveal their knowledge gap, which ultimately decreases information
search. Similarly, Bishop and Barber (2012) hypothesize that (i) individuals with low self-
esteem tend to depend more on internal sources of information (i.e., themselves), and further
that (ii) individuals with low self-esteem are more likely to consult external resources, including
other persons (e.g., kin) or impersonal sources of information (e.g., Internet, product
discussions, advertisement). Both hypotheses were supported by empirical data. The same
relationship seems to surface in this study.

By using the same sample, nevertheless, H3 was supported by a significant positive relationship
(r = .51, p < .001) between eWOM trust and consumer risk propensity (measured with 2 items
adopted from Chang et al. (2008); 7-Point Likert scale). This implied that the more a consumer
is a risk taker, the more he/she is willing to trust information given in online customer reviews.

This research also predicted that consumer susceptibility to interpersonal information (H9),
attention to social comparison information (H10), eWOM influence (H14a), eWOM seeking
(H14b), and eWOM passing (H14c) would be positively correlated with eWOM trust. The
following scales were used: 16 items taken from Bearden et al. (1989) to measure CSII; 7 items
from Bearden and Rose (1990) to measure ATSCI; and a total of 18 items adopted from Park
and Lee (2009), Chu and Kim (2011), Bailey (2005), and Lee and Koo (2012) to measure the
different kinds of eWOM usage. All constructs were measured on a 7-Point Likert scale ranging
from 0 (“I strongly disagree”) to 6 (“I strongly agree”). According to the data obtained from the
sample (n = 120), all the correlations pertaining to these predictions were significant. The
weakest correlation was observed between eWOM trust and ATSCI (r = .19); however, it was
still significant on the p < .05 level.

H11 predicted that eWOM trust and eWOM skepticism (measured in the questionnaire with 9
items adopted from Obermiller and Spangenberg (1998); 7-Point Likert scale) were negatively
related. Accordingly, the sub-sample (n = 135) provided adequate support for this relationship
(r = -.93, p < .001). In total, as the majority of the hypothesized relationships behaved as
predicted, strong empirical evidence for the nomological validity of the scale across separate
samples is suggested. Table 43 provides an overview of the hypotheses tests.

272
Table 43: Nomological Validity (Sample 5)

Internal Correlation
Mean SD Consistency1 with eWOM Sample
Trust
eWOM Trust 2.78 1.21 .97 n.a.
Perceived Tie Strength 1.68 1.37 .95 .77*** (n=268)
Reviewer Homophily 2.26 1.31 .98 .68***
eWOM Trust 3.15 1.12 .97 n.a.
Reviewer Altruism 3.60 1.19 .95 .76*** (n=274)
Reviewer Egotism 2.10 1.12 .72 -.12
eWOM Trust 3.34 1.16 .97 n.a.
Consumer Self-confidence 2.49 .95 .83 .29* (n=80)
Risk Propensity 2.03 1.44 .81 .51***
eWOM Trust 2.82 .98 .94 n.a.
CSII (Informative) 3.17 1.16 .86 .34***
ATSCI 2.49 1.39 .88 .19*
(n=120)
eWOM Influence 2.30 1.42 .95 .74***
eWOM Seeking 2.42 1.54 .95 .75***
eWOM Passing 2.35 1.72 n.a. .59***
eWOM Trust 3.33 1.19 .97 n.a.
(n=135)
eWOM Skepticism 3.37 1.20 .91 -.93***
Notes: 1= Internal consistency estimates (Cronbach’s alpha); Pearson correlation, 2-tailed;
*** = p < .001; * = p < .05.

Social Desirability Bias

Questionnaire-based research always runs the risk that answers from respondents may be
distorted by a social desirability bias. Here, individuals are strongly motivated to provide
favourable responses in order to present themselves in a way that society regards as positive
(DeVellis, 2012; Steenkamp et al., 2010). Nowadays, being a eWOM truster might be evaluated
as socially desirable, as having trust in other persons or, more specifically, in the community of
fellow online shoppers, represents a personal characteristic valued by the modern consumption
society. Therefore, socially desirable responding may falsify responses to the new eWOM trust
scale. In order to assess the magnitude of this negative effect, this research followed the
conventions of earlier research (e.g., Morgan et al., 2009; Web et al., 2008) and correlated the
eWOM trust scale with the Crowne-Marlowe social desirability (CM) scale (Crowne &
Marlowe, 1960). For this purpose, data from a convenient student sample (n = 133) was used.
Relating the reduced 6-item CM scale to the overall eWOM trust score resulted in a low, non-
significant correlation (r = .17, p > .05), and also the correlations with the five first-order
constructs (ranging from .08 to .18) were statistically insignificant (p > .05). Only three out of

273
22 manifest indicators were significantly - but weakly - correlated with the CM scale: (1) In4
(r = .12, p < .05); (2) In6 (r = .18, p < .05); and (3) Ab9 (r = .20, p < .05); suggesting that the
validity of the scale is not seriously imperiled by socially desirable responding.

Supplemental Reliability and Generalizability Assessments

Test-Retest Reliability

In an effort to provide additional evidence for the reliability of the new eWT-S (besides split-
half reliability, etc.), the following research question was asked:

RQ 7: Is there a significant and considerable correlation between the measurements of

eWOM trust on two different assessment occasions?

Responses to the eWOM trust scale were collected on two occasions separated by three weeks
from a convenience sample of undergraduate business students. On the first occasion, a paper-
and-pencil questionnaire was administered to the respondents during five minutes of class time.
The students were informed that the survey related to their perceptions of online customer
reviews in general, and on the front page of the questionnaire a definition of customer reviews
was provided. A similar description was given at the second point of time. No additional
information was provided concerning the reasons for the second administration and no student
questioned it. All respondents participated for course credits. Because of absences and not
assignable questionnaires, a final sample consisted of 50 persons. The estimates of internal
consistency reliability (Cronbach’s alpha) ranged from .74 to .94 at the first phase, and from
.70 to .92 at the second phase (see Table 44).

The correlation between the first and the second administration for the overall eWOM trust
score was .73 and was significant on the .001 level. The test-retest correlations for the five sub-
dimensions ranged from .64 to .84. All of these correlations were significant (p < .001), and
thus they provide some evidence for the stability of the scale over time.

Generalizability to different Online Sources

Another concern in the context of this research is the applicability of the new scale to different
forms of eWOM sources. Hence, the following research question was proposed:

RQ 8: (a) Can the developed measurement of trust in eWOM reliably and validly measure
trust in specific eWOM platforms? (b) Does trust in eWOM vary across different
platforms? (c) On which platform is eWOM most trusted?

274
Table 44: Test-Retest Reliability

Internal Test-Retest
Sample Mean SD Sign.
Consistency1 Correlation
Test 3.46 .82 .95
eWOM Trust .73 ***
Retest 3.50 .73 .93
Test 3.18 .88 .92
Factor 1: Integrity/Honesty .73 ***
Retest 3.16 .80 .87

Factor 2: Test 4.28 .92 .93

.68 ***
Ability Retest 4.14 .87 .92
Test 2.04 1.33 .90
Factor 3: Benevolence .84 ***
Retest 2.36 1.28 .85

Factor 4: Willingness to Test 4.64 1.01 .94

.64 ***
rely Retest 4.67 .84 .87

Factor 5: Willingness to Test 3.18 1.21 .74

.65 ***
depend Retest 3.18 .96 .70
Notes: Sample: 5 (n=50); 1= Internal consistency estimates (Cronbach’s alpha); Pearson correlation, 2-tailed;
*** = p < .001.

Empirical data for answering this question was gained with a paper-and-pencil questionnaire
which comprised an adapted version of the eWOM trust scale. First, respondents were furnished
with descriptions and examples of the two major types of eWOM sources, namely consumer-
and marketer-developed sites (CDS vs. MDS). Then the questionnaire contained two sections
targeting the measurement of trust in these alternative sources of eWOM by using an adopted
version of the new eWOM trust scale. For instance, the item “The information given in online
customer reviews is credible” was transformed to “The information given in consumer-
developed eWOM sites is credible” and “The information given in marketer-developed sites is
credible” respectively, in order to fit the research context. As usual, these questions were
measured on a 7-Point Likert scale ranging from 0 (“I strongly disagree”) to 6 (“I strongly
agree”). Two different versions of the questionnaire with different ordering of these two
sections were used in order to control for ordering effects. In addition, nine items were included
adopted from literature (Choi & Rifon, 2002; Flanagin & Metzger, 2000; Johnson & Kaye,
2009) in order to measure the credibility of the two platforms on a 7-Point Likert scale, and 11
items measuring perceived reviewer credibility on a 7-Point semantic differential adopted from
Dou et al. (2012) originating from Ohanian (1990). The questionnaire also contained
eWOM/Internet usage patterns and demographic variables. Every fifth passerby at five heavily
frequented places in Vienna was interviewed after verifying that the prospective respondent
was an Austrian Internet user, had some experience with online customer reviews/was familiar
with eWOM on the platforms, and whether he/she exhibited the required demographic variables

275
in order to fit the quotas (gender, age group) on Internet users based on figures from Statistik
Austria (2012). In total, interviews with 176 consumers were conducted; 50.6 % were females
and the average was 31.2 years (range: 16 to 58 years).

The reliability and validity of the scale was assessed independently for the two types of review
sources. In both cases, the 22-item, second-order confirmatory factor model was estimated
using LISREL 8.5 and resulted in a significant chi-square test for both scales (χ2Cons = 406.80,
dfCons = 204, pCons < .001; χ2Mar = 504.43, dfMar = 204, pMar < .001), but acceptable measurement
fit (RMSEA = .06 (.07); NFI = .90 (.88); NNFI = .95 (.91); CFI = .95 (.91); SRMR = .05 (.07)
– results for the MDS in parentheses). Table 45 reports the measures to assess reliability, as
well as validity, of the measure. On the item level, all factor loadings were significant, as all of
them (except Be2 for marketer sites) exceeded the .70 threshold suggested by Nunally (1978).
Additionally, the corrected item-to-total correlations were acceptable and also the individual
items’ squared multiple correlation with their respective factors were .50 or better and,
therefore, surpass the recommended cut-off value of .50 proposed by Fornell and Larcker
(1981).

Concerning the reliability of the five first-order constructs, construct reliability was above the
minimum threshold of .60 recommended by Fornell and Larcker (1981) and, in both types of
sites, it exceeded the the preferred level of .70 (Churchill, 1979). Specifically, it ranged from
.87 to .96 for consumer sites, and from .74 to .95 for marketer sites. Other estimates of internal
consistency were also estimated for the first-order constructs. The Cronbach alphas were
considerable (ability: .94 (.93), integrity/honesty: .92 (.93), benevolence: .87 (.74), willingness
to rely: .96 (.95), and willingness to depend: .90 (.88)) and were all above Nunally’s (1978)
recommended level (.70) and in most cases also above the stricter .80 cut-off (Netemeyer et al.,
2003). It is agreed that the values of the average variance extracted – assessing the amount of
variance captured by a construct’s measure relative to random measurement error – of .50 or
above provide further evidence of the internal consistency of a construct’s measure (Fornell &
Larcker, 1981). For both types of sites, all the average variance extracted estimates achieved
this criterion on the first-order construct level. In addition the average inter-item correlation
ranged from .58 (.65) to .87 (.83) across the two sources of eWOM. On the second-order level,
evidence for internal consistency was provided by adequate levels of construct reliability and
AVE. The construct reliability measures exceeded the .80 threshold (Netemeyer et al., 2003)
with values of .91 in both cases and also the .50 cut-off value (Fornell & Larcker, 1981)
(.68/.66). Taken collectively, item loadings, composite reliability, variance extracted, and item-
to-total correlations provided support for the internal consistency of the eWOM trust measure
on its diverse construct levels.

Convergent validity was assessed by following Bagozzi and Yi’s (1988) recommendation that
all items should load considerably on their hypothesized dimensions, and that the estimates are

276
positive and significant. Table 45 presents factor loadings of all items, which turned out be
greater than .70 (.51), and a review of the t-tests for the factor loadings (greater than twice their
standard error) suggested that they all were significant. A similar pattern was observed for the
relationship between the first- and the second-order constructs, where the completely
standardized loadings ranged from .67 (in both samples) to .97 (.99), were significant at the
.01 level, and the parameter estimates were 10 to 20 times as large as the standard errors
(Anderson & Gerbing, 1988). Inspection of the squared factor loadings on the individual item
as well as first-order level revealed that the majority of the relationships were equal to or better
than .50. In addition, correlations among the five sub-dimensions of eWOM trust were
considerable, indicating high levels of convergence among the dimensions to measure overall
eWOM trust. These results added up to the conclusion that the new eWOM trust scale was able
to measure reliably as well as validly consumer trust in diverse forms of eWOM sources. As
the second part of the research question emphasized the level of trust in customer reviews on
CDS compared to MDS, a paired-samples t-test was deemed appropriate to evaluate potential
differences. The results indicated that the trust level was not significantly higher for CDS (M =
2.81, SD = 1.09) than for MDS (M = 2.74, SD = 1.05), t(175) = 1.38, p > .05. These results
suggested that while consumer-developed sites were found to posses a slightly higher level of
consumer trust, the difference in means between the two types of eWOM sources is
unremarkable, implying that online reviews are equally trusted. A further investigation
comparing individual review and commercial websites (e.g., Epinions) could provide a more
detailed insight.

RQ 9: Are consumer perceptions of (a) eWOM platforms’ credibility and (b) the
credibility of reviewers on these platforms associated with trust in eWOM on that
platform?

Two multiple regression analyses were used to test if platform credibility and reviewer
credibility significantly predicted respondents’ level of eWOM trust. The results of the first
regression concerning trust in online reviews on CDS indicated that the two predictors
explained 63.60% of the variance (R²Adj = 63.20, F(2,173) = 151.24, p < .001). It was found
that perceived platform credibility significantly impacted eWOM trust (βs = .42, p < .001), as
did perceived review credibility on these sites (βs = .44, p < .001). A similar result was observed
in the context of MDS. Here, platform credibility (βs = .58, p < .001) and reviewer credibility
(βs = .24, p < .05) likewise predicted eWOM trust on such sites. The two variables explained a
significant proportion of the variance in eWOM trust scores, R²Adj = 62.30, F(2,173) = 145.63,
p < .001).

277
278
Squared Corrected Average
Cronbach’s Construct Construct
Sub-dimension γ Item λ Multiple Item-to-Total Inter-Item AVE AVE
Alpha Reliability Reliability
Correlation Correlation Correlation
Ab7 .89 (.84) .78 (.71) .86 (.82)
Ab8 .88 (.81) .78 (.65) .84 (.76)
Ability .97 (.88) Ab9 .90 (.91) .81 (.82) .87 (.86) .77 (.72) .94 (.93) .94 (.93) .77 (.73)
Ab10 .90 (.85) .80 (.73) .87 (.81)
Ab11 .82 (.85) .68 (.73) .79 (.82)
In2 .80 (.83) .64 (.69) .70 (.80)
In3 .70 (.80) .50 (.63) .40 (.75)
In4 .88 (.82) .77 (.67) .79 (.78)
Integrity/
.93 (.99) In5 .85 (.81) .72 (.65) .66 (.79) .58 (.65) .92 (.93) .92 (.93) .64 (.66)
Honesty
In6 .88 (.80) .77 (.63) .78 (.76)
In9 .79 (.81) .63 (.66) .69 (.77)
In10 .86 (.81) .75 (.66) .77 (.74) .91 (.91) .68 (.66)
Be1 .85 (.81) .73 (.65) 58. (.47)
Benevolence .75 (.81) Be2 .79 (.51) .62 (.50) 62. (.40) .70 (.66) .87 (.74) .87 (.74) .69 (.50)
Be3 .85 (.82) .73 (.68) 65. (.52)
Wi1 .93 (.90) .86 (.80) .90 (.87)
Willingness to Wi4 .93 (.93) .86 (.87) .91 (.90)
.76 (.68) .87 (.83) .96 (.95) .96 (.95) .87 (.83)
rely Wi5 .94 (.92) .89 (.85) .92 (.89)
Wi8 .93 (.90) .86 (.81) .90 (.87)
Wi2 .79 (.78) .62 (.61) .81 (.72)
Willingness to
Table 45: Psychometric Properties of the eWT-S (Alternative Online Sources)

.67 (.67) Wi6 .89 (.85) .79 (.73) .74 (.76) .75 (.71) .90 (.88) .91 (.88) .76 (.71)
depend
Wi7 .94 (.90) .88 (.80) .86 (.81)
Notes: Table shows results for consumer-driven sites (CDS); Results for the marketer-developed sites (MDS) in parentheses; γ = Completely standardized second-order
loading; λ = Completely standardized first-order loading; AVE = Average variance extracted.
Generalizability to individual Customer Reviews

This research does not assume that eWOM trust as a consumer’s generalized tendency to rely
on online customer reviews is able to perfectly predict his/her response to every individual
review. For the impact of individual reviews, contextual variables such as diverse perceptions
of the review (e.g., wording), the reviewer itself (e.g., perceived competence, credibility) and
situation-specific personal conditions of the reader (e.g., mood) become pivotal. But, in
principle, generalized trust in eWOM should at least determine tendencies. Therefore, where
the eWOM trust scale is able to forecast reactions to a series of customer reviews, additional
evidence for the scale’s criterion validity and function is provided. Accordingly, the following
two research questioners were asked:

RQ 10: Is eWOM trust in general significantly and positively related to eWOM trust in
individual customer reviews?

RQ 11: Is eWOM trust in general significantly and positively related to attitude towards
individual customer reviews?

A convenience sample of students from two universities was asked to complete a paper-based
questionnaire, including the eWOM trust scale, in class for course credit. Three weeks later,
they received an email in which they were asked to visit a website containing a series of
consumer reviews and to respond to several questions targeting their attitude towards and trust
in the reviews just read. Respondents’ attitude towards the individual reviews was assessed
with six 7-Point semantic differentials adopted from earlier research (Bezjian-Avery et al.,
1998; Mitchell & Olson, 1981; Muehling, 1987; Olney et al., 1991) (e.g., “The reviews just
read are … good – bad”, “… convincing – not convincing”). In contrast, situational trust was
measured with a single Likert-like item measure (“I trust the information given in these
reviews.”) ranging from 0 (strongly disagree) to 6 (strongly agree). It was expected that these
items would represent responses to online customer reviews which should be related to eWOM
trust. Additionaly, a four-item Likert scale for product involvement (Zaichkowsky, 1994) and a
single question evaluating product category knowledge (Lee et al., 2008) were included. In
order to control for response bias, subjects were requested to indicate prior knowledge of the
reviews and familiarity with the brands discussed. Each respondent was exposed to a series of
four review sets, each one consisting of four reviews. The number of reviews (4), as well as
their length (average review length: 50 words), corresponded to regular review consumption
behaviors (BrithtLocal, 2013). The review sets discussed four fictitious products/brands (i.e.,
for a digital camera, a notebook, a hotel, a MP3 player). These products were selected because
of three criteria: first, the products should have at least some relevance to the respondents;
second, respondents should also use customer reviews at least sometimes when making
purchases in the investigated product category; third, both experience and search goods should
be included. In a pretest, it was verified that these two conditions were fulfilled for the selected
279
products. The review sets were displayed sequentially. The order of exposure was randomly
selected to control for order effects. Participants were instructed to read the reviews for each
product in the intensity and duration they regularly are used to. After that, they were asked to
respond to the scales described above before the next review set was presented.

If eWOM trust is a suitable predictor of responses to individual customer reviews, then the
correlation between the eWOM trust scale and the response measures should be significant and
positive, given that higher ratings in each of these scales indicate more generalized trust and
more favourable responses respectively. Table 46 reports the relationships. The correlation
between eWOM trust and average trust in the reviews was considerable and significant (r =
.66, p < .001). The positive relationship manifested itself in all four review sets. Correlations
were large (Cohen, 1992), ranging from .53 to .60, and were all significant on the .001 level. A
similar relationship was identified between eWOM trust and average attitude towards the
reviews (r = .58, p < .001), as well as the attitude towards the individual reviews (correlations
ranging from .48 to .53; all p < .001). Results suggest that respondents who are willing to trust
customer reviews in general found the reviews, on average, more favourable and more
trustworthy. This provides further empirical evidence for the predictive power of the eWT-S.
That is, situational influences (e.g., cues) can only to some extent predict consumers’
perceptions, attitudes and reactions towards OCR.

Table 46: Influence of Generalized eWOM Trust on Perceptions of Individual Reviews

Correlation
Internal with
Review Sets Mean SD Sign.
Consistency1 Generalized
eWOM Trust
Reviews Trust 3.59 1.10 n.a. .66 ***
1 (Digicam) 3.74 1.21 n.a. 0.60 ***
Individual Review
2 (Notebook) 3.59 1.29 n.a. 0.55 ***
Trust
3 (Hotel) 3.50 1.36 n.a. 0.53 ***
4 (MP3 Player) 3.52 1.32 n.a. 0.58 ***
Reviews Attitude 3.57 0.90 .91 .58 ***
1 (Digicam) 3.61 1.00 .90 0.53 ***
Individual Review
2 (Notebook) 3.60 1.04 .91 0.50 ***
Attitude
3 (Hotel) 3.52 1.09 .93 0.48 ***
4 (MP3 Player) 3.56 1.05 .91 0.50 ***
Notes: Sample: 5 (n=117); 1 = Internal consistency estimates (Cronbach’s alpha); Pearson correlation, 2-
tailed; *** = p < .001; n.a. = not available.

280
Generalizability to different Languages and Cultures

The eWOM trust scale was developed in German, with samples from both Austria and
Germany. To test the extent to which the proposed measurement approach is stable and
generalizable across alternative languages and cultures, the following research question was
advanced:

RQ 12: Is the eWOM trust scale applicable to different languages and cultures?

In order to provide first insights into this research topic, the thesis at hand investigates the
applicability of an English version of the new scale, using a sample of US consumers. While it
is recognized that insights from this research can only be regarded as first evidence for a cross-
cultural and cross-lingual application of the scale, testing it in additional contexts was beyond
the scope of this research and remains open for future investigation. However, testing the
robustness of the eWOM trust scale in the United States was appealing due to several
considerations: first, the United States offers an interesting study setting, as US consumers are
increasingly turning to online customer reviews (BrightLocal, 2013); second, the United States
is to some degree culturally different from Austria and Germany (i.e., Kogut and Singh’s (1988)
Cultural Distance Index (CDI)); third, a valid English version of the scale would enable a
relatively large research community to apply the new scale in their research. This said, data
discussed in this section stems from a representative sample of US online consumers (sample
6) and was collected by the procedures and research instruments described in Chapter 4 earlier.
The final sample was composed of 517 consumers using the Internet, of which 53.2 % were
females. The average age was 43.9 years (SD = 15.0) and ranged from 18 to 74 years (for details
see Chapter 4).

Collected data on the 22 scale items was first subject to an exploratory factor analysis using
principal components analysis with a Promax oblique rotation method. In order to test for
similarities in data structure, the number of factors to extract was set to five. This extraction
attempt resulted in a solution that deviated from the earlier findings in that a single item (Be2,
“social”) loaded on an individual factor. This was perhaps due to variability in the term’s
meaning in the two cultural hemispheres. Accordingly, the item was regarded as being biased
and was deleted from the item set. What followed was a re-estimation of the exploratory factoral
model for the remaining 21 items. The results of this procedure are provided in Table 47.

The same structure of the scale’s first-order constructs emerged, mirroring the earlier findings.
The five factors explained 77.73% of the total variance in the data (with the integrity/honesty
dimension explaining almost 60%) and an a-posteriori parallel test showed that a five-factor-
solution was most appropriate. All 21 items loaded considerably and significantly (with
loadings above .46) on their hypothesized factors and showed negligible cross-loadings with
the other dimensions. The majority of items had an item loading in the .80’s with acceptable

281
communality levels (≥ .71), providing first evidence for the robustness of the scale in alternative
languages/cultures (at least on the fist-order construct level). Before progressing to
confirmatory factor analysis, the item set was also subject to a review of the individual items’
properties (e.g., item means and variances, theoretical range vs. actual range) in accordance to
scale development guidelines (e.g., DeVellis, 2012). However, this analysis was not able to
identify problematic indicators and the investigation into the generalizability of the new scale
was continued.

Table 47: Results of the EFA (Sample 6)

Variance
Sub-
Explained/ Item Factor Loading Communality MSA
dimension
Eigenvalue
Ability 4.00% / .84 Ab7 .76 .81 .96
Ab8 .73 .71 .98
Ab9 .80 .83 .96
Ab10 .69 .74 .98
Ab11 .54 .76 .98
Integrity/ 59.06% / 12.40 In2 .75 .77 .98
Honesty In3 .59 .68 .98
In4 .85 .82 .97
In5 .87 .78 .98
In6 .80 .78 .97
In9 .60 .76 .98
In10 .88 .79 .98
Benevolence 3.52% / .74 Be1 .81 .84 .93
Be3 .86 .81 .94
Willingness to 7.26% / 1.73 Wi1 .92 .80 .96
rely Wi4 .84 .80 .94
Wi5 .80 .75 .95
Wi8 .91 .78 .97
Willingness to 2.89% / .61 Wi2 .71 .73 .98
depend Wi6 .46 .78 .96
Wi7 .58 .80 .96
Notes: Total variance explained: 77.73%; Extraction method: Principal Component Analysis; Rotation method:
Promax with Kaiser Normalization; Rotation converged in 9 iterations; Kaiser-Meyer-Olkin Measure of
Sampling Adequacy (MSA) .97 and Bartlett’s test of Sphericity: sig. .001.

To test the reliability and validity of the extracted dimensions, as well as the theorized structure
of the proposed scale, a confirmatory factor analysis (CFA) with LISREL 8.5 was used. Here,
a second-order CFA measurement model with the five first-order trust factors was specified,
similar to the earlier conceptualizations. The scale exhibited satisfactory psychometric
characteristics similar to the earlier findings. While the chi-square test was again significant (χ2
= 701.11, df = 184, p < .001), the absolute, incremental and parsimonious fit measures revealed
an adequate representation of the data by the model (GFI = .94, AGFI = .91, RMSEA = .07,
SRMR = .05; CFI = .95, NNFI = .95, NFI = .93; normed chi-square: 3.8, PFI = .70). This
suggested that the data fitted a higher-order model well.
282
As shown in Table 48, results indicated a good internal consistency of the measure on its
multiple levels. To assess the reliability of the observable items as indicators of the first-order
constructs, all path coefficients turned out to be substantial (ranging from .60 to 91), positive,
and significant, with t-values associated with the correlations that exceeded the critical value
(2.75) for .01 significance. Additional support for the existence of the hypothesized
relationships on the first-order construct level was provided by an examination of the squared
multiple correlation (R²) (Bollen, 1989). Fornell and Larcker (1981) suggest that an R² equal
to or above .50 indicates reliability of the measure, as a majority of the variance in the
(observable) measure can be attributed to its intended latent variable. The majority of the items
showed an R² well above the recommended threshold, with most items ranging in the .70s and
.80s. However, a single indicator (Ab8) fell short in this measure (.40) and also showed a
weaker (but nevertheless considerable and significant) factor loading (.60).

On the first-order level, all coefficient alphas of the individual sub-dimensions surpassed the
.80 level by ranging from .85 to .94 and achieved satisfactory levels of construct reliability.
Here, the construct reliability of the benevolence dimension was the lowest, however, showing
a value of .84, which was again considerably above the recommended .70 threshold. The AVE
of the five sub-dimensions also met the .50 minimum level by ranging from .65 to .73. The
review of path coefficients also indicated positive and significant loadings between the first-
and the second-order constructs. Additionally, the corresponding squared multiple correlations
turned out to be equal to or to surpass the recommended .50 threshold (range: .50 to .86). The
reliability of the first-order constructs as indicators of the higher-order eWOM trust construct
was evaluated in line with previous discussion by reviewing the construct reliability (.94) and
AVE (.74), which indicated a reasonable internal consistency across the five aspects of trust.
Taken together, the results provided reasonable evidence for the scale’s reliability.

According to Bagozzi and Yi (1991) and others (Anderson & Gerbing, 1988; Fornell & Larcker,
1981; MacKenzie et al., 2011), weak evidence for convergent validity is given by significant
correlation between an item and its hypothesized (higher-order) latent construct. As previously
discussed, all factor loadings (λ) were equal to or greater than .60 and showed t-values greater
than 2.57 (ranging from 15.21 to 27.26), implying a significant loading of sub-dimensions on
its observable indicators on the .01 significance level. The relationships between the second-
order construct and the first-order constructs showed a similar pattern, supporting the second-
order hypothesized structure. A more demanding indicator for the existence of convergent
validity is the squared factor loading, for which it is recommended that a threshold of .50 should
be exceeded. 20 of the 21 items had squared factor loadings (λ2) above this threshold. Only Ab8
fell short on this criterion. However, all five first-order constructs turned out to have a squared
factor loading (γ²) ranging from .50 to 86. Convergent validity can also be assessed in terms of
the extent to which the first-order constructs are correlated (Bagozzi and Yi 1991). All
correlations turned out to be considerable (ranging from .58 to .93) and significantly related (p

283
< .01), suggesting that they converge into a common underlying construct. All these analyses
(including the satisfactory AVEs on all levels) yield evidence that the English version of the
eWOM trust scale possessed satisfactory convergent validity for measuring consumer trust in
online customer reviews.

By now, one was able to argue for appropriate (convergent) validity, as well as reliability, of
the adopted scale. However, it is also advisable to test its robustness. In order to ensure that the
reported results were not biased by the idiosyncrasy of the development samples and that the
measure is applicable across different languages, the scale’s invariance (i.e., equivalence) was
successionally tested. This research’s approach follows the suggestions put forward by Doll et
al. (1998), who advocate that testing for invariance across samples is a particularly demanding
but insightful test of the instrument’s robustness. While the assessment of a measure’s
invariance regularly provides important implications for its construct validity and hence is
increasingly demanded by scale development guidelines (e.g., Netemeyer et al. 2003), its
application in literature is sparse (see Delgado-Ballester, 2004 for a notable exception;
Netemeyer et al., 1996).

The basic idea behind this test is that the relationships between the observed scores and the
latent variables shall be identical across independent groups. If the psychometric properties are
similar across the two groups (e.g., Austrian/German consumers (sample 4; n = 526) vs. US
consumers (sample 6; n = 517)), the eWOM trust scale is said to have equivalent meaning. This
consequently enhances the generalizability of the scale (Bollen, 1989; Marsh, 1995; Steenkamp
& Baumgartner, 1998). For this test, the same measurement model was applied to the data from
the two language groups in order to simultaneously assess the model’s equivalency across these
groups by constraining different sets of model parameters to be equal in both groups
(Worthington & Whittaker, 2006). Specifically, the measurement invariance was evaluated
hierarchically. A group-model (Model 1) – for which parameters were estimated across both
groups – was used as a baseline for comparison with subsequent models (Model 2-4) in the
hierarchy. Table 49 presents the fit estimates for this model.

Configurational invariance for the 21-item scale was established, as the baseline model showed
acceptable fit (χ2 = 1,334.42 (df = 368, p < .001), RMSEA = .07, CFI = .95, NNFI = .95, NFI
= .93). Subsequently, a model in which the factor loadings were constrained to be equal across
both groups was estimated (Model 2). This model also yielded acceptable fit (χ2 = 1,355.81 (df
= 384, p < .001), RMSEA = .07, CFI = .95, NNFI = .95, NFI = .93). A chi-square difference
test between this constrained model and the unconstrained baseline model performed to be non-
significant (χ2Diff = 21.40, dfDiff = 10, p > .05), implying full metric invariance across the
samples.

284
 Squared Corrected Average
Sub- Cronbach’s Construct Construct
γ Item λ Multiple Item-to-Total Inter-Item AVE AVE
dimension Alpha Reliability Reliability
Correlation Correlation Correlation
Ab7 .87 .76 .81
Ab8 .60 .40 .56
Ability .93 Ab9 .86 .74 .81 .64 .90 .90 .65
Ab10 .82 .66 .78
Ab11 .85 .73 .78
In2 .87 .75 .83
In3 .78 .61 .76
In4 .90 .81 .87
Integrity/
.92 In5 .84 .71 .81 .71 .94 .95 .71
Honesty
In6 .87 .75 .81
In9 .81 .66 .79
In10 .85 .71 .83 .94 .75

Be1 .91 .82 .73

Benevolence .71 .73 .85 .84 .73
Be3 .80 .64 .73
Wi1 .87 .76 .81
Willingness Wi4 .88 .77 .82
.83 .72 .91 .91 .73
to rely Wi5 .82 .67 .77
Table 48: Psychometric Properties of the eWOM Scale (Sample 6)

Wi8 .85 .71 .79

Wi2 .70 .50 .64
Willingness
.92 Wi6 .88 .78 .76 .66 .85 .87 .69
to depend
Wi7 .88 .78 .77
Notes: γ = Completely standardized second-order loading; λ = Completely standardized first-order loading; AVE = Average variance extracted.

285
For Model 3, factor loadings and factor variances were set equivalent. Here, the chi-square
difference test with the baseline model was significant (χ2Diff = 81.94, dfDiff = 21, p < .001),
suggesting that there is some non-chance lack of invariance. However, it has to be noted that
the same limitations apply to these results as to any other statistical test evaluating confirmatory
models. This means that “invariance constraints are a-priori false when applied to real data with
a sufficiently large sample size” (Marsh, 1995, p. 12). Therefore, the invariance should be
investigated also by considering alternative fit indices. If, according to these indices, the model
fits the data satisfactorily, sufficient evidence for model invariance exists (Marsh, 1995; Marsh
& Hocevar, 1985). In line with this argumentation, it was concluded that due to adequate fit
levels of Model 3, the factor loadings as well as factor variances were classified equal in both
samples.

The last model estimated (Model 4) assumed the equivalence of factor loadings, factor
variances, and error terms. Again, this model was compared with the baseline model. The
difference in the chi-squares was χ2Diff = 263.01, dfDiff = 42 and significant (p < .001). However,
the levels of fit for this (considerably) constrained model were acceptable (χ2 = 1,597.42 (df =
410, p < .001), RMSEA = .07, CFI = .94, NNFI = .95, NFI = .92), suggesting invariance across
the two samples. Simultaneously, the CAIC did not decrease considerably across the four
models, indicating comparable model adequacy. In summary, some evidence for measurement
invariance across the two languages/cultures existed which built further confidence in the
generalizability of the new scale.

Table 49: Measurement Invariance (Sample 4 and 6)

Competing Baseline Model Model Fit Indices

Model Chi Chi
df p df
Description Square Square Sign. RMSEA CAIC NFI NNFI CFI
Difference
Difference
1 Baseline 1,334.42 368 .001 n.a. n.a. .07 2,082.78 .93 .95 .95
2 Factor 1,355.81 384 .001 21.40 16 n.s. .07 2,010.51 .93 .95 .95
loadings
invariant
3 Factor 1,416.36 389 .001 81.94 21 *** .07 1,997.53 .93 .94 .95
loadings and
factor
variances
invariant
4 Factor 1,597.42 410 .001 263.01 42 *** .07 2,011.41 .92 .94 .94
loadings,
factor
variances,
and variance
of error terms
invariant
Notes: *** = p <.001; n.s. = not significant; n.a. = not available.

286
Scale Finalization and Development of Norms
A concluding consideration in the scale development process targets scale length optimization
(DeVellis, 2012; Netemeyer et al., 2003). The final eWOM trust scale comprised 22 (English
version: 21) items, each one mirroring one of the five sub-dimensions: ability: 5 items;
integrity/honesty: 7 items; benevolence: 3 items; willingness to rely: 4 items; and willingness to
depend: 3 items. In this thesis’ introduction, it was mentioned that the new scale should also be
a source for the further development of sub-scales, such as for the measurement of eWOM
trustworthiness. Hence, it was advisable to approach scale finalization by considering scale
length of the individual factors. Three of five factors were measured by four or more items,
allowing (individual) one-factor CFA models to be over-identified. It followed that all these
constructs were measured in accordance to general scaling recommendations put forward, for
example, by Netemeyer et al. (2003). Benevolence as well as willingness to depend dimensions
included three items each, implying that one-factor CFA models would be perfectly identified.
Three items represent the lower end of the number of indicators for valid and reliable construct
measurement. All constructs achieved considerable internal consistency. The three-item
constructs achieved a Cronbach alpha greater than .80, while the other constructs were in the
.90s range. This suggests that the later constructs had (theoretically) some “reliability to sparse”
for a possible trade-off between scale length and internal consistency (DeVellis, 2012;
Worthington & Whittaker, 2006). However, this research refrained from doing so due to several
reasons. First, all items performed considerably well across multiple samples, giving no reason
to exclude them without support from additional sample data. Second, the total eWOM trust
scale was not considered as too lengthy but placing a burden on the respondents which is still
adequate. Hence, the new scale possesses practicability. Respondents’ convenience was not
regarded to be a valued compensator of scale reliability. Third, Bagozzi (according to
Netemeyer et al., 2003) suggests that, for scales that target the measurement of personality traits
or similar constructs, each dimension should be ideally measured by 8 to 10 items. As eWOM
trust represents such a construct, a higher number of items was preferred to less. In sum, total
and factors’ scale length were considered adequate.

Finally, this research recognized the development of norms as a valuable final step of the scale
development process. Scholars such as Spector (1992) and more recently MacKenzie et al.
(2011) advance the notion that researchers should provide a frame of reference which aids in
the interpretation of the observed scores on a scale. In social sciences, the distribution of scores
represents such a frame of reference to which the score from an individual is regularly
compared. Investigating a population’s distribution of scores means administering the scale to
a representative, as well as large enough, sample of respondents of the population of interest.
The kind of data collection used in this research to obtain responses from sample 4
(Austrian/German consumers) and 6 (US consumers) suggested that these two samples are, to
a considerable degree, representative in respect to age and gender of the two target populations

287
(i.e., online consumers/Internet users). Additionally, as each sample includes more than 500
respondents, both samples are assumed to be of sufficient size to provide scores that are quite
stable. Table 50 provides the means and standard deviations of the scores for these two samples.
Additionally, the table also includes information concerning the shape of the distribution (i.e.,
skewness and kurtosis).

Table 50: Construct and Item Descriptives (Sample 4 and 6)

Construct Item Mean SD Skewness Kurtosis

eWOM Trust 3.73 (4.30) .89 (.95) -.41 (-.70) .80 (1.11)
Ability 4.31 (4.59) 1.01 (.97) -.69 (-.93) .16 (1.59)
Ab7 4.41 (4.81) 1.18 (1.14) -.66 (-1.21) .40 (2.26)
Ab8 4.19 (4.06) 1.18 (1.28) -.58 (-.64) .36 (.65)
Ab9 4.46 (4.79) 1.17 (1.13) -.76 (-1.06) .74 (1.53)
Ab10 4.16 (4.52) 1.16 (1.13) -.53 (-.74) .46 (1.13)
Ab11 4.33 (4.79) 1.11 (1.12) -.56 (-.93) .46 (.83)
Integrity/ 3.39 (4.12) 1.01 (1.08) -.33 (-.52) .34 (.53)
Honesty In2 3.50 (4.18) 1.24 (1.26) -.21 (-.57) .16 (.37)
In3 3.37 (4.28) 1.19 (1.16) -.28 (-.39) -.07 (-.02)
In4 3.33 (4.04) 1.18 (1.25) -.35 (-.45) .38 (.40)
In5 3.54 (4.13) 1.14 (1.30) -.30 (-58) .13 (.43)
In6 3.36 (4.14) 1.23 (1.29) -.37 (-.57) .24 (.34)
In9 3.24 (4.14) 1.18 (1.25) -.13 (-.65) .20 (.80)
In10 3.36 (3.96) 1.14 (1.26) -.20 (-.33) .47 (-.03)
Benevolence 2.89 (3.57) 1.16 (1.27) -.08 (-.06) .16 (-.13)
Be1 2.68 (3.63) 1.44 (1.41) -.03 (-.22) -.35 (-.27)
Be21 3.19 (3.74) 1.30 (1.54) -.25 (-.44) .25 (-.12)
Be3 2.81 (3.50) 1.34 (1.34) -.76 (.09) .02 (-.16)
Willingness to rely 4.58 (4.89) 1.10 (1.08) -.71 (-1.31) .44 (2.24)
Wi1 4.58 (4.87) 1.25 (1.20) -.79 (-1.28) .27 (1.97)
Wi4 4.58 (4.96) 1.20 (1.21) -.66 (-1.50) .33 (2.83)
Wi5 4.58 (4.78) 1.22 (1.30) -.71 (-1.38) .21 (2.26)
Wi8 4.60 (4.95) 1.23 (1.16) -.82 (-1.34) .52 (2.33)
Willingness to 3.45 (4.33) 1.28 (1.17) -.38 (-.75) -.10 (.83)
depend Wi2 3.27 (4.19) 1.56 (1.47) -.45 (-.77) -.36 (.32)
Wi6 3.50 (4.47) 1.49 (1.24) -.43 (-.99) -.33 (1.44)
Wi7 3.50 (4.34) 1.41 (1.30) -.43 (-.81) -.12 (.57)
Notes: Results for US respondents (sample 6) in parentheses; 1 = item not included in final US scale.

288
Scale Application: Segmenting Online Consumers
Segmenting Consumers on eWOM/Online Ad Trust

Another critical issue of this research is to demonstrate the usefulness of the newly developed
eWT-S in a greater research context. Therefore, the last research question concerns the
applicability of the new eWOM trust scale for developing an empirically-based typology of
online consumers in respect to their level of trust in online customer reviews (eWOM Trust) and
online advertising (oADTrust). Specifically, the following question was proposed:

RQ 13: (a) Can online consumers be meaningfully segmented according to their trust in
online customer reviews and online adverstising? (b) How can segments be
profiled?

As demonstrated earlier, research suggests that the levels of trust that consumers have
developed towards eWOM and online advertising as the two major sources of market
communication are likely to determine their responses. Answers to this research question not
only shed light on the influence of these two critical forms of online market communication
(e.g., persuasiveness, purchase influence) but also will enable marketers to implement target-
group specific communication instruments in order to stimulate positive market responses. As
it seems reasonable to group consumers according to the extent to which they rely on these
information sources, cluster analysis was employed to segment US online consumers according
to their responses to the new eWOM trust and online advertising trust scales. In its basic form,
cluster analysis is a multivariate statistical procedure that meets the objective of exploring
groupings of individuals in reference to certain characteristics they possess (Aldenderfer &
Blashfield, 1984; Hair et al., 2010). In this research, the mean scores of generalized trust in
online customer reviews as well as online advertising for each of the 517 respondents were used
as a basis for a two-step clustering procedure (Punj & Stewart, 1983). For the initial step, the
Ward’s hierarchical clustering method with squared Euclidean distances was applied in order
to get an insight into a potential clustering solution. For an adequate representation of the real
structure within the data, the elbow criterion was employed to determine the number of clusters.
It was concluded that the sample is best described by four clusters. During the second step, a
non-hierarchical, k-means clustering procedure (MacQueen, 1967) was used for developing a
four-cluster solution. Here, the group centroids calculated in the hierarchical procedure were
designated as the initial clusters for the subsequent k-means clustering. Tables 51 and 52
describe the identified empirical segments.

In general, the four distinct segments can be labeled as follows. Pure review trusters (19% of
respondents): consumers in this cluster express a high level of trust in online customer reviews
while showing a low level of trust in B2C market communication. Non-trusters (10%): the
smallest group in the sample is consumers who have trust neither in online customer reviews

289
nor online advertisements. It appears that these respondents tend to rely more on alternative
information sources, including internal (i.e., their knowledge) and external sources (e.g.,
friends). Moderately trusting customers (41%): the largest segment is composed of consumers
who score moderately on both trust in online customer reviews and online advertising. It can
be assumed that in this group the impact of and trust in eWOM/Online advertising is here
mainly triggered by situational circumstances. High trusters (30%): consumers in this group
show the highest level of eWOM trust while at the same time have high trust in online
advertising. Together with moderately trusting consumers, in terms of size these two clusters
represent more than two thirds of the sample (71%). Further analysis demonstrated that within
each cluster, eWOM trust was significantly higher than Ad trust. For instance, a paired-samples
t-test for the high-truster segment indicated that scores were significantly higher for the eWOM
scale (M = 5.13, SD = .50) than for the Ad trust scale (M = 4.79, SD = .68), t(153) = 5.42, p <
.001.

In accordance with the suggestions made by Maute and Dubé (1999), the cluster solution was
subject to a multivariate analysis of variance (MANOVA) to test its internal validity. More
specifically, the four clusters were compared in terms of the newly developed 22 trust-in-online-
customer-reviews items (α = .94), as well as the 9 trust-in-online-advertising items (α = .97). A
review of the Hotelling’s trace statistic showed significant differences between the identified
clusters (F(6,1022) = 463.25, p < .01; Hotelling’s trace = 5.44, partial η2 = .73) which therefore
represents reasonable empirical evidence for the internal validity of the proposed four-cluster
solution.

Profiling the Segments

Subsequently, the identified segments were first profiled in terms of group demographics and
followed by perceptual, attitudinal and habitual characteristics. Analysis of variance (ANOVA)
showed a significant difference in age between the groups (F(3,513) = 5.88, p < .001). Post-hoc
analyses using Tukey’s HSD indicated that consumers belonging to the fourth segment (i.e.,
high trusters) were the youngest (M = 39.81, SD = 13.39) when compared to the three remaining
groupings. By using a chi-squared difference test, it was found that clusters did not significantly
vary either in age (χ2(3) = 4.41, p > .05) or in education (χ2(18) = 27.13, p > .05). The same was
true for maritial status (χ2(6) = 6.33, p > .05), gender (χ2(3) = 4.41, p > .05) (however, pure
review trusters seem to slightly tend to be female), as well as ethnicity of cluster members
(χ2(12) = 16.00, p > .05). By using a one-way ANOVA, the clusters were found to not differ in
net income (F(3,490) = .64, p > .05). However, by applying a similar procedure, this research
found a significant difference in current employment among the segments (χ2(21) = 40.65, p <
.01). That is, white-collar workers and public servants were typically well represented in the
high-truster cluster, while retired persons were very unlikely to be found in this segment. On
the other hand, unemployed persons are often pure review trusters. Another chi-square test

290
showed that the clusters significantly varied in the living area (χ2(6) = 21.94, p < .001). High
trusters were well represented in the urbanized area, while the members of the remaining
segments more typically lived in the suburban area.

While some of these findings provide first insights into the characteristics of online consumers,
the results also suggested that demographic variables only could not meaningfully differentiate
the consumer segments. Prior segmentation studies agree in this aspect and advocate the
inclusion of further contextual variables (Schoefer & Diamantopoulos, 2009; Walsh et al.,
2010).

Table 51: Cluster Demographics

Cluster 1 Cluster Cluster Cluster

2 3 4

Moderately
Pure review Non-
trusting High trusters
trusters trusters
consumers

CLUSTER SIZE
Absolute 99 54 210 154
Percentage 19% 10% 41% 30%
DEMOGRAPHICS
Age
Mean 44.76 47.52 45.46 39.81
SD 15.81 16.50 14.84 13.39
Gender
male 37.4% 50.0% 49.0% 48.7%
female 62.6% 50.0% 51.0% 51.3%
Employment
Student 5.2% 5.6% 6.2% 3.9%
Employed for wages (white-collar 35.1% 27.8% 31.6% 47.4%
worker, public servant)
Employed for wages (blue-collar worker) 8.2% 13.0% 15.8% 14.9%
Self-employed 9.3% 13.0% 11.0% 7.1%
Retired 17.5% 22.2% 20.6% 9.7%
Unemployed (looking for work) 2.1% 7.4% 5.7% 5.8%
Unemployed (not looking for work, 19.6% 5.6% 5.3% 7.8%
unable to work)
Other 3.1% 5.6% 3.8% 7.8%
Monthly Net Income
Under USD 20,000 15.8% 14.0% 10.6% 8.0%
USD 20,000-29,999 14.7% 10.0% 11.1% 12.7%
USD 30,000-39,999 10.5% 12.0% 12.1% 12.0%
USD 40,000-49,999 10.5% 6.0% 14.6% 11.3%
USD 50,000-74,999 18.9% 34.0% 26.1% 27.3%
USD 75,000-99,999 16.8% 10.0% 17.6% 15.3%
Above USD 100,000 12.6% 14.0% 8.0% 13.3%

291
Table 51 (Cont.): Cluster Demographics

Cluster Cluster Cluster Cluster

1 2 3 4

Pure Moderately
Non-
review trusting High trusters
trusters
trusters consumers

Marital Status
Single 19.4% 24.5% 26.7% 25.8%
Married, living with another 68.4% 58.5% 57.8% 64.9%
person
Divorced, separated, widowed 12.2% 17.0% 15.5% 9.3%
Ethnicity
African-American (Non- 6.1% 11.1% 4.8% 13.2%
Hispanic)
Asian or Pacific Islander 3.0% 1.9% 4.8% 7.9%
White/Caucasian (Non-Hispanic) 85.9% 81.5% 85.1% 73.7%
Latino or Hispanic 3.0% 3.7% 3.8% 4.6%
Native American or Aleut 2.0% 1.9% 1.4% 0.7%
Living Area
Urban 19.2% 32.1% 21.0% 40.5%
Sub-urban 54.5% 49.1% 52.4% 39.9%
Rural 26.3% 18.9% 26.7% 19.6%

Accordingly, the four clusters were compared in terms of their perceptions, attitudes and habits
concerning online customer reviews and online advertising in general. As discussed earlier in
this thesis, consumers with heightened trust in eWOM were theoretically expected to exhibit
positive reactions towards online customer reviews than non-trusters. Therefore, this research
expected to identify higher positive perceptions and attitudes as well as cooperative behaviors
among pure review trusters and high trusters compared to the other two segments. However,
the circumstance that trust in B2C communication was found to differ in the high eWOM trust
segments makes certain differences likely.

Using one-way ANOVA, significant divergence in the consumers’ attitude towards online
customer reviews (α = .92) was found across the four clusters (F(3,513) = 97.75, p < .001). In
particular – and in line with expectations – the two clusters having high scores of eWOM trust
(i.e., pure review and high trusters) were shown to score significantly higher on the trait than
the other two clusters. Additionally, pure review trusters and high trusters exhibited similar
attitudes towards eWOM in general (see Table 52). Another insight of the nomological network
is that consumer’s trust in online customer reviews is formed by experiences over time,
implying that the perceived valence of these interactions has a rudimentary impact on later
developed generalized trust. This research therefore expected that individuals who have earlier
benefited from online customer reviews are therefore likely to perceive heightened trust (due
to operational conditioning) (Hogg & Vaughan, 2011). To investigate this assumption, another
one-way ANOVA was conducted showing significant differences between the clusters

292
(F(3,513) = 76.76, p < .001), in particular demonstrating that pure review trusters as well as
high trusters both seemed to have more positive experiences (α = .71) with eWOM than
consumers in the other two segments in the past.

It was also discussed earlier that positive relationships between eWOM trust and review usage
habits are likely to exist. The research at hand was able to demonstrate such linkages by using
a series of one-way ANOVAs. That is, variations in review seeking behavior (α = .90) between
the segments were identified (F(3,513) = 84.47, p < .001). There was also a significant impact
of segment membership and the impact of online reviews on purchase decisions (α = .91)
(F(3,513) = 117.67, p < .001). In both cases, pure review trusters, as well as high trusters,
showed a significant heightened positive behavior when compared to the other segments, but
did not vary essentially from each other in these respects. Another assumption was that high
eWOM trust will guide consumers to forward review content to others (i.e., review passing).
This research was able to demonstrate that this expectation is also mirrored by empirical data
(F(3,513) = 66.55, p < .001). However, the consumers who are most likely to forward comments
and recommendations from fellow shoppers are not pure review trusters but high trusters (M =
4.88, SD = 1.17). Pure review trusters, in contrast, possessed the second highest likeability but
they did not significantly differentiate from moderate-trusting consumers. Conducting another
ANOVA analysis provided the insight that differences did exist between the identified
segments in terms of review posting (F(3,513) = 25.46, p < .001). Here, high trusters were
(again) the most active consumers.

A further investigation of the segments’ characteristics revealed differences concerning the

perceived levels of risk in using online customer reviews (α = .87) (F(3,513) = 11.51, p < .001).
Here, as theorized, post-hoc analyses indicated that pure review trusters exhibited significantly
lower levels of risk when compared to the other segments. High trusters showed the highest
risk perceptions. So this group of consumers seems to be quite aware of eWOM’s perils. Using
another one-way ANOVA also showed significant differences between the segments in terms
of review avoidance behavior (α = .87) (F(3,513) = 13.23, p < .001). In line with expectations,
the segment of pure review trusters showed the smallest fear of contact (M = 1.01, SD = 1.19),
which was significantly smaller than in the other three clusters. High trusters showed a
significantly higher review avoidance behavior when compared to this group but also when
compared to moderately trusting persons.

Referring to the usage of online shops providing customer reviews, the analysis of variance
showed an effect of segment membership on site visits, F(3,513) = 10.46, p < .001. Post-hoc
analyses indicated that shop usage was higher within the segments with moderate and high
eWOM trust. Here, high trusters exhibited the most positive behavior, being more favourable
than in most other segments. However, there was no significant difference in respect to the pure

293
review trusters’ cluster. Low trusters are more likely to refrain from visiting commercial
websites.

Table 52: Cluster Description

Cluster Cluster Cluster Cluster

1 2 3 4

Pure Moderately
Non-
review trusting High trusters
trusters
trusters consumers

I. SEGMENTATION VARIABLES
Trust in Online Reviews (eWOM Trust)
Mean 4.57 2.69 3.96 5.13
SD .58 .82 .63 .50
Trust in Online Advertising (Ad Trust)
Mean 1.49 1.17 3.28 4.79
SD .72 .88 .50 .68
II. ONLINE REVIEWS ATTITUDES AND BEHAVIORS
Attitude towards Online Reviews
Mean 5.14 3.18 4.41 5.24
SD .77 1.20 .87 .67
Positive Review Experience
Mean 4.76 2.86 4.12 5.03
SD .87 1.21 .98 .93
Perceived Review Usage Risk
Mean 1.58 2.52 2.18 2.66
SD 1.29 1.05 1.21 1.96
Review Avoidance
Mean 1.01 1.96 1.57 2.20
SD 1.18 1.15 1.26 2.08
Review Seeking
Mean 5.12 3.15 4.44 5.33
SD .85 1.30 1.06 0.61
Review Purchase Influence
Mean 4.90 2.66 4.15 5.23
SD .88 1.25 1.01 .70
Review Forwarding
Mean 3.37 1.83 3.34 4.88
SD 1.79 1.55 1.52 1.17
Review Posting
Mean 2.32 2.06 2.47 3.19
SD .91 1.00 .99 1.17
Visiting Online Shops with Online
Customer Reviews
Mean 3.87 3.43 3.75 4.11
SD .77 1.00 .87 .78
(Continued on next page.)

294
Table 52 (Cont.): Cluster Description

III. ONLINE ADVERTISEMENT ATTITUDES AND BEHAVIORS

Attitude towards Online Advertisement
Mean 1.54 1.13 3.13 4.62
SD .92 .83 .71 .86
Online Advertising Usage
Mean 2.12 2.04 2.52 3.45
SD .87 .96 .93 1.03
IV. INTERNET ATTITUDES AND BEHAVIORS
General Internet Attitude
Mean 5.27 4.60 5.00 5.52
SD .89 1.29 .94 .61
General Attitude towards Online Shopping
Mean 5.13 4.41 4.69 3.69
SD 1.20 1.73 1.47 2.35
Online Shopping Behavior
Mean 3.72 3.26 3.49 3.77
SD .98 1.14 .98 .96
V. MARKET PERCEPTIONS AND CONSUMER PSYCHOGRAPHICS
Consumer Alienation
Mean 4.27 3.91 3.68 .87
SD 1.02 1.49 .88 1.26
Perceived General Product Quality
Mean 2.62 2.80 2.84 2.59
SD 1.29 1.41 1.06 1.63
Perceived General Price Fairness
Mean 2.52 2.11 3.20 4.22
SD 1.22 1.49 1.00 1.34
Self-Esteem
Mean 5.13 4.49 4.51 5.26
SD .80 1.32 1.00 .67
Self-Confidence (Decision Making)
Mean 4.02 3.98 3.43 2.75
SD 1.23 1.11 1.10 1.91
Self-Confidence (Persuasion Knowledge)
Mean 4.91 4.28 4.41 4.93
SD .83 1.38 .88 .78
CSII (Informational Influence)
Mean 3.06 2.44 3.39 4.32
SD 1.38 1.24 1.13 1.31
CSII (Normative Influence)
Mean 1.49 1.21 2.35 3.53
SD 1.35 1.12 1.32 1.86
Disposition to Trust
Mean 3.86 3.31 3.94 4.81
SD 1.25 1.24 1.00 .92

295
Due to a similar reasoning, this research also expects that high trusters or consumers that have
high beliefs in the reliability of online advertising also have favourable attitudes and show
positive approach behaviors towards this medium. Drawing on an analysis of variance, this
research was able to demonstrate that consumers’ attitude towards online advertising (α = .92)
is not equally distributed over the four samples (F(3,513) = 407.35, p < .001). More specifically,
ad attitude had a significant positive relationship with oADTrust. That is, samples with the lowest
oADTrust (i.e., pure review trusters, non-trusters) had a more negative attitude, while moderate
and high trusting individuals were characterized by a reasonably higher level of devotion. As
expected from ad literature, high trusters scored highest on this trait (M = 3.06, SD = 1.49).

Referring to the usage of online advertising information, the same literature proposes that
consumers that posses a heightened level of generalized oADTrust also more often seek advice
from this kind of market information (e.g., Soh, 2007). Assuming that high trusters find online
advertising information more valuable and are deliberately looking to consume advertised
products, it seems reasonable that these consumers belong to the segment of consumers who
are likely to use ad information for their purchase decisions. Therefore, this research expected
a positive relationship between ad trust and the indicated use of online advertising. Investigating
this assumption using a one-way ANOVA, significant differences between the clusters were
found (F(3,513) = 55.20, p < .001). In particular, the data demonstrated the hypothesized
increase in usage patterns, with high trusters showing significantly more attention to advertising
information than the other three segments. Usage behaviors did not differ significantly between
the segments with low oADTrust.

Another group of segment variables targeted the description of the perceptions of the Internet
as a research and shopping instrument. According to a variance analysis, significant differences
among the clusters existed in respect to general Internet attitude (F(3,513) = 18.54, p < .001).
More specifically, the two segments with increased trust in eWOM (i.e., pure review trusters
and high trusters) both also had a more favourable attitude towards the online medium.
Compared to this, low trusters were more opposed to the Internet. Another interesting
characteristic of different groups of online consumers is their attitude towards and usage of
online shopping. This research found that the four segments also significantly differ in these
respects. More precisely, pure review trusters showed the most favourable shopping attitude
which was, however, insignificantly higher compared to non-trusters and moderately trusting
consumers. In contrast, high trusters demonstrated an online shopping attitude that was
significantly lower when contrasted to the remaining three segments. The research further
demonstrated significant differences concerning online shopping behaviors (F(3,513) = 4.98, p
< .01). Subsequent analysis, however, showed that only the difference between the high trusters
and the low trusters segment was significant, where the former was found to have increased
usage patterns. The rest of the sample was homogeneous in this respect. In the course of the
empirical investigation, the consumers’ attitude towards online shopping in general was also

296
examined and a one-way ANOVA was able to find significant differences (F(3,513) = 15.72, p
< .001). Here, pure review trusters exhibited the most positive attitude (M = 5.13, SD = 1.20),
which nevertheless was only significantly higher when compared with the high truster cluster.
The latter was positioned on the other end of the continuum and was characterized with very
low scores (M = 3.69, SD = 1.84). In addition, non-trusters as well as moderately trusting
consumers were both found to feel more positively towards online shopping compared to high
trusters.

The final group of segment characteristics emphasized perceptions and attitudes towards the
marketplace and additional consumer psychographics. Dramatically changing consumer
attitudes make the topic of consumer alienation worth considering. According to literature, an
alienated person bears a profound feeling of separation, exclusion or estrangement towards a
particular social institution (Seeman, 1959). Interactions with this social entity are at the same
time typically associated with unpleasant and unfavourable feelings (Krishnan et al., 2009).
Some authors argue that alienation is a domain-specific phenomenon; as a consequence the
topic was discussed among a variety of disciplines, including marketing (e.g., Balasubramanian
& Kamakura, 1989; Burns, 2010; Johnson, 1996; Krishnan et al., 2009. Marketing scholars
have heavily drawn on Seeman’s conceptualization and have investigated the extent to which
shoppers feel alienated by manufacturers and vendors in the market place (Allison, 1978;
Lambert, 1980). This thesis adopts the definition of Mady (2011) and views consumer
alienation as the consumer’s “feelings of separation from the norms and values that
characterize the typical marketplace” (p. 194). Here marketplace includes all social institutions
that are involved with the offering of products/services to the consumer and all activities these
entities conduct (Johnson, 1996). The shift in market power has led to an increased importance
of the topic within the last years. Classic as well as current literature suggests that consumer
alienation causes several negative business outcomes. For instance, Lambert (1980) argues that
consumer alienation from the marketplace leads to increased dissatisfaction, as well as mistrust
towards the company. In addition, alienated consumers demonstrate a variety of avoidance
behaviors, such as switching to alternative market offerings, minimizing the interactions with
the company and negative word-of-mouth. These consumers refrain from identifying with the
market institutions, their outputs (e.g., products, brands), and are unlikely to accept usual
market practices (e.g., advertising) (Pruden et al., 1974; Shuptrine et al., 1977). Mady (2011) is
able to demonstrate a negative relationship between consumer alienation and sentiment toward
marketing. Therefore, this research expected consumer groups that have more adversarial
feelings towards the marketplace to be over-represented in the segments that show low levels
of oADTrust. A one-way analysis of variance showed a significant effect of segment membership
on the degree of consumer criticism of business principles (i.e., consumer alienation, α = .85)
(F(3,513) = 6.46, p < .001). As expected, pure review trusters were the most alienated
consumers when compared to moderately and high trusting segments. They strive to bypass the
information given by companies.
297
In contrast, further data analysis was not able to identify differences among the clusters
concerning perceived product quality; however, the segments showed divergences in respect to
perceived general price fairness (F(3,513) = 61.84, p < .001). While there was no difference
between the pure review trusters and the non-trusting segment, perceived price fairness seemed
to increase with trust in reviews/ads. Here, high trusters were most likely to believe that the
general market prices are justified.

As discussed earlier, generalized trust is likely to be dependent on a person’s self-esteem, as

well as self-confidence (see Chapter 3). Accordingly, these variables were also included in the
segmentation approach (α: self-esteem = .86; self-confidence (decision making) = .89; self-
confidence (product knowledge) = .88) and significant differences were identified due to a
series of variance analyses. Here, one of the outcomes was that pure review trusters, as well as
high trusters, were both characterized by a heightened degree of self-esteem, which
differentiates them from the remaining two segments but not from each other. As discussed
before, this implies that individuals with high self-esteem have no problem consulting external
information sources and demonstrating their lack of knowledge. In contrast, individuals scoring
low on the trait are more likely to use internal sources (Bishop & Barber, 2012). Segments
which showed a minor degree of oADTrust (i.e., pure review trusters and non-trusters) were
characterized by a high self-confidence in making adequate purchase decisions, while high
trusters had little confidence. In contrast, the segments also varied concerning the level of self-
confidence in persuasion knowledge (F(3,513) = 15.22, p < .001). Here, the two segments
which put high trust in online reviews scored significantly higher. This implies that these
consumers feel confident that they are able to identify misleading and fake messages.

As eWOM and online advertising are both likely to represent different kinds of social influence,
it also was reasonable to include variables measuring this degree. Specifically, variations
among the segments could be demonstrated concerning consumer susceptibility to
interpersonal influence (CSII) both for informational (CSIIInfo, α = .86) (F(3,513) = 40.01, p <
.001) as well as normative influence (CSIINorm, α = .91) (F(3,513) = 53.32, p < .001). In respect
to the former, non-trusters scored significantly lower, while high trusters indicated the highest
level (M = 4.32, SD = 1.31). Normative influence was also greatest in the high truster segment,
as expected. Pure review trusters and non-trusters scored relatively low on this trait. The high
truster segment is additionally characterized by the desire to trust others. Disposition to trust
(α = .90) varied significantly between the four segments (F(3,513) = 36.10, p < .001) and was
highest (as expected) in the high truster cluster, followed by moderately trusting consumers.
However, this segment did not differ from the level of interpersonal generalized trust measured
in the pure review truster group.

This thesis closes with a discussion of the results and an interpretation of the above-mentioned
findings in the final chapter.

298
Chapter 6

Discussion, Implications and Future Research

Summary of Research
Regardless of the insights of the majority of formal academic disciplines which have already
acknowledged the key role that trust plays in the context of interpersonal communication as
well as influence, research in the field of online consumer research put little emphasis on a
systematic investigation of the trust concept in the context of inter-consumer communication.
This situation may be not attributable to the fact that online scholars haven’t recognized the
construct’s general importance also in their own research area, but is explainable by a latent
conceptual confusion of what trust in eWOM really is. A variety of industry reports and
academic studies have nevertheless attempted to measure the construct or closely related
concepts (e.g., Briggs et al., 2002; Pan & Chiou, 2011; Racherla et al., 2012; Sen, 2007; Smith
et al., 2005). While this research stream represents a valuable contribution, especially to the
recognition of the important role of the trust concept for review research, these writings are
commonly characterized by a lack of understanding of the construct’s content and scope.
Besides these shortcomings in conceptualizing the focal construct, methodological problems in
deriving adequate measurement instruments limit application of existing self-reporting scales
in further research. To the best of the author’s knowledge, no measure was developed which
meets the standards and requirements of contemporary scale development literature until today.
In fact, online research fails by still using single-item measures or ad-hoc multiple item scales
that have not been evaluated in terms of their reliability and validity. The lack of an adequate
measure for a key concept of eWOM impact leads to confinements in the quality of scientific
research and the limited acceptance of the discipline’s insights by the academic community, as
well as practitioners. Most of all, a new measure would provide meaningful insights into the
role of trust as one of the most important critical mental mechanisms that facilitates online
interaction among consumers and ensures that the function of OCR as a valuable third-party
evaluation mechanism is maintained.

299
© Springer Fachmedien Wiesbaden 2017
W. Weitzl, Measuring Electronic Word-of-Mouth Effectiveness,
DOI 10.1007/978-3-658-15889-7_6
Given this state of affairs, this thesis intends to overcome these serious restrictions by
introducing a framework that is able to theoretically describe the construct of consumer
generalized trust in information conveyed in online customer reviews and recommendations,
developing an instrument of measuring it, and applying the new measure for investigating the
construct’s role in eWOM influence among clusters of consumers. The particular objectives of
the research are: (1) to clarify and advance the theoretical conceptualization of eWOM trust by
investigating the construct’s conceptual meaning, as well as scope, and to provide evidence on
the conceptual uniqueness of the construct which discriminates it from related but diverging
eWOM concepts; (2) to introduce a reliable, generalizable, valid and practicable multi-item
scale (eWT-S) that quantitatively captures trust in eWOM and is the result of a rigid
measurement development process; (3) to investigate the role of eWOM trust in an e-
commerce/trust framework by evaluating the concept’s antecedents, correlates, as well as
consequences; and (4) to apply the new scale to segment online consumers in accordance with
their generalized trust in C2C (eWOMTrust), as well as B2C communication (oADTrust), and
profile the identified segments in terms of attitudinal, perceptual and habitual characteristics.

To meet the aforementioned objectives, this thesis adhered to the given standards for the
accomplishment of the research process advanced by acknowledged scholars (Churchill, 1979;
DeVellis, 2012; Netemeyer et al., 2003) and cherished, approved research techniques. The
research process was characterized by a multi-stage/multi-sample design and was itself guided
by a series of 13 research questions, as well as 15 hypotheses which all targeted the further
assessment of the psychometric properties and applicability of the new measurement scale.

The first research question (RQ1) intended to evaluate the degree to which the eWOM trust
construct can be understood as a complex, higher-order construct that is best described in terms
of several facets. Further, this question also emphasized the investigation of the true nature of
the focal construct, as it wanted to clarify that the theorized ingredients of eWOM trust are also
mirrored by the empirical reality. Based on a careful review of research results from various
intellectual disciplines, as well as the insights gained from the qualitative and quantitative stages
of the research process, this thesis was able to globally describe eWOM trust as a consumer’s
general confidence that the information conveyed in electronic word-of-mouth (eWOM) is
reliable. In the context of this research, eWOM is equated and limited to online customer
reviews and recommendations which are understood as peer-generated, text-based evaluations
of market offerings posted on company or third-party websites by former, actual, or potential
customers (Hennig-Thurau et al., 2004; Mudambi & Schuff, 2010).

This said, eWOM trust in this research is formally and more specifically defined as the extent
to which a consumer believes that information given in online customer reviews and
recommendations is (a) useful, (b) honest, and (c) benevolent; (d) is willing to rely on or use
online customer reviews in general; and (e) is willing to depend on this kind of market

300
information. As this thesis defines eWOM trust as a person’s generalized disposition toward
relying on online customer reviews and recommendations, the focal construct is regarded as a
relatively stable, trait-like attitudinal concept that predisposes individuals to react to this kind
of information with internal perceptions as well as attitudes, but also determines the way
consumers react in terms of their behaviour. Accordingly, eWOM trust is conceptualized here
as an element of the human personality that triggers a consistent response tendency that can be
observed across a wide range of situations.

Further, this definition mirrors the understanding of eWOM trust as a psychological construct,
being best measured as a second-order construct that itself is reflected by five distinct types of
trust – each of these five dimensions belong to one of the three underlying general trust elements
(i.e., cognitive, affective, or behavioural). Hence, the conceptualization is consistent with the
classic three-component view of attitude (Hogg & Vaughan, 2011). This research further
concluded that on the first-order level, eWOM trust is most appropriately assessed by five
separate sub-dimensions: (1) (eWOM) ability (usefulness), (2) (eWOM) integrity/honesty, (3)
(eWOM) benevolence, (4) willingness to rely (use), and (5) willingness to depend (risk). eWOM
trust is therefore conceptualized as a psychological collective of these five factors of trust,
which themselves represent relatively stable confident beliefs in review information, as well as
behavioural intentions to use and rely on online reviews under conditions of uncertainty, risk,
and interdependence that are typical for this kind of interpersonal communication.

After rigid scale item purification and cross-validations, the thesis proposes a set of 22
observable indicators which seem to best capture the content and scope of the five eWOM trust
dimensions. Table 53 reports the final scale items for the English and German eWT-S.

301
Table 53: eWOM Trust Scale (eWT-S)

Subdimension Items

The information given in online customer reviews is …

Informationen in Online Kundenrezensionen sind …

Ability (eWOM helpful useful informative

Usefulness) hilfreich (Ab7) nützlich (Ab9) informativ (Ab11)
instructive applicable
aufschlussreich (Ab8) brauchbar (Ab10)

Integrity/Honesty genuine credible correct

(eWOM Reliability) echt (In2) glaubwürdig (In5) richtig (In10)
honest dependable
ehrlich (In3) verlässlich (In6)
true serious
wahr (In4) seriös (In9)

eWOM Benevolence caring social* kind-hearted

fürsorglich (Be1) sozial (Be2) gutherzig (Be3)

Willingness to rely on When I want to purchase a specific product/service, I am willing to visit online
(use) customer reviews to get relevant information.
Wenn ich ein bestimmtes Produkt/eine bestimmte Dienstleistung kaufen möchte, bin
ich bereit, Online Kundenrezensionen zu besuchen, um relevante Informationen zu
erhalten. (Wi1)
If I want to know more about the experiences of other customers with a
product/service/vendor, I am willing to look at online customer reviews.
Im Fall, dass ich mehr über die Erfahrungen anderer Konsumenten mit einem
Produkt/einer Dienstleistung/einem Händler erfahren möchte, bin ich bereit, Online
Kundenrezensionen zu besuchen. (Wi4)
I am prepared to consult online customer reviews in order to find out more about the
attributes of a product/service/vendor.
Ich bin bereit, Online Kundenrezensionen heranzuziehen, wenn ich mehr über die
Eigenschaften eines Produktes/einer Dienstleistung/eines Händlers erfahren möchte.
(Wi5)
I would use online customer reviews to find out more about a
product/service/vendor.
Ich würde Online Kundenrezensionen verwenden, um mehr über ein
Produkt/eine Dienstleistung/einen Händler herauszufinden. (Wi8)

Willingness to depend I’m willing to recommend products/services/vendors which have been recommended
on (risk) in online customer reviews to my friends and family.
Ich bin bereit, Produkte/Dienstleistungen/Händler, die in Online Kundenrezensionen
empfohlen wurden, meinen Freunden oder Familienmitgliedern weiterzuempfehlen.
(Wi2)
I am willing to make relevant purchasing decisions based on online customer
reviews.
Ich bin bereit, wichtige Kaufentscheidungen aufgrund von Online
Kundenrezensionen zu treffen. (Wi6)
When I make purchasing decisions, I am willing to rely on online customer reviews.
Wenn ich Kaufentscheidungen treffe, bin ich bereit, mich auf Online
Kundenrezensionen zu verlassen. (Wi7)

302
A series of research questions aided the evaluation of the final scale’s properties in terms of its
reliability, as well as validity. To assess the degree of concurrent validity, the second research
question (RQ2) addressed the capability of the scale to discriminate among groups of
consumers that indicated varying levels of trust in online customer reviews and
recommendations on an alternative, single measure of eWOM trust. For the validity assessment,
consumers’ scores on the new eWOM trust scale of two consumer groupings were contrasted.
The first group (high eWOM trusters) demonstrated a high level of trust on the single-item
measure, while the second group (low eWOM trusters) showed a low level of trust by using the
same single-item measure. Further analysis indicated that, by using the new measure, the mean
within the high eWOM truster group was significantly above, compared to the average score of
the low truster group. This result provided evidence that the new eWOM trust scale possesses
concurrent validity.

Convergent validity targets the extent to which a new measure corresponds with alternative
measurement approaches of the same construct. The third research question (RQ3) assessed
this issue by comparing the eWOM trust scale with four different measures: (1) a non-diagnostic
single item measure, (2) a Likert, multi-item measure of overall eWOM trust, (3) a semantic
differential, multi-item measure of overall eWOM trust, and (4) a qualitative measure of overall
eWOM trust. The empirical results derived from two independent samples showed significant
and strong correlations, demonstrating that the alternative measures are very likely to assess
one and the same construct. This supported convergent validity.

The multi-part research question four (RQ4a-d) intended to find a proof for the scale’s
discriminant validity. In respect to research questions 4a and 4b, eWOM trust was theorized to
differ from eWOM credibility (RCred), as well as consumer attitude towards eWOM in general
(RAtt) – two concepts which seem to be closely related but also conceptually different from the
focal construct. In the course of the research, the relationship between these three constructs
was subject to an empirical examination by using different methods. On the global level, the
constructs’ interrelation was tested particularly by the analysis of the correlation matrix, by
reviewing a Multi-trait Multi-method matrix (MTMM), as well as by comparing alternative
measurement models with confirmatory factor analysis. For the MTMM, consumer responses
for the three constructs were included in the form of two alternative scale formats: (a) Likert-
format scale and (b) semantic differential scale. Analysis provided additional evidence for the
convergent validity of the new scale, as well as strong support for its discriminant validity. The
latter finding was also supported by confirmatory factor analysis, which also demonstrated that
the measure of eWOM trust is distinct from the measures intended to assess consumer general
attitude towards online reviews, as well as perceived online review credibility. Further analysis
also included an assessment of the relationships between the individual eWOM trust sub-
dimensions and the other two constructs. Taken together, the analyses indicated that the eWOM
trust scale is related to eWOM attitude, as well as eWOM credibility, but more importantly

303
awards the new scale properties that cannot be attributed to the alternative constructs. This
research was able to show that the eWOM trust scale has some overlap with the measure of
eWOM credibility, as its dimension of integrity/honesty is conceptualized very similarly. In
addition, it is also agreed that the focal construct is not completely conceptually independent
from eWOM attitude, as some dimensions demonstrated meaningful inter-correlations.
However, the definition of eWOM trust advanced in this research also includes other facets that
are regularly not found in measurement approaches of the two alternative constructs; that is,
benevolence, willingness to rely (use), and willingness to depend.

Research question 4c and 4d strived for answers that should provide evidence that consumers
develop trust in online reviews that is different from their personal disposition to trust others
and is specific to the kind of trusting object discussed in this thesis. By means of confirmatory
as well as exploratory factor analysis, it was shown that consumers typically differ in their
internal tendency to put trust in generalized others and online reviewers. However, there exists
a significant relationship between these separated constructs, as more disposition to trust leads,
as expected, to increased levels of trust in eWOM. In addition, by measuring the level of trust
in different forms of market information (e.g., traditional word-of-mouth, information from
salespersons, advertising), this thesis was also able to demonstrate that consumers develop trust
that is specific to different kinds of market communication and the development of mental
attitudes takes place independently for the most part. Considered together, the results
empirically demonstrate the construct’s distinctiveness from related eWOM, as well as personal
characteristics constructs supporting the scale’s discriminant validity. This insight should alert
scientists as well as practitioners to carefully distinguish between the constructs when including
them in their survey research.

Another indicator of the scale’s quality is its known group validity (RQ5), which is the extent
to which a measure differs as predicted between groups which should score low and high on a
trait. Earlier contributions in communications research provided the necessary theoretical
ground for the assumption that faculty members share characteristics (e.g., intelligence,
education, age) that negatively impact this group’s general trust level in customer reviews and
therefore separates them from university students that were hypothesized to score higher on the
trait. As expected, the faculty sample showed an average eWOM trust level that was
significantly lower than for the student sample, thus supporting known group validity.

The scale’s predictive as well as postdictive validity was addressed in research question six
(RQ6). While predictive validity can be defined as the measure’s ability to effectively predict
a temporally later-occurring criterion, postdictive validity in contrast assesses the new scale’s
capability to predict a temporally earlier-occurring criterion. In the context of this research, this
criterion was referred to as future/earlier trusting behaviours (e.g., opinion seeking, adoption,
and giving). This was theoretically grounded on well-established research insights which have

304
demonstrated close relationships between an individual’s trust and subsequent trusting or risk-
taking behaviours. By using a sequential study design and data from multiple samples, this
thesis was able to find significant as well as reasonable linkages between these constructs. The
empirical results therefore provide evidence for the existence of predictive as well as postdictive
validity.

In order to test the nomological validity of the eWOM trust scale, fifteen hypothesized
relationships (H1-H15) between eWOM trust and a variety of theorized antecedents, correlates,
and response variables were assessed. Amongst others, it was hypothesized that the eWOM
trust construct is positively and significantly related with (a) eWOM involvement, (b) eWOM
usage, (c), perceived reviewer credibility, (d) perceived tie strength, (e) perceived reviewer
homophily, and (f) risk propensity. On the other hand, the construct was theorized to be
negatively related with (g) reviewer distrust, (h) eWOM avoidance, and (i) eWOM scepticism.
The majority of the defined relationships (13 out of 15) found empirical proof, as they showed
significant correlations as expected. Consequently, the eWOM trust scale was demonstrated to
have nomological validity. However, this research was not able to empirically demonstrate a
negative relationship of the focal construct with perceived reviewer egotism. This may be
explainable by the failure to capture negative attitudes towards eWOM sources by the
measurement instrument. Respondents may not like to express negative feelings or aversions
in respect to others. Additionally, the relationship between eWOM trust and consumer self-
confidence turned out to be significant but in the opposite direction to that theorized (for a
possible explanation see below).

As this research proposes a new scale that is applicable to paper-based as well as online
questionnaires, answers from respondents are not immune to common problems, such as the
adulteration by a social desirability bias. This research tested the scale by correlating it with the
Crowne-Marlowe social desirability scale. As this correlation was shown to be weak and
insignificant, it was concluded that the new scale is not seriously imperilled by socially
desirable responding.

The remaining research questions targeted supplemental reliability, as well as generalizability

assessments, which also included an evaluation of the scale’s test-retest reliability (RQ7). Here,
the main emphasis lies on the investigation of the measure’s temporal stability across a series
of sequenced surveys. By using data from a sample of 50 respondents and the administration of
the new scale in two survey occasions three weeks apart, this research was able to demonstrate
significant correlations between the two responses. This supported test-retest reliability, as the
new scale seemed to be capable of robustly measuring a personal characteristic across time.

Research question eight (RQ8) asked whether the eWOM trust scale could be generalized to
validly measure consumer trust in online reviews on different eWOM platforms. Two types of
online platforms were considered: consumer- (CDS) and marketer-driven (MDS) eWOM sites.
305
By applying confirmatory factor analysis, reasonable support for the WOM trust scale with the
original factor structure was gained for both contexts. Besides adequate characteristics in
respect to the scale’s convergent as well as discriminant validity, it also showed desirable
reliability properties, indicating that the scale is able to adequately assess trust in reviews on
consumer- (e.g., online discussion forums) as well as marketer-developed (e.g., merchant
websites) review sites. Consumers showed similar trust in both types of sites. This finding is in
contrast with other research (e.g., Bray & Schetzina, 2006). However, this study maybe was not
able to find differences as the respondents were prompted to think about the two categories in
general and not specific examples. Additional evidence for the scale’s validity was provided by
investigating the relationships between platform-specific eWOM trust and platform’s eWOM
credibility (RQ9a), as well as perceived credibility of reviewers on these two platforms (RQ9b),
which are both theorized to represent correlates/antecedents of the focal construct. Empirical
evidence mirrors a reasonably strong and significant relationship between the constructs.
Hence, this study led to the acceptance of the scale’s ability to generalize over different platform
contexts.

This research has conceptualized eWOM trust as a personal characteristic of the consumer that
predisposes the individual to rely on online customer reviews and recommendations. In its
function therefore, the construct is assumed not to perfectly predict his/her response to every
individual review but explain to some degree behavioural and attitudinal tendencies towards
specific reviews. Accordingly, research questions 10 and 11 both targeted this issue by striving
for evidence for a positive relationship between generalized eWOM trust and trust in (RQ10)
and attitude towards individual reviews (RQ11). In the course of an empirical investigation,
respondents were first asked to indicate their personal level of eWOM trust in the new scale
before later evaluating different sets of online reviews for four alternative products (brands).
The results showed – as assumed – that eWOM trust can be regarded as a good predictor of
individual review perception and/or attitude.

The fact that the eWOM trust scale was developed in German – by using empirical data from
Austrian as well as German respondents – may limit its application to this cultural and language
context. In order to overcome this restriction, this thesis intended to demonstrate the robustness
of the scale by distributing an English version to a sample of US online customers and to test
its psychometric properties (RQ12). By means of exploratory as well as confirmatory factor
analysis, this research was able to provide initial evidence for the scale’s invariance across
different languages and countries. All but a single item (i.e., “social”) were demonstrated to be
integral elements of the adopted scale showing adequate psychometric properties. The same
was also true for the proposed overall measurement model, which gives confidence that the
proposed eWOM trust scale is likely to capture the nature and the occurrence of the construct
among consumers, independent of the language context of its application. The fresh sample of
US respondents also aided the establishment of multi-sample scale norms and the

306
demonstration of differences in the level of trust among consumer groupings (i.e., Americans
placed more trust in eWOM than Germans/Austrians), which may be attributable to the cultural
and social background.

The last research question (RQ13) emphasized the new scale’s contribution to meaningfully
segmenting online consumers. Here, the thesis at hand investigated a typology based on trust
in online customer reviews (eWOMTrust) and trust in online advertising (oADTrust) as
independent dimensions. The implication of this examination is that contemporary consumers
typically have access to two major forms of online market communication – that is, online
customer reviews (eWOM) as consumer-to-consumer information and online advertising
(Online Ads) as business-to-consumer information. As consumers develop trust that is specific
to these communication forms – making them more or less amenable to their contents and
specific claims – both orientations need to be included in order to form groups that vary in
respect to their persuasibility by marketer-driven vs. public-driven communication. The cluster
analysis resulted in a four-segment solution. Pure ad trusters, who solely rely on information
provided by companies, were not found in the sample. Instead, a segment with high trust in
online reviews and non-existing trust in online advertisements was uncovered (Pure review
trusters, 19% of respondents). The remaining three segments were characterized by an
increasing amount of eWOM, as well as ad, trust. For instance, Low trusters (10%) formed a
segment within which trust in both forms was not existing. In contrast, a third (Moderately
trusting consumers, 41%) and a fourth segment (High trusters, 30%) were identified,
suggesting that a large proportion of consumers are relatively open to information given online
and are likely to integrate it in their purchase decision; however, trust in the opinions and
recommendations of fellow shoppers was significantly higher compared to its marketer-driven
counterpart (in every cluster).

The variety of research questions and selection of hypotheses enabled this research to conduct
a rigid scale development process that included a series of demanding tests for the scale’s
reliability, validity, generalizability, practicability, as well as cross-national stability. These
tests conjointly provided evidence for the psychometric soundness of the scale. Therefore, the
eWOM trust scale provides a parsimonious, reliable, and valid measurement instrument which
could be used both for scientific theory testing endeavours, just as well as for practical
identification of eWOM trusting consumers.

Conclusions and Discussion of Major Findings

In this thesis, the author advances the conceptualization of eWOM trust as a key construct in
the consumer research domain. The following section discusses the four major findings in
greater detail.

307
The Five Sub-Dimensions of eWOM Trust

By drawing from an integration of interdisciplinary literature, trust in online customer reviews

and recommendations (eWOMTrust) is conceptualized as an enduring orientation or personality
characteristic which can be specified as a higher-order construct comprising the dimensions of
(1) (eWOM) ability (usefulness), (2) (eWOM) integrity/honesty, (3) (eWOM) benevolence, (4)
willingness to rely on (use), and (5) willingness to depend (risk). Given such a
conceptualization, the new eWOM trust scale (eWT-S) overcomes problems of extant
measurement instruments by being theoretically anchored, as well as specific to the domain of
online consumer research, and by reflecting all dimensions relevant for measuring
interpersonal, impersonal trust among humans on the Internet. The above-mentioned
dimensions to a large degree conform to the original structure derived from literature. However,
the empirical assessment demonstrated that eWOM trust is not reflected in a purely affective
attitude dimension (i.e., likeability) leading to its exclusion. Further, eWOM trust also triggers
a variety of behavioural intentions that vary on a risk continuum and also represent different
kinds of risks or vulnerability. Accordingly, the initial willingness to rely dimension was split
into two separate sub-dimensions. The research demonstrates that the identified dimensions
mirror the typical content and the nature of a trust construct, but also that the actual factor
structure depends on the specific kind of trust object, demonstrating the necessity to introduce
a specific eWOM trust measure.

The first dimension, (eWOM) ability, refers to a consumer’s belief about how useful,
knowledgeable, and adequate the information given in online reviews and recommendations is
for purchase decision-making. The concept is a critical component of eWOM trust, as it reflects
the generalized degree of perceived practical information value of customer reviews, enabling
consumers to make their purchasing decisions more effectively, and thereby optimizing the
purchase outcomes. In trust research, the concept of ability or the similar concepts like
competence or expertise are typically included in conceptualization of trust. Ability is often
generally defined as the group of skills, competencies, and other characteristics which enable a
trusted party to have influence within some specific domain (Mayer et al., 1995). The very
similar notion of competence surfaces when comparing the nature of the ability construct with
the domain of competence. The latter is regularly defined as the trustee’s ability to solve a
specific task successfully. Therefore, the concept generally reflects the extent of how well the
trustee performed a certain job (Butler & Cantrell, 1984; McKnight & Chervany, 2002). Barber
(1983), a typical representative of this school of thought, defines the concept as the individual’s
belief that the interaction partner (or trust object) exhibits a technically competent role
performance. In the context of eWOM, role performance can be assessed by judging eWOM’s
ability to fulfil its primary function for consumer decision-making; that is, furnishing the
consumer with valued and essential market information on products, services, brands, etc. and
giving adequate advice. The dimension of eWOM ability can therefore be conceptualized as the

308
consumer’s relatively stable belief that online customer reviews and recommendations are a
useful source of purchase-related information.

The second sub-dimension of eWOM trust, labelled (eWOM) integrity/honesty, is defined as a

consumer’s belief that the information conveyed in online customer reviews and
recommendations is honest, truthful, verified, and adheres to moral standards. This research
conceptualizes a variety of classic trust concepts, such as the concept of integrity (i.e., the
consumer’s belief that the trust object adheres to moral standards) and honesty (i.e., the
consumer’s belief that the trust object is not lying and is believable) to be represented in a single
facet which may be also named reliability, or the consumer’s belief that online customer
reviews and recommendations keep their promise and consumers can depend on them when
making purchasing decisions. In trust research it is quite common to assume that concepts like
integrity, morality, credibility, reliability, honesty, believability, and dependability represent
the same construct (Doney & Cannon, 1997; Larzerele & Huston, 1980; McKnight & Chervany,
2002). Others focus on a single concept and regularly explain it in terms of the other concepts.
For example, Chen and Dhillon (2003) include integrity in their trust conceptualization and
define it as the trust object’s characteristic of acting in a consistent, reliable, and honest manner.
In the present research, terms like honest, genuine, true, credible, and dependable all
demonstrate that the domain of this sub-dimension is the consumer’s evaluation of the
information given in eWOM in respect to its truthfulness which triggers his/her volition to rely
on its content.

eWOM benevolence, as the third dimension, mirrors the consumer’s belief that eWOM is guided
or motivated by the reviewers’ favourable and positive intentions towards the welfare of fellow
shoppers and not by the reviewers’ manipulative- or self-interest only. Online reviews are a
special kind of social (although impersonal) interaction whose information content is directly
associated with the characteristics as well as intentions of its creators. Trust in eWOM is
therefore inter-social trust, and literature agrees that in such a context the inclusion of a
benevolence dimension mirroring social-affective aspects of the communication is reasonable
(e.g., Cheung & Lee, 2006; Doney & Cannon, 1997; Le & Turban, 2001). Similarly to this
literature stream, it is here argued that the extent of perceived benevolence is a critical
component of eWOM trust, since without a positive reciprocation and general detection of
reviewers’ altruism, the truster would not consider this kind of information either genuine or
believable. eWOM is regularly provided by anonymous individuals who possess the
opportunity to deliberately steer fellow shoppers wrong by recommending faulty products or
by making false statements about the product’s quality. Therefore, a consumer’s general
proposition to appraise the reviewers’ underlying motivations as courteous or deceitful is
regarded as a good indicator for the presence of trust.

309
This research also finds that two forms of behavioural trust intentions mirror eWOM trust; that
is, the consumer’s willingness to rely on (or use) the information given in online reviews, as
well as his/her willingness to depend on it and take serious risks. Both constructs mirror the
consumer’s intention to act (e.g., make a purchase decision) due the contents and
recommendations given in review. However, the two constructs are shown to be separated due
to varying degrees of involved risk. Trust in eWOM therefore is mirrored both by the
consumer’s intention to generally use online reviews during the decision-making process and
accepting a potential (limited) loss (e.g., being misguided) (i.e., search risk), but also by the
extent of how much the consumer is willing take serious risks by depending solely on the
information given in online reviews and recommendations (i.e., purchase risk). The latter also
includes social risk. Hence, the two behavioural dimensions differ in risk quality (high vs. low),
as well as risk type. In general, various scholars have formed the conceptualization of the trust
construct as being intrinsically tied to the concept of risk-taking (e.g., Coleman, 1980; Deutsch,
1958; Rosseau et al., 1998) and there are numerous studies which included behavioural
intentions in their measurement instrument (e.g., Albrecht & Travaglione, 2003; Mayer &
Davis, 1999; McAllister, 1995; Moorman et al., 1993). The research at hand is consistent with
these contributions and extends it with respect to a differentiation of risk nature.

In contrast to other conceptualizations of trust (e.g., Swan et al., 1988), eWOM trust does not
include a separate eWOM likeability dimension. A possible explanation might be that the kind
of trust object ultimately determines the content and the nature of the individual’s trust in it.
Online customer reviews and recommendations are typically characterized by their strong
informative, topic-focused character and their subordinate affective content. Emotions in
respect to the contents, as well as to the reviewers, are hardly evoked on the side of the reader.
Rather, readers are typically induced to purely process the information cognitively. Therefore,
affective evaluations in terms of eWOM’s likeability or enjoyability seem to hardly be
associated with the level of trust in eWOM. The exclusion of the purely emotional attitude
dimension leads to a more parsimonious view of the eWOM trust that more exactly fits the trust
conceptualizations of mainstream research when measuring comparable phenomena (e.g.,
Mayer et al., 1995).

The Three Elements of eWOM Trust

This research has provided empirical evidence that eWOM trust is best regarded as a multi-
dimensional construct of higher order. In Chapter 3 it was shown theoretically that the various
sub-dimensions of eWOM trust can be categorized in accordance with their intrinsic
characteristic, which mirrors a three-component view (Hogg & Vaughan, 2011). Various trust
scholars propose that the concept is a matter of (i) cognitive thoughts or rational expectations,
together with (ii) affective processes or emotional bonds towards the trust object, as well as (iii)
behavioural intentions (e.g., Barber, 1983; Cummings & Bromiley, 1996; Granovetter, 1985;

310
Mishra, 1996). Accordingly, it was hypothesized that (eWOM) ability and (eWOM)
integrity/honest both reflect the truster’s cognitive beliefs about the trust object. In contrast,
perceived (eWOM) benevolence is a kind of cognitive expectation which is based on the
truster’s social-affective orientations, hence it was classified as an emotional element of trust.
(eWOM) likeability or affect was theorized as another emotional element mirroring the stable
personal feelings of the truster towards online reviews. The third element of eWOM trust was
trusting motivations or specifically the consumer’s willingness to rely on the review
information, given latent risk. The empirical results demonstrated that such a theoretical
conceptualization is likely to represent the true nature of the construct, as almost all elements
of trust are represented in the final trust scale. As such, this thesis is consistent with the
contributions of earlier researchers that regarded trust as a composition of trusting beliefs and
behavioural trusting intentions (Lewis & Weigert, 1985; Luhmann, 1979; McKnight et al.,
1998; Moorman et al., 1993), as well as to the findings of those scholars who conceptualize
trust in terms of a mixture of cognitive as well as emotional aspects (e.g., Ohanian, 1990; Swan
et al., 1988).

However, a person’s confidence that the information given in online customer reviews is
reliable (i.e., eWOM trust) is not reflected by his/her feelings evolved over time towards this
kind of communication. Hence, eWOM affect was excluded from the construct’s measurement.
In general, it seems that eWOM trust is mainly mirrored by the cognitive aspects of the
constructs. The majority of trust items belong to eWOM ability as well as eWOM
integrity/honesty, which belong to the cognitive dimensions. In addition, the two constructs
explained most of the total variance of the construct. Therefore, one can say that eWOM trust
is mainly cognitive-based. This supposition corresponds with the findings of earlier research,
which asserts that the kind of social relationship between the truster and the trustee or the trust
object finally determines the content of the trust construct. eWOM trust is conceptualized here
as personality characteristic, which deals with a relatively stable consumer orientation towards
relying on a specific kind of market communication. Therefore, the constructs assess the extent
of how much an individual develops a relationship with a group of anonymous fellow shoppers
(a virtual community) and with a specific kind of social institution; or, in other words, a
relationship with a secondary group. Under such conditions, literature generally proposes that
the cognitive aspect of trust is more relevant than in primary group relationships (e.g., in-person
interactions with friends or acquaintances) – here, the emotional aspects tend to be more
important (Johnson-George & Swap, 1982; Rempel et al., 1985). In literature, multiple
examples for comparable secondary group relationships emphasize that trust is more based on
cognitive expectations or evaluations. For instance, in the widely cited work of Mayer et al.
(1995), the researchers measure a person’s trust in an organization in terms of cognitive beliefs.

311
The Characteristics of eWOM Trusters

In order to test the scale’s nomological validity, the construct’s relationship with other
constructs was tested. The process also enables this research to characterize eWOM trusting
consumers on a variety of variables. These insights can be used to address these consumers as
a target segment by buzz marketing. In this thesis, eWOM trust is conceptualized as a relatively
stable, internally-caused individual difference characteristic and it is agreed with trust literature
that the construct is dependent on several antecedents. These include more basic personality
traits, prior consumer experiences, introspective and extrospective observations, as well as
social-affective attitudes.

Among the personality-based trust antecedents, a person’s willingness to depend on others and
his/her general tendency to take risks in life are both positively related to eWOM trust. Research
demonstrates that disposition to trust and risk propensity are also determinants of critical
consumption-related constructs, such as trust in e-vendors (e.g., Gefen, 2000). High eWOM
trusters exhibit both a heightened propensity to trust others and are also more likely to take risks
in general. In contrast to expectations, this research has found a significant negative relationship
between consumer self-confidence and his/her trust in online customer reviews; that is,
individuals with high self-confidence are more likely to trust eWOM. A considerable body of
literature proposes an inverse relationship. For instance, Lee and Neville (1994) argue that
individuals with a low self-confidence are more likely to trust another, external source for
information over themselves. In this research, self-confidence was defined as a consumer’s
confidence in his/her capabilities in respect to making educated marketplace decisions and
behaving as a smart consumer. A possible explanation for the observed relationship (besides
the one already proposed) may be that self-confident individuals regard online review
information only as affirmation of their already-made decision or prior, firm expectation. They
are more likely to depend on themselves. Therefore, they have fewer reasons to mistrust
reviewers, as the information created by them is regarded as inferior. Further, self-confident
consumers may be believed to possess the ability to identify misleading market communication
(e.g., reviews) and are able to react appropriately to it. Future research could help to clarify the
background of the identified relationship.

This research was able to demonstrate that eWOM trust is also dependent on cognitive-based
trust antecedents which represent cognitive evaluation processes which are associates with self-
observations or self-perceptions, such as reviewer and recommender characteristics (e.g.,
reviewer credibility), and self-interests. eWOM trusters have a positive picture of review
creators in general, mirrored by favourable perceptions of reviewers’ credibility, as well as low
reviewer distrust. Hence, high eWOM trusters regard reviewers in general as very competent
to give advice, as well as very believable. The confidence in eWOM is also affirmed by multiple
affective-socially-based trust antecedents which refer to the emotional and social bonds

312
between the reader and the reviewer. High trusters generally assume that the act of review
writing is driven by the reviewers’ motivation to do something good for others and serve fellow
shoppers without anticipating any extrinsic reward.

eWOM trust is further positively influenced by the positive perception of social ties between
the communication parties. High eWOM trusters consider that a strong social tie binds them
with the members of the shopper community, so that they are not regarded as complete
strangers. The close social relationship is strengthened by consumers’ perceptions of attitudinal,
demographical, and behavioural similarities between them and eWOM creators. Hence,
reviewers are commonly regarded as an externalized “self” of the trusting consumer. They are
perceived as individuals who share the same interests and have similar demands from market
offerings. High eWOM trusters identify themselves with reviewers.

Further, the level of eWOM trust is affected by the individual’s susceptibility to interpersonal
influence or, in other words, with his/her willingness to interact with others regarding
consumer-related topics. Accordingly, high eWOM trusters are shown to have a considerable
tendency to accept information from others as evidence about reality. Simultaneously, this
group of consumers is likely to be particularly sensitive to reactions of others to their
consumption behaviour. Persons with a heightened level of eWOM trust typically care about
what others think of them and hence tend to obtain social clues from various sources in the
marketplace, including online customer reviews. Within this kind of information, they find
valuable information about the appropriateness of the consumption of different market offerings
by fellow shoppers that represent a referent or aspiration group. Online reviews, thus, can have
an important normative influence. By investigating the construct’s correlates, the thesis was
able to show that high eWOM trust is associated with personal involvement in this form of
communication. eWOM trusters feel that the information given in reviews has high personal
relevance for them, as it mirrors valued characteristics. Consequently, scepticism towards
eWOM is unincisive among high trusters, leading to increased review usage. Here, this
consumer grouping not only passively consults reviews during information search, but is also
more likely to purchase recommended products. Additionally, there is a positive relationship
between eWOM trust and information-forwarding and review writing. In a nutshell, high
eWOM trusters appear to represent a very distinct group of online consumers forming a
community of shoppers who share an affinity for the exchange of market knowledge and
collectively strive for individual consumption wisdom. They are very conscious, self-aware
consumers for whom the opinions and claims of others trigger their own consumption
experience.

Profiling Recipients of Online Market Communications

This research also alerts researchers as well as practitioners to carefully distinguish between
segments of online consumers who rely on different kinds of market communication. The
313
rationale of the segmentation is that specific groups of individuals can score low (high) on trust
towards the two major forms of online communication – that is marketer-driven online
advertising (oADTrust) and customer online reviews (eWOMTrust). Both orientations have to be
included as segmentation bases in order to target specific audiences with adequate marketing
communication instruments.

Pure review trusters (19% of respondents) are strongly oriented towards information given by
other shoppers. More specifically, they have developed a high confidence in online customer
reviews, while having no trust in online advertising messages at all. Such consumers appear to
be females for the most part; office workers as well as unemployed persons who are not looking
for a job are well represented in this segment and a major proportion is living in a relationship.
Like most other segments, Pure review trusters typically reside in a suburban area.
Unsurprisingly, this kind of online consumers is characterized by positive perceptions, attitudes
as well as behaviours towards consumer-to-consumer communication, while having quite a
negative stance towards online advertising in general. However, this segment does not hold the
most positive attitude towards reviews – this is reserved to the High trusters cluster. Anyhow,
Pure review trusters have a favourable stance towards reviews which is for the most part
triggered by prior positive experiences with this information source and a positive disposition
towards others. They are typically less critical towards reviews and think that their usage comes
with very limited risk. Hence, they are much more likely to accept eWOM claims and
recommendations. These persons passively consume and actively seek purchase-related
information published by fellow shoppers and this information also largely impacts their own
purchase decisions. Pure review trusters also are involved in contributing review postings and
feel safe to forward recommendations and claims read to others. The segment holds a
favourable attitude towards the Internet in general and shops online regularly. Its attitude to e-
commerce is tremendously well developed and clearly the best of all segments. These
consumers’ low trust in online advertising seems to be related to an intense feeling of consumer
alienation, which guides this segment to have a relatively hostile attitude towards companies
and the marketplace. Further, Pure review trusters are characterized by a high self-esteem and
self-confidence – they especially think that they are able to make the right purchase decision.
However, the normative influence of others is very limited compared to other segments.

The cluster of Non-trusters comprises consumers who put no trust in any form of online market
information. With a cluster size of 10%, it is by far the smallest segment in which members
seem to be quite non-responsive to regular market communication. These consumers appear to
consult their own knowledge, as well as other external information sources (e.g., family
members, magazines), during the purchase process. Non-trusters are slightly older than the
members of the other segments which is also mirrored by a heightened rate of retirees. They
benefit from a relatively high income. Naturally, this segment holds a relatively negative
attitude towards reviews, as well as online advertising, and generally refrains from using these

314
kinds of information in their purchasing decisions. Non-trusters do not have much affinity with
the Internet and stand out from the other segments with unfavourable attitudes towards the
Internet in general, and online shopping in specific. This also possibly explains their limited
intentions to purchase online. They feel alienated by the companies, but not as much as pure
review trusters. In parallel, they are sceptical in respect to the general price fairness on the
market. This segment shows the smallest susceptibility towards normative as well as
informative influence by others, which is accompanied by a slightly lower than the average
extent of trust in others.

Moderately trusting consumers represent, with 41%, the biggest cluster. It shares some
demographic characteristics with the Non-truster segment, but members are more likely to be
found in rural areas. However, this segment demonstrates a more positive attitude and
behaviours towards both online customer reviews, online advertising, as well as online
activities. Their trust in eWOM as well as online advertising seems to be situation-specific,
which highlights the importance of perceptual cues and contextual evaluations of the content,
message, etc.

The segment High trusters (30% of respondents) is a grouping of consumers which is rich in
contrast. Here, members have developed a high level of trust in both online reviews and
advertising, which makes them open for both kinds of communication. It is the second largest
segment. Its members are relatively young and live above-average in the urban area. Their
attitude towards online reviews is more favourable than in the Pure review trusters segment.
However, their scepticism towards this medium seems also to be more distinct, as they are more
aware of the potential risks embedded and they also have a tendency to avoid reviews for some
purchasing decisions. In contrast, High trusters source information given by fellow shoppers
and belong to a segment of consumers that is most active in distributing this information to
others. Members of the cluster possess by far the most favourable attitude towards online ads
and regularly consult them while being online. The Internet in general seems to have an
important role in their lives, as they can be regarded as “Internet literates”. Individuals of the
Generation Y or the Digital natives cohort typically belong to this segment. Their online
shopping behaviour is comparable to the characteristics of the Pure review trusters segment,
while visiting online shops with customer reviews more often. High trusters have an open and
favourable orientation towards companies and their business practices but also tend to accept
given market prices. However, they are slightly more sceptical concerning product quality,
which may cause them to collect additional information via various sources. They have a well
developed self-esteem (which guides them to external information search), but their trust in the
various forms of market communication seems to be triggered by a low self-confidence in
making purchasing decisions on their own. As in the High trusters segment, members have
confidence in their ability to identify misleading or false communication. Further, such
consumers appear to be very susceptible to both normative as well as informative influence.

315
Hence, in the eyes of High trusters, they face the serious problem of being reliant on external
information and advice in order to come up with a smart purchase decision. Here, both
marketer- as well as consumer-driven communication are two valued sources to overcome this
burden. However, when it comes to the edge, eWOM is more trusted than online advertising
(as in the other segments). This influence parallels with the consumers’ strong tendency to
generally rely on others. In a nutshell, this research indicates that a large body of online
consumers is still open to online advertising. However, this coexists with a similar openness
towards eWOM and, if it comes to the crunch, eWOM seems to be more influential.

Contribution and Implications

This research makes a theoretical as well as practical contribution. From the theoretical
perspective, this thesis is one of the first of its kind to introduce a reliable and valid
measurement instrument which can be used in the context of customer online research. As
outlined in the introductory chapter, this research string has already recognized the potential
explanatory power of the eWOM trust construct. However, contextual research is characterized
by the usage of poorly assessed scales which often mirror the conceptual confusion among
researchers concerning what to include in the construct and which kinds of dimensions most
appropriately mirror consumers’ general reliance on online customer reviews and
recommendations. Typically, the applied scales have not been tested for internal reliability or
their necessary validity. Under such conditions, this research is able to make multiple
contributions. First, the development of a reliable and valid scale, in general, brings the desired
standardization to the study of the focal construct and enables scholars to investigate its
characteristics and relationship with critical concepts of consumer behaviour more
systematically. By providing a basis for common conceptual agreement, the new instrument
can also stimulate further research collaborations among international marketing scholars.
Further, through the theoretical recognition of the construct’s role and the collective usage of
an adequate measurement instrument, research benefits as this adds to its interpretability and
the acceptance of its findings.

More specifically, the application of the eWOM trust scale representing a valid and reliable
indicator of the extent of consumers’ confidence in online customer reviews aids researchers to
gain a profound knowledge of how and which consumers are influenced by others on the
Internet. Preliminary evidence from online customer research suggests that personal
characteristics and general orientations of the consumer impact his/her reactions to online
reviews (e.g., De Valck et al., 2009; Park & Kim, 2008; Zou et al., 2011). This research has
demonstrated that generalized trust in reviews has a far-reaching impact on the perceptions of
individual reviews in terms of developed trust in and attitude towards these reviews. Both the
recognition of the role of generalized eWOM trust as a determinant of review perceptions and
the inclusion of direct eWOM trust effects can make a valuable contribution to the research of

316
consumer responses to social information available on the Internet. The introduction of the
eWOM trust construct in research frameworks and its measurement with the new scale may
also enable researchers to explain some earlier inexplicable phenomena, as trust potentially has
a critical role as a moderator between review perceptions and outcomes or as a mediator.

Further, this research makes a theoretical contribution, as it is specific to online research.

However, the insights of interdisciplinary research helped to develop a theoretically sound
conceptualization of eWOM trust. As Chapter 2 has shown, various disciplines have deeply
investigated the concept of trust. Nevertheless, online consumer research has only recently
discovered the construct’s explanatory power, leading to little systematic research in this field.
As the study of consumer trust in online shopping (e.g., in e-vendors, in the Internet as a
purchasing channel) has already shown, scientific progress can be stimulated by an in-depth
knowledge of the nature of the core concept and an analysis of the eWOM trust phenomenon
from a broad interdisciplinary perspective (e.g., Gefen, 2002b; Gefen & Heart, 2006; Mayer et
al., 1995). The eWOM trust scale is theoretically anchored and derived from multiple scientific
disciplines. This circumstance represents a launching pad for the explanation of diverse
eWOM-related consumer behaviours from the viewpoint of interdisciplinary research.

In literature, contributions for segmenting consumers in respect to their communication patterns

and attitudes are surprisingly scarce. While its key role for social relationships has been
recognized for a long time, the concept of trust – in a similar way – is hardly used to identify
target segments for marketing actions n(see Dimitriadis et al., 2011; Hultén, 2007 for notable
exceptions), while being used in other contexts such as political voting (e.g., Schiffman et al.,
2002). This research overcomes some of these limitations by making a first step towards
grouping online consumers according to their reliance on the two major forms of online
marketing communications. As such, it completes our knowledge about online consumers
advanced by earlier research (e.g., Barnes et al., 2007; Bhatnagar & Ghose, 2004; Brengman et
al., 2005; Ganesh et al., 2010).

From a managerial perspective, the new scale potentially enables some practical insights. Prior
research, as well as the results of this thesis, imply that consumer trust recognizably affects the
evaluation of review information and its impact on shopping and business outcomes. Empirical
findings support the theoretical assumption that high trusters respond more positively
(negatively) to positive (negative) reviews or recommendations than low trusting consumers
(e.g., Bambauer-Sachse & Mangold, 2013; Tormala & Petty, 2007). The developed scale can
help to anticipate the impact of social information in diverse communication channels and
forms among different consumer groups and target audiences. The scale, here, not only enables
marketers to assess generalized review trust, but furthermore can help to measure trust in
specific messages, reviewers, and online channels.

317
By developing this thought further, this research has demonstrated that using the new scale can
provide meaningful insights into essential topics of online market communications. The
findings of the segmentation approach suggest that the identification of groupings of
communication recipients who differ in terms of trust in online reviews and advertising and
their subsequent targeting with group-specific communication efforts can trigger desired
consumer responses. More specifically, literature as well as recent studies suggest that
consumer scepticism towards advertising is constantly rising and represents a serious problem
for marketers (Nielsen, 2012). The research at hand, however, shows that a relatively small
amount of online consumers (29%) have low levels of trust in online advertising. A third of
these consumers seem to be not responsive to any kind of market communication whatsoever.
Hence, they are also unresponsive to negative eWOM (e.g., shit storm). The other two-thirds
can be labelled pure review trusters. The high level of trust in online reviews results in a high
influenceability of these consumers by opinions of fellow shoppers and the social climate
towards a product or brand on the Internet. The social information is accented without
questioning it. This is true for positive, but also negative, reviews. Companies face the challenge
that they are completely unable to steer direct communications and thus can hardly elicit
consumer responses. Negative eWOM can have an especially serious impact on business
outcomes. This is intensified by the circumstance that pure review trusters appear to have a-
priori strong hostile feelings towards companies and their offerings in general.

Low trusters, in contrast, seem to neglect information given by other shoppers, as well as
marketers. This can be explained by a low susceptibility to informative as well as normative
information and the orientation towards alternative information sources like personal
experiences. The research has shown that a large proportion of almost two-thirds of online
consumers has moderate or high trust in online reviews and online advertising and thus is
responsive to marketer communication efforts. Especially, high trusters represent an interesting
and actionable consumer grouping. Such consumers appear to have high trust in reviews;
however, this trust comes with some scepticism towards its characteristics and reasonable risk
perceptions of using this kind of information. The segment is also very active in sharing
personal experiences with products and brands. This information transfer, however, may be
also influenced by advertising contents, to which high trusters hold a very favourable and open-
minded stance. That is, these consumers can be used as a trigger of viral marketing. They also
have a very positive attitude towards companies, their products and the marketplace. Purchasing
decisions are typically characterized by a strong usage of social information. This may be
attributable to a low self-confidence in decision-making, as well as an openness to informational
and normative influence. These characteristics imply that high trusters highly consider both
consumer- as well as marketer-issued product clues, but they are particularly receptive to
product claims distributed homogeneously in various communication channels and by various
originators. Communication messages targeting this consumer group should therefore integrate
the two major online communication forms and should additionally resonate well with the
318
characteristics of the high trusters. The emotional, cognitive as well as behavioural
characteristics of the high truster segment provide some hints that its members show heightened
customer engagement (Brodie et al., 2013) and activism. This surely needs further scientific
clarification. However, if a relationship exists, companies may benefit from a number of
consequences of customer engagement, including increased trust (e.g., Hollebeek, 2011),
satisfaction (e.g., Bowden, 2009b), loyalty (e.g., Bowden, 2009a), and commitment (e.g., Chan
& Li, 2010). The findings, on the other hand, also warn marketers to be cautious, as no group
of consumers exists that solely relies on online advertising. Additionally, consumers – across
all segments – typically trust eWOM more than product information provided by companies.
This is especially critical when conflictive information is available.

Besides planning and managing communication efforts that are target group-specific
(advertising and buzz management), the new scale also enables marketers to develop more
efficient social recommendation systems. When implementing social instruments on their
websites, marketers have to decide, for instance, in which form information is provided (e.g.,
star ratings, pictures) and which complementary cues (e.g., information on the reviewer’s name,
origin, purchasing history) it should feature. Similarly, they have to decide which channels to
use. For instance, recommendation systems can be included on the corporate website or on an
independent, external site (e.g., [Link]). By asking consumers to evaluate
alternative forms and appearances of consumer reviews in terms of felt trust and other
perceptions, marketers are able to find the best solution valued by consumers.

Finally, this thesis also provides some insights for policy makers. For instance, it is widely
acknowledged that online customer reviews are an important form of third-party evaluations.
This reputation instrument relieves consumers from risks associated with online shopping.
Getting the advice from trusted individuals gives guidance and enables individuals to make
well-informed and better shopping choices (Zhou et al., 2013). However, unethical behaviours
are picked out as a central theme by the trade press recently, as some companies are using fake
reviews in order to boost their sales or harm competitors. Steering consumers wrong with
manipulated OCR is an unfair business practice. False online reviews seduce consumers to
make consumption choices they wouldn’t have done if they had known the truth. A variety of
firms have suffered from untruthful review content already and it is very likely that online
shopping in general will be harmed in the long-run if OCR loses its credibility. Accordingly,
various organizations and companies have begun to invest in authentication mechanisms that
aim to ensure the correctness of the claims made in reviews. Researchers support this
endeavour, for instance, by developing new instruments to discover lies in eWOM (e.g.,
Hancock, et al., 2007). By recognizing that untruthful eWOM jeopardizes consumers’
confidence in Internet shopping, national and international public authorities have put the topic
of eWOM trust on their schedule (e.g., EU’s 2012 Consumer Agenda) and have started to
develop new standards, guidelines and programs which all strive to ensure the authenticity of

319
customer reviews and consequently consumer safety on online shopping portals (European
Commission, 2014). The new scale here can make a valuable contribution, as it can be used to
evaluate the effectiveness of such efforts in the long-run. The scale’s practicability allows its
inclusion in longitudinal surveys permitting the capture of the eWOM/online social shopping
dynamics in our consumption society.

Limitations and Directions of Future Research

Even though the thesis at hand has contributed to online customer research by furnishing it with
a conceptually, as well as psychometrically, sound measurement scale to assess individuals’
trust in online customer reviews and recommendations, several important issues can be
identified that are currently beyond the scope of this study but nevertheless make sense to be
investigated by further research.

First, the scale’s development procedure reflects general recommendations of scale

development literature to evaluate its properties with multiple samples. However, further
research may be required to re-test and replicate the scale with additional samples in order to
confirm the scale’s validity, as well as reliability. Such an evaluation would provide meaningful
insights into the stability of the instrument’s psychometric properties across various
populations.

Second, while this research has embedded the construct of eWOM trust within a comprehensive
nomological framework, it has tested only a part of the theorized relationships empirically.
Additionally, there was no further investigation into the simultaneous influences and the
relative importance of various concepts included in this framework. Future research, for
example, could substantially investigate additional antecedents of eWOM trust, such as
personality characteristics and consumer experiences. Research demonstrates that individual
differences, such as need for cognition (Haugtvedt & Petty, 1992), need for closure (Kruglanski
et al., 1993), need to evaluate (Jarvis & Petty, 1995), and preference for consistency (Cialdini
& Trost, 1998), influence a person’s persuasiveness and should be related to trust development.
Moderator variables (e.g., product category) should also be considered (Hogg & Vaughan,
2011). In addition, scholars could tie up with prior research on the investigation of differences
in individuals’ susceptibility to interpersonal influence and trust based on gender and age. Prior
research regularly identifies trust differences attributable to gender (e.g., Foubert & Sholley,
1996). Here, for example, it is claimed that females typically possess a stronger openness to
interpersonal influence than males (Solomon, 1963). The same literature string posits that
females also tend to interact with their peers concerning purchase- and consumption-related
topics and they are generally more susceptible to social influence (e.g., consumption patterns
are strongly influenced by social motivations) than males (Churchill & Moschis, 1979). Results
from persuasiveness research on the gender topic, however, remain mixed. On the other hand,
research was able to demonstrate variations in the influence of social relationships over a
320
person’s lifetime. Here, susceptibility to interpersonal influence seems to have a curvilinear
relationship, with age having the highest impact on behaviours during teenage and young
adulthood. On the other hand, research has also demonstrated that, concerning informational
influence, individuals generally tend to consult others (e.g., family members) as information
sources with advancing age (Phillips & Sternthal, 1977). The theoretical framework also
suggests that eWOM trust is strongly rooted in personality (e.g., extraversion, openness).
Additional research on these and alternative antecedent variables such as a person’s inner-other
directedness (Kassarjian, 1962) or self-monitoring (Lennox & Wolfe, 1984) would be valuable
in contributing to the current discussion about how trust in online reviews is formed.

Third, future research that tests the scale’s applicability beyond its current domain to other
forms of C2C communication (e.g., communication in online brand communities or social
media platforms (e.g., Facebook), expert reviews, online video reviews (e.g., on Youtube) is
desirable. The focal scale was developed in the specific thematic domain to measure trust in
anonymous, text-based customer reviews. On the Internet, however, similar forms of online
market communication exist, but the scale’s generalizability to these regularly used information
sources is still questionable. Additionally, there is a recognizable trend among social media
platforms to furnish consumers with personalized recommendations from known individuals
(e.g., friends, online contacts). The scale has to be assessed with regard to whether it is capable
of capturing the trust construct also in such cases.

Fourth, the eWOM trust scale was developed in two German-speaking countries and was
applied in three different cultures. This furnishes this research with valuable information and
initial evidence for the scale’s stability in other country settings. However, additional cross-
country studies are still necessary. This would help to uncover additional or alternative trust
dimensions attributable to cultural differences. As further assessments of the scale’s English
version are advisable, the same research string could also help to overcome latent limitations
of this research.

Fifth, eWOM trust is here conceptualized by following a trait-based definition and regarding
the construct as a cumulative tendency to rely on information given in online customer reviews
that is relatively stable but changeable. Such a conceptualization includes the assumption that
an individual’s trust is likely to be enduring over a longer time period, but is also subject to
change over time as a result of personal experience. The conceptualization of the construct
emphasized in this thesis parallels the understanding of Sitkin and Weingart (1995), who regard
a personal propensity as a changeable trait which is persistent and enduring; however, the trait
can be “learned or inherited” (corsini & Osaki, 1984; p. 542-543). Hence, this view recognizes
people’s ability to adapt. However, it does not deny that, as individuals gather more experience,
they may be less susceptible to contextual influences and are more likely to show cross-
situational patterns. This said, the present study nevertheless has conceptualized eWOM trust

321
as a relative stable orientation, rather than a context-specific state. Further research is surely
needed to empirically verify this theoretical assumption. In addition, periodical measurement
of the construct is needed in order to assess dynamics which may be attributable to the changing
role of the Internet in humans’ life. Another interesting question is which events in the life of
an online consumer (e.g., eWOM scandals) can alternate his/her relationship to this kind of
online information.

Sixth, while the measurement model applied in this research is conceptually justified, as well
as empirically supported, the evaluation of alternative models should maybe also be
reconsidered using additional data. There is still an open discussion concerning how to
conceptualize trust accordingly (e.g., generalized vs. situational trust). However, this research
recognizes that different types, contexts and objects of trust exist and the construct has to be
conceptualized differently in respect to the demands of the various appearances of the trust
construct. Research endeavours in the same direction could also further test the relationship of
the various trust components and their differential effects on consumer behaviour. The proposed
scale could provide valuable basics for the future development of sub-scales and related
eWOM-scales. For instance, future research should assess whether an adopted version eWT-S
(one that eventually applies an alternative measurement model – e.g., a type II model) is
applicable to quantify situational trust in specific reviews.

Seventh, the new scale should also be useful in experimental research. Here, a consumer’s trust
in reviews can be manipulated in order to investigate effects on review perceptions and
outcomes. Another interesting stream of research could be initiated by studying the determining
effects of dispositional eWOM trust on situational trust. Such an approach would parallel
various contributions in advertising research (e.g., Tan & Chia, 2007). An adopted version of
the scale deemed to be capable of measuring situational eWOM trust could also be used to
compare various types of eWOM messages, such as rational (negative) vs. emotional (positive)
reviews in different product categories and their impact on consumer behaviour.

Eigth, this research has investigated a variety of profiling variables aimed at describing online
consumers as an actionable segment of different forms of market communications. However,
there are still additional characteristics that seem to be critical to understanding the different
segments and which should be emphasized in future research; amongst them, for instance,
opinion leadership (e.g., Childers, 1986), risk perceptions (e.g., Sitkin & Weingart, 1995), self-
efficacy (e.g., Bishop & Barber, 2012; Dunning, 2007), price consciousness (e.g., Lichtenstein
et al., 1993), consumer knowledge (e.g., Park & Kim, 2008; Zou et al., 2011), shopping goal
(e.g., Lee & Ariely, 2006) and gender (e.g., Awad & Ragowsky, 2008; Barber, 2008).
Additionally, further insights into media preferences, channel usage, extent of online search
and consumption patterns would provide interesting insights for marketing managers. This
thesis’ segmentation approach was limited to the US market. Conducting similar research in

322
other countries and contrasting the research at hand with these findings would surely provide
interesting insights into cultural differences. For instance, eWOM trust patterns are likely to
differ between Western cultures (e.g., United States), in which the individualistic self is valued
and Eastern cultures (e.g., Japan, India), whose members are striving for collectivism and
interdependent self and where group loyalty and honesty are more valued (Markus & Kitayama,
1991). By drawing on earlier studies, eWOM trust can be expected to differ across cultures.
This research would tie in already existing cross-cultural eWOM/online research (Fong &
Burton, 2006; Pavlou & Chai, 2002).

Finally, this thesis’ segmentation approach has shown that persons with a high trust in eWOM
only can be contrasted from a group of media literates who possess high trust in online reviews,
as well as online advertising, by means of psychological characteristics, consumption-related
attitudes, behavioural intentions, as well as behaviours. Nevertheless, this research only makes
a first contribution into the investigation of consumers’ willingness to purchase ad- or eWOM-
recommended products, the real consumption of such market offerings, and the role of eWOM
information contrasted from ad claims in purchasing decisions. Hence, future research should
address additional outcome variables, including online consumers’ willingness to pay,
intentions to purchase or to switch a product, and discrete perceptions of market offerings – to
name a few. This research may also investigate the coherent effects between consumers’ online
and offline shopping patterns. Taken together, this should enable research to comprehensively
describe online consumers and their shopping habits.

323
References

Abernethy, A. M. (1991). Differences between advertising and program exposure for car radio
listening. Journal of Advertising Research, 31(April/May), 33-42.
Adler, P. S. (2001). Market, hierarchy and trust: The knowledge economy and the future of
capitalism. Organization Science, 12(2), 215-235.
Ahrholdt, D. C. (2011). Empirical identification of success-enhancing web site signals in e-
tailing: An analysis based on known e-tailers and the theory of reasoned action. Journal
of Marketing Theory and Practice, 19(4), 441-458.
Aiken, K. D., & Bousch, D. M. (2006). Trustmarks, objective-source ratings, and implied
investments in advertising: Investigating online trust and the context-specific nature of
internet signals. Journal of the Academy of Marketing Science, 34, 308-323.
Ajzen, E. (1991). The theory of planned behavior. Organizational Behavior and Human
Decision Processes, 50(2), 179-211.
Ajzen, I. (1988). Attitudes, personality and behaviors. Chicago, IL: Dorsey Press.
Ajzen, I., & Fischbein, M. (1980). Understanding attitudes and predicting social behavior.
Englewood Cliffs, NJ: Prentice-Hall.
Ajzen, I., & Madden, T. J. (1986). Prediciton of goal-directed behavior: Attitudes, intentions,
and perceived behavioral control. Journal of Experimental Social Psychology, 22, 453-
474.
Akaike, H. (1987). Factor analysis and AIC. Psychometrika, 52, 317-332.
Alam, S. S., & Yasin, N. M. (2010). The antecedents of online brand trust: Maylaysian
evidence. Journal of Business Economics and Management, 11(2), 210-226.
Alba, J., Lynch, J., Weitz, B., Janiszewski, C., Lutz, R., & Sawyer, A. (1997). Interactive home
shopping: Consumer, retailer, and manufacturer incentives to participate in electronic
marketplaces. Journal of Marketing, 61, 38-53.
Albers, S., & Hildebrandt, L. (2006). Methodische Probleme bei der Erfolgsfaktorenforschung:
Messfehler, formative versus reflektive Indikatoren und die Wahl des
Strukturgleichungs-Modells. Schmalenbachs Zeitschrift für betriebswirtschaftliche
Forschung, 58(1), 2-33.
Albrecht, S., & Travaglione, A. (2003). Trust in public sector senior management during times
of turbulent change. International Journal of Human Resource Management, 14(2), 1-
17.
Aldenderfer, M. S., & Blashfield, R. K. (1984). Cluster analysis: Sage Publications, Inc. .

325
© Springer Fachmedien Wiesbaden 2017
W. Weitzl, Measuring Electronic Word-of-Mouth Effectiveness,
DOI 10.1007/978-3-658-15889-7
Aldrich, H. E., & Fiol, M. C. (1994). Fools rush in? The institutional context of industry
creation. Academy of Management Review, 19, 645-670.
Allen, V. L. (1965). Situational factors in conformity. In L. Berkowitz (Ed.), Advances in
experimental psychology New York, NY: Academic Press.
Allison, N. K. (1978). A psychometric development of a test for consumer alienation from the
marketplace. Journal of Marketing Research, 15(4), 565-575.
Allport, G. W., & Odbert, H. S. (1936). Trait names: A psycho-lexical study. Psychological
Monographs, 47(211).
Alpern, K. D. (1997). What do we want trust to be? Some distinctions of trust. Business and
Professional Ethics Journal, 16(1-3), 29-46.
Alwin, D. F. (1992). Information transmission in the survey interview: Number of response
categories and the reliability of attitude measurement. Sociological Methodology, 22,
83-118.
Alwin, D. F. (1997). Feeling thermometers versus 7-point scales: Which are better?
Sociological Methods Research, 25(3), 318-340.
Alwitt, L. F., & Prabhaker, P. R. (1994). Identifying who dislikes television advertising: Not
by demographic alone. Journal of Advertising Research, 34(November/December), 17-
29.
Amos, C., Holmes, G., & Strutton, D. E. (2008). Celebrity endorser source effectsand
effectiveness in advertising: A quantiative synthesis. International Journal of
Advertising, 27, 209-234.
Andaleeb, S. S., & Anwar, S. F. (1996). Factors influencing customer trust in salespersons in a
developing country. Journal of International Marketing, 4(4), 35-52.
Andaleeb, S. S., Lee, W., & Gruneisen, D. U. (1992). Building bridges: A model of partnership
between buyers and sellers. Paper presented at the AMA Summer Educators
Conference.
Anderson, E., & Gerbing, D. W. (1988). Structural equation modeling in practice - A review
and recommended two-step approach. Psychological Bulletin, 103(3), 411-423.
Anderson, E., Lodish, L. M., & Weitz, B. A. (1987). Resource allocation behavior in
conventiional channels. Journal of Marketing Research, 24(2), 85-97.
Anderson, E., & Weitz, B. (1989). Determinants of continuity in conventional industrial
channel dyads. Marketing Science, 8(Fall), 310-323.
Anderson, J., & Narus, J. (1990). A model of distributor firm and manufacturer firm working
partnerships. Journal of Marketing, 54(January), 42-58.
Anderson, J. C., & Gerbing, D. W. (1982). Some methods for respecifying measurement models
to obtain unidimensional construct measurement. Journal of Marketing Research, 19,
453-460.
Anderson, J. C., & Gerbing, D. W. (1991). Predicting the performance of measures in a
confirmatory factor analysis with a pretest assessment of their substantive validities.
Journal of Applied Psychology, 76(5), 732-740.
Anderson, M. (2013a). 2013 Study: 79% of consumers trust online reviews as much as personal
recommendations. Retrieved 2014-05-17, from [Link]
study-79-of-consumers-trust-online-reviews-as-much-as-personal-recommendations-
164565
Anderson, M. (2013b). Local consumer review survey 2013.
Anderson, N. H. (1971). Integration theory and attitude change. Psychological Review, 78, 171-
206.
Andreoli, V., & Worchel, S. (1978). Effects of media, communicator, and message position on
attitude change. Public Opinion Quarterly, 42(1), 59-70.
Andrews, J. C. (1989). The dimensionality of beliefs toward advertising in general. Journal of
Advertising, 18(1), 26-35.
326
Archak, N., Ghose, A., & Ipeirotis, P. G. (2011). Deriving the price power of product features
by mining consumer reviews. Management Science, 57(8), 1485-1509.
Arndt, J. (1967). Word of mouth advertising and infomal communication. In D. F. Cox (Ed.),
Risk taking and information handling in consumer behavior (pp. 188-239). Boston, MA:
Harvard University Press.
Arnold, T. J., Landry, T. D., & Reynold, J. K. (2007). Retail online assurances: Typology
development and empirical analysis. Journal of Marketing Theory and Practice, 15(4),
299-313.
Aronson, L. (1985). Attitudes and attitude change. The handbook of social psychology. New
York, NY: Psychology Press.
Awad, N. F., & Ragowsky, A. (2008). Establishing trust in electronic commerce through online
word of mouth: An examination across genders. Journal of Management Information
Systems, 24(4), 101-121.
Ba, S. (2001). Establishing online trust through a community responsibility system. Decision
Support Systems, 31(4), 323-336.
Ba, S., & Pavlou, P. A. (2002). Evidence of the effect of trust building technology in electronic
markets: Price premiums and buyer behavior. MIS Quarterly, 26(3), 243-268.
Backhaus, K., Erichson, B., Plinke, W., & Weiber, R. (2006). Multivariate Analysemethoden
(11 ed.). Berlin, Heidelberg, New York: Springer.
Bae, S., & Lee, T. (2011). Product type and consumers' perception of online consumer reviews.
Electronic Markets, 21, 255-266.
Bagozzi, R. P. (1982). Theoretical concepts, measurements, and meaning. In C. Fornell (Ed.),
A second generation of multivariate analysis. Measurement and evaluation. New York,
NY.
Bagozzi, R. P. (1998). Marketing Management. Upper Saddle River, NJ: Prentice-Hall.
Bagozzi, R. P., & Dholakia, U. M. (1999). Goal setting and goal striving in consumer behavior.
Journal of Marketing, 63(4), 19-32.
Bagozzi, R. P., & Dholakia, U. M. (2002). International social action in virtual communities.
Journal of Interactive Marketing, 16(2), 2-21.
Bagozzi, R. P., & Heatherton, T. F. (1994). A general approach to representing multifaceted
personality constructs: application to sate self-esteem. Structrual Equation Modeling,
1(1), 35-67.
Bagozzi, R. P., Tybout, A. M., Craig, C. S., & Sternthal, B. (1979). The construct validity of
the tripartite classification of attitudes. Journal of Marketing Research, 16(1), 88-95.
Bagozzi, R. P., & Yi, Y. (1988). On the evaluation of structural equation models. Academy of
Marketing Science. Journal, 16, 74-94.
Bagozzi, R. P., & Yi, Y. (1991). Multitrait-multimethod matrices in consumer research. Journal
of Consumer Research, 17(4), 426-439.
Bagozzi, R. P., Yi, Y., & Phillips, L. W. (1991). Assessing construct validity in organizational
research. Administrative Science Quarterly, 36(3), 421-458.
Baier, A. (1986). Trust and antitrust. Ethics, 96, 231-260.
Baier, A. C. (1994). Moral prejudices: Essays on ethics. Harvard University Press. Cambridge,
MA.
Bailey, A. A. (2005). Consumer awareness and use of product review websites. Journal of
Interactive Advertising, 6(1), 68-81.
Bailey, T. (Producer). (2002, 2012-04-26). On trust and philosophy. The philosophy of trust.
[[web page]] Retrieved from [Link]
Bakos, J. Y. (1997). Reducing buyer search costs: Implications for electronic marketplaces.
Management Science, 43, 1676-1692.
Balasubramanian, S. K., & Kamakura, W. A. (1989). Measuring consumer attitudes toward the
marketplace with tailored interviews. Journal of Marketing Research, 26(3), 311-328.
327
Bambauer-Sachse, S., & Mangold, S. (2011). Brand equity diluation through negative online
word-of-mouth communication. Journal of Retailing and Consumer Services, 18, 38-
45.
Bambauer-Sachse, S., & Mangold, S. (2013). Do consumers still believe what is said in online
product reviews? A persuasion knowledge approach. Journal of Retailing and
Consumer Services, 20, 373-381.
Bandura, A. (1977a). Social learning theory. Englewood Cliffs, NJ: Prentice-Hall.
Bandura, A. (1977b). Towards a unifying theory of behavioral change. Psychological Review,
84(2), 191-215.
Bandura, A. (1982). Self-efficacy mechanism in human agency. American Psychologist, 37(2),
122-147.
Bandura, A. (1986). Social foundations of thought and action: A social cognitive theory.
Englewood Cliffs, NJ: Prentice Hall.
Bandura, A. (1991). Social cognitive theory of self-regulation. Organizational Behavior and
Human Decision Processes, 50, 248-287.
Banoma, T. V. (1976). Conflict, cooperation, and trust in three power systems. Behavioral
Science, 21(6), 499-514.
Bansal, A. (1998). Self-efficacy. In V. S. Ramachaudran (Ed.), Encyclopedia of human
behavior (pp. 71-81). New York, NY: Academic Press.
Bansal, H. S., & Voyer, P. A. (2000). Word-of-mouth process within services purchase decision
context. Journal of Service Research, 3(2), 166-177.
Barber, B. (1983). The logic and limits of trust. Brunswick, NJ: Rutgers University Press.
Barber, N. (2008). How self-confidence and knowledge effects the source of information
selected during purchase situations. Texas Tech University, Lubbock, TX.
Barnes, S. J., Bauer, H. H., Neumann, M. M., & Huber, F. (2007). Segmenting cyberspace: A
customer typology for the internet. European Journal of Marketing, 41(1/2), 71-93.
Barney, J. B., & Hansen, M. H. (1994). Trustworthiness as a source of competitive advantage.
Strategic Management Journal, 15, 175-190.
Barrick, M. R., & Mount, M. K. (1991). The big five personality dimensions and job
performance: A meta-analysis. Personnel Psychology, 44, 1-26.
Bartlett, C. A., & Ghoshal, S. (1995). Rebuilding behavioral context: Turn process
reengineering into people rejuvenation. Sloan Management Review, Fall, 11-23.
Basuroy, S., Chatterjee, S., & Ravid, S. A. (2003). How critical are critical reviews? The box
office effects of film critics, star-power, and budget. Journal of Marketing, 67(4), 103-
117.
Bauer, R. A. (1967). Consumer risk as risk taking. In D. F. Cox (Ed.), Risk taking and
information handling in consumer behavior. Boston, MA: Harvard University.
Bauer, R. A., & Greyser, S. A. (1968). Advertising in America: The consumer view. Boston,
MA: Harvard University, Graduate School of Business Administration, Division of
Research.
Baumgartner, H., & Homburg, C. (1996). Applications of structural equation modeling in
marketing and consumer research: A review. International Journal of Research in
Marketing, 13(2), 139-161.
Baxter, R. (2009). Reflective and formative metrics of relationship value: A commentary essay.
Journal of Business Research, 62(12), 1370-1377.
Bearden, W., Hardesty, D. M., & Rose, R. L. (2001). Consumer self-confidence: Refinements
in conceptualization and measurement. Journal of Consumer Research, 28(1), 121-134.
Bearden, W. O., & Etzel, M. J. (1982). Reference group influence on product and brand
purchase decisions. Journal of Consumer Research, 9(9), 183-194.
Bearden, W. O., & Netemeyer, R. G. (1998). Handbook of marketing scales: Multi-item
measures for marketing and consumer behavior research. Thousand Oaks, CA: Sage.
328
Bearden, W. O., Netemeyer, R. G., & Teel, J. E. (1989). Measurement of consumer
susceptibility to interpersonal influence. Journal of Consumer Research, 15(4), 473-
481.
Bearden, W. O., Netemeyer, R. G., & Teel, J. E. (1990). Further validation of the consumer
susceptibility to interpersonal influence scale In M. E. Goldberg (Ed.), Advances in
Consumer Research (Vol. 17, pp. 770-776). Provo, UT: Association for Consumer
Research.
Bearden, W. O., & Rose, R. L. (1990). Attention to social comparison information: An
individual difference factor affecting consumer conformity. Journal of Consumer
Research, 16(March), 461-471.
Becerra, E. P. (2006). Influence of consumers' trust beliefs on intentions to transact in the online
environment: An e-tail study. Flordia Atlantic University, Fort Lauderdale, FL.
Becerra, E. P., & Korgaonkar, P. (2011). Effects of trust beliefs on consumers' online intentions.
European Journal of Marketing, 45(6), 936-962.
Bei, L. T., Chen, E. Y. I., & Widdows, R. (2004). Consumers' online information search
behavior and the pheonomenon of search vs. experience products. Journal of Family
and Economic Issues, 25(4), 449-467.
Beldad, A., De Jong, M., & Steehouder, M. (2010). How shall I trust the faceless and the
intangible? A literature review on the antecedents of online trust. Computers in Human
Behavior, 26, 857-869.
Beltramini, R. F., & Evans, K. R. (1985). Perceived believability ofresearch results information
in advertising. Journal of Advertising, 14(3), 18-31.
Benamati, J., Serva, M. A., & Fuller, M. A. (2006). Are trust and distrust distinct constructs?
An empirical study of the effects of trust and distrust among online banking users. Paper
presented at the Proceedings of the 39th Hawaii International Conference on System
Sciences.
Benedicktus, R. L. (2011). The effects of 3rd party consensus information on service
expectations and online trust. Journal of Business Research, 64(8), 846-853.
Benedicktus, R. L., Brady, M. K., Darke, P. R., & Voorhees, C. M. (2010). Conveying
trustworthiness to online consumers: Reactions to consensus, physical store presence,
brand familirity, and generalized suspicion. Journal of Retailing, 86(4), 322-335.
Bente, G., Baptist, O., & Leuschner, H. (2012). To buy or not to buy: Influence of seller photos
and reputation on buyer trust and purchase behavior. International Journal of Human-
Computer Studies, 70(1), 1-13.
Bentler, P. M. (1990). Comparative fit indexes in structural models. Psychological Bulletin,
107(2), 238-246.
Bentler, P. M. (1995). EQS: Structural equations program manual. Encino, CA: Multivariate
Software.
Bentler, P. M., & Bonett, D. G. (1980). Significance tests and goodness of fit in the analysis of
covariance structures. Psychological Bulletin, 88, 588-606.
Bentler, P. M., & Chou, C.-P. (1987). Practical issues in structural modeling. Sociological
Methods & Research, 16, 78-117.
Berger, J., & Milkman, K. L. (2012). What makes online content viral? Journal of Marketing
Research, 49(2), 192-205.
Berry, L. L. (1995). Relationship marketing of services: Growing interest, emerging
perspectives. Journal of the Academy of Marketing Science, 23(4), 236-245.
Betra, R., & Ray, M. (1985). Affective response mediating acceptance of advertising. Journal
of Consumer Research, 3, 234-249.
Bettman, J. R. (1973). Perceived risk and its components: A model and empirical test. Journal
of Marketing Research, 10, 184-190.

329
Bezjian-Avery, A., Calder, B. J., & Iacobucci, D. (1998). New media interactive advertising vs.
tradtional advertising. Journal of Advertising Research, 38(4), 23-32.
Bhatnagar, A., & Ghose, A. (2004). Segmenting consumers based on the benefits and risks of
Internet shopping. Journal of Business Research, 57, 1352-1360.
Bhattacharya, R., & Devinney, T. M. (1998). A formal model of trust based on outcomes.
Academy of Management Review, 23(3), 459-474.
Bhattacherjee, A. (2002). Individual trust in online firms: Scale development and initial test.
Journal of Management Information Systems, 19(1), 211-241.
Bickart, B., & Schindler, R. M. (2001). Internet forums as influential sources of consumer
information. Journal of Interactive Marketing, 15(3), 31-40.
Bigley, G., & Pearce, J. (1998). Straining for shared meanings in organization science:
Problems of trust and distrust. Academy of Management Review, 23(3), 405-421.
Bishop, M., & Barber, N. (2012). A market segmentation approach to esteem and efficacy in
information search. Journal of Consumer Marketing, 29(1), 13-21.
Bitkom. (2013). Online-Shopping in Deutschland besonders beliebt [Press release]. Retrieved
from [Link]
Blakeney, R. N. (1986). A transactional view of the role of trust in organizational
communication. Transactional Analysis Journal, 16(1), 95-98.
Blalock, H. M. (1964). Causal inferences in nonexperimental research. Chapel Hill, NC:
University of North Carolina Press.
Blalock, H. M. (1969). Multiple indicators and the causal approach to measurement error.
American Journal of Sociology, 75, 264-272.
Blau, P. (1964). Exchange and power in social life. New York, NY: John Wiley & Sons, Inc.
Blois, K. J. (1999). Trust in business to business relationships: an evaluation of its status.
Journal of Management Studies, 36(2), 197-215.
Bluhm, L. H. (1987). Trust, terrorism and technology. Journal of Business Ethics, 6, 333-342.
Bok, S. (1978). Lying: Moral choice in public and private life. New York, NY: Pantheon Books.
Bollen, K. (1984). Multiple indicators: Internal consistency or no necessary relationship?
Quality and Quantity, 18(4), 377-385.
Bollen, K. (1989). Structural equations with latent variables. New York, NY: John Wiley &
Sons.
Bollen, K. (2011). Evaluating effect, composite, and causal indicators in structural equation
models. MIS Quarterly, 35(2), 359-372.
Bollen, K., & Lennox, R. (1991). Conventional wisdom on measurement: A structural equation
perspective. Psychological Bulletin, 110(2), 305-314.
Bolton, R. N., Lemon, K. N., & Verhoef, P. C. (2004). The theoretical underpinnings of
customer asset management: A framework and propositions for future research. Journal
of the Academy of Marketing Science, 32(3), 271-292.
Bolton, R. N., Parasuraman, A., & Hoefnagels, A. (2013). Understanding generation Y and
their use of social media: A review and research agenda. Journal of Service Marketing,
24(3), 245-267.
Boon, S. D., & Holmes, J. G. (1991). The dynamics of interpersonal trust: Resolving uncertainty
in the face of risk. In R. A. Hinde & J. Groebel (Eds.), Cooperation and prosocial
behavior (pp. 190-211). New York, NY: Cambridge University Press.
Boush, D. M., Friestad, M., & Rose, G. M. (1994). Adolescent skepticism toward TV
advertising and knowledge of advertiser tactics. Journal of Consumer Research, 21(1),
165-175.
Boush, D. M., Kim, C.-H., Kahle, L. R., & Batra, R. (1993). Cynicism and conformity as
correlates of trust in product information sources. Journal of Current Issues and
Research in Advertising, 15(2), 71-79.

330
Bowden, J. L. H. (2009a). Customer engagement: A framework for assessing customer-brand
relationships: The case of the restaurant industry. Journal of Hospitality Marketing and
Management, 18(6), 574-596.
Bowden, J. L. H. (2009b). The process of customer engagement: A conceptual framework.
Journal of Marketing Theory and Practice, 17(1), 63-74.
Bowers, K. S. (1973). Situationism in psychology - Analysis and a critique. Psychological
Review, 80, 307-336.
Bowlby, J. (1973). Attachment and loss Anxiety and anger. London, UK: Hogarth Press.
Bowlby, J. (1982). Attachment and loss, Volume I: Attachment. New York, NY: Basic Books.
Brace, I. (2004). Questionnaire design. How to plan, structure and write survey material for
effective market research. London, UK: Kogan Page.
Bradach, J. L., & Eccles, R. G. (1989). Price, authority, and trust: From ideal types to plural
forms. In W. R. Scott & J. Blake (Eds.), Annual Review of Sociology (pp. 97-118).
Bramall, C., Schoefer, K., & McKechnie, S. (2004). The determinants and consequences of
consumer trust in e-retailing: A conceptual framework. International Journal of Internet
Marketing and Advertising, 1(4), 388-412.
Braunsberger, K., & Munch, J. M. (1998). Source expertise versus experience effects in hospital
advertising. Journal of Services Marketing, 12(1), 23-38.
Bray, J., & Schetzina, C. (2006). Travel 2.0: Harnessing the power of user-generated content
and tagging. Hospitality Upgrade, Fall, 28-29.
Brehm, S. S. (1992). Intimate relationships. New York, NY: McGraw-Hill.
Brengman, M., Geuens, M., Weijters, B., Smith, S. M., & Swinyard, W. R. (2005). Segmenting
Internet shoppers based on their web-usage-related lifestyle: A cross-cultural validation.
Journal of Business Research, 58, 79-88.
Brettel, M., & Spilker-Attig, A. (2010). Online advertising effectiveness: A cross-cultural
comparison. Journal of Research in Interactive Marketing, 4(3), 176-196.
Breuer, R., & Brettel, M. (2012). Short- and long-term effects of online advertising: Differences
between new and existing customers. Journal of Interactive Marketing, 26, 155-166.
Brewer, P. (2007). Operationalizing psychic distance: A revised approach. Journal of
International Marketing, 15(1), 44-66.
Briggs, P., Burford, B., Angeli, A., & Lynch, P. (2002). Trust in online advice. Social Science
Computer Review, 20(3), 321-332.
BrightLocal. (2013). Local consumer review survey 2013. Retrieved 2014-05-31, from
[Link]
Bristor, J. M. (1990). Enhanced explanations of word of mouth communications: The power of
relationships. Research in Consumer Behavior, 4, 51-83.
Brock, T. C. (1965). Communicator-recipient similarity and decision change. Journal of
Personality and Social Psychology, 1, 650-654.
Brockhaus, R. H. (1980). Risk-taking propensity of entrepreneurs. Academy of Management
Journal, 23, 509-520.
Brodie, R. J., Ilic, A., Juric, B., & Hollebeek, L. (2013). Consumer engagement in a virtual
brand community: An exploratory analysis. Journal of Business Research, 66, 105-114.
Bromiley, P., & Cummings, L. L. (1992). Transaction costs in organizations with trust.
Working paper, No. 28. University of Minnesota, Strategic Management Research
Center. Minneapolis.
Bromiley, P., & Curley, S. P. (1992). Individual differences in risk taking. In J. F. Yates (Ed.),
Risk taking behavior (pp. 87-132). Chichester, UK: Wiley.
Bronner, F., & Hoog, R. d. (2010). Consumer-generated versus marketer-generated websites in
consumer decision making. International Journal of Market Research, 52(2), 231-248.

331
Brown, J., Broderick, A. J., & Lee, N. (2007). Word of mouth communication within online
communities: Conceptualizing the online social network. Journal of Interactive
Marketing, 21(3), 2-20.
Brown, J. J., & Reingen, P. H. (1987). Social ties and word-of-mouth referral behavior. Journal
of Consumer Research, 14, 350-362.
Browne, G. J., Pitts, M. G., & Wetherbe, J. C. (2007). Cognitive stopping rules for terminating
information search in online tasks. MIS Quarterly, 31(1), 89-104.
Browne, M. W., & Cudeck, R. (1993). Alterantive ways of assessing model fit. In K. A. Bollen
& J. S. Long (Eds.), Testing structural equation models (pp. 136-162). Newbury Park,
CA: Sage.
Bucklin, R. E., Lattin, J. M., Ansari, A., Gupta, S., Bell, D., Coupey, E., Steckel, J. (2002).
Choice on the Internet: From clickstream to research stream. Marketing Letters, 13(3),
245-258.
Burgess, S., Sellitto, C., Cox, C., & Buultjens, J. (2011). Trust perceptions of online travel
information by different content creators: Some social and legal implications.
Information Systems Front, 13, 221-235.
Burke, P. J., & Stets, J. e. (1999). Trust and commitment through self-verification. Social
Psychology Quarterly, 62(4), 347-360.
Burnkrant, R. E., & Cousineau, A. (1975). Informational and normative social influence in
buyer behavior. Journal of Consumer Research, 2(12), 206-215.
Burns, D. J. (2010). Consumer alienation and attitudes toward consumer free riding. The
Journal of Business Inquiry, 9(1), 22-36.
Busch, P., & Wilson, D. T. (1976). An experimental analysis of a salesman's expert and referent
bases of social power in the buyer-seller dyad. Journal of Marketing Research, 23(2),
3-11.
Businessweek. (2009). Amazon: Turning consumer opinions into gold. Retrieved from
[Link]
Butler, J. K. (1986). Reciprocity of dyadic trust in close male-female relationships. The Journal
of Social Psychology, 126(5), 579-591.
Butler, J. K. (1991). Toward understanding and measuring conditions of trust: Evoluation of a
conditions of trust inventory. Journal of Management, 17, 643-663.
Butler, J. K., & Cantrell, R. S. (1984). A behavioral decision theory approach to modeling
dyadic trust in superiors and subordinates. Psychological Reports, 55, 19-28.
Büttner, O. B., & Göritz, A. S. (2008). Perceived trustworthiness of online shops. Journal of
Consumer Behaviour, 7(1), 35-50.
Byrne, B. M. (1994). Structural equation modeling with EQS and EQS/Windows. Thousand
Oaks, CA: Sage Publications.
Cadogan, J. W., Diamantopoulos, A., & Pahud de Mortanges, C. (1999). A measure of export
market orientation: Scale development and cross-cultural validation. Journal of
International Business Studies, 30, 689-707.
Calder, B. J. (1977). An attribution theory of leadership. In B. M. Stwa & G. R. Salancik (Eds.),
New directions in organizational behavior. Chicago, IL: St. Clair.
Calder, B. J., Phillips, L. W., & Tybout, A. M. (1982). The concept of external validity. Journal
of Consumer Research, 9(December), 240-244.
Calfee, J. E., & Ringold, D. J. (1994). The 70% majority: Enduring consumer beliefs about
advertising. Journal of Public Policy & Marketing, 13(2), 228-238.
Campbell, D. T. (1960). Recommednations for APA test standards regarding construct, trait, or
discriminant validity. American Psychologist, 15, 546-553.
Campbell, D. T., & Fiske, D. W. (1959). Convergent and discriminant validity by the multitrait-
multimethod matrix. Psychological Bulletin, 56, 81-105.

332
Casalo, L. V., Flavian, C., & Guinaliu, M. (2007). The impact of participation in virtual brand
communities on consumer trust and loyalty. Online Information Review, 31(6), 775-
792.
Casalo, L. V., Flavian, C., & Guinaliu, M. (2008). Fundamentals of trust management in the
development of virtual communities. Management Research News, 31(5), 324-338.
Casalo, L. V., Flavian, C., & Guinaliu, M. (2011). Understanding the intention to follow the
advice obtained in an online travel community. Computers in Human Behavior, 27, 622-
633.
Cattell, R. B. (1966). The screen test for the number of factors. Multivariate Behavioral
Rresearch, 1, 245-276.
Cattell, R. B. (1978). The scientific use of factor analysis. New York, NY: Plenum Press.
Chaiken, S. (1980). Heuristic versus systematic information processing and the use of source
versus message cues in persuasion. Journal of Personality and Social Psychology,
39(5), 752-766.
Chaloupka, W. (1999). Everybody knows: Cynicism in America. Minneapolis: University of
Minnesota Press.
Chan, K. W., & Li, S. Y. (2010). Understanding consumer-to-consumer interactions in virtual
communities: The salience of reciprocity. Journal of Business Research, 63, 1033-1040.
Chang, H. H., & Chen, S. W. (2008). The impact of online store environment cues on purchase
intention. Online Information Review, 32(6), 818-841.
Chang, L., Lee, Y.-J., & Huang, C.-L. (2008). The influence of e-word-of-mouth on the
consumer's purchase decision: A case of body care products. Working Paper. Takming
University.
Chau, P. Y. K., Hu, P. J.-H., Lee, B. L. P., & Au, A. K. K. (2007). Examining customers' trust
in online vendors and their dropout decisions: An empirical study. Electronic
Commerce Research and Applications, 6(2), 171-182.
Chaudhuri, A., & Holbrook, M. B. (2001). The chain of effects from brand trust and brand
affect to brand performance: The role of brand loyalty. Journal of Marketing, 65(2), 81-
93.
Cheema, A. (2008). Surcharges and seller reputation. Journal of Consumer Research, 35, 167-
177.
Cheema, A., & Papatla, P. (2010). Relative importance of online versus offline information for
Internet purchases: Product category and Internet experience effects. Journal of
Business Research, 63, 979-985.
Chen, C.-J., & Hung, S.-W. (2010). To give or to receive? Factors influencing members'
knowledge sharing and community promotion in professional virtual communities.
Information & Management, 47, 226-236.
Chen, J., & Dibb, S. (2010). Consumer trust in the online retail context: Exploring the
antecedents and consequences. Psychology & Marketing, 27(4), 323-346.
Chen, P., Dhanasobhon, S., & Smith, M. (2008). All reviews are not created equal: The
disaggregate impact of reviews on sales on [Link]. Working paper. Temple
University. Retrieved from [Link]
abstract_id=918083
Chen, Q., Griffith, D. A., & Wang, F. (2005). The behavioral implications of consumer trust
across bric-and-mortar and online retail channels. Journal of Marketing Channels,
11(4), 61-87.
Chen, S. C., & Dhillon, G. S. (2003). Interpreting dimensions of consumer trust in e-commerce.
Information Technology and Management, 4, 303-318.
Chen, Y.-H., & Barnes, S. (2007). Initial trust and online buyer behavior. Journal of Industrial
Management & Data Systems, 107(1), 21-36.

333
Cheong, H. J., & Morrison, M. A. (2008). Consumers' reliance on product information and
recommendatiion found in UGC. Journal of Interactive Advertising, 8(2), 38-49.
Cheung, C. M. K., & Lee, M. K. O. (2001). Trust in Internet shopping: Instrument development
and validation through classical and modern approaches. Journal of Global Information
Management, 9, 23-35.
Cheung, C. M. K., & Lee, M. K. O. (2012). What drives consumers to spread electronic word
of mouth in online consumer-opinion platforms. Decision Support Systems, 53, 218-
225.
Cheung, C. M. K., & Lee, M. K. O. (n.a.). An integrative model of consumer trust in internet
shopping. Hong Kong.
Cheung, C. M. K., Lee, M. K. O., & Rabjohn, N. (2008). The impact of electronic word-of-
mouth: The adoption of online opinions in online customer communities. Internet
Research, 18(3), 229.
Cheung, C. M. K., & Thadani, D. R. (2012). The impact of electronic word-of-mouth
communication: A literature analysis and integrated model. Decision Support Systems,
forthcoming.
Cheung, M., Luo, C., Sia, C., & Chen, H. (2009). Credibility of electronic word-of-mouth:
Informational and normative determinants of on-line consumer recommendations.
International Journal of Electronic Commerce, 13(4), 9.
Cheung, M. K., & Lee, M. K. O. (2006). Understanding consumer trust in Internet shopping: A
multidisciplinary approach. Journal of the American Society for Information Science
and Technology, 22, 479-492.
Chevalier, J. A., & Mayzlin, D. (2003). The effect of word of mouth sales: Online book reviews.
Yale SOM Working Paper Series.
Chevalier, J. A., & Mayzlin, D. (2006). The effect of word of mouth online: Online book
reviews. Journal of Marketing Research, 43(3), 345-354.
Childers, T. L. (1986). Assessment of the psychometric properties of an opinion leadership
scale. Journal of Marketing Research, 23(2), 184-188.
Childers, T. L., & Rao, A. R. (1992). The influence of familial and peer-based reference groups
on consumer decisions. Journal of Consumer Research, 19(2), 198-211.
Chiles, T. H., & McMackin, J. F. (1996). Integrating variable risk preference, trust, and
transaction cost economics. Academy of Management Review, 21, 73-99.
Chiou, J.-S. (2004). The antecedents of consumers' loyalty toward Internet service providers.
Information & Management, 41, 685-695.
Chiu, C. M., Hsu, M. H., & Wang, T. G. (2006). Understanding knowledge sharing in virtual
communities: An integration of social capital and social cognitive theories. Decision
Support Systems, 42, 1872-1888.
Cho, J., & Lee, J. (2006). An integral model of risk-reducing strategies. Journal of Business
Research, 59(1), 112-120.
Choi, S. M., & Rifon, N. J. (2002). Antecedents and consequences of web advertising
credibility: A study of consumer response to banner advertisements. Journal of
Interactive Advertising, 3(1), 12-24.
Chu, S.-C., & Kim, Y. (2011). Determinants of consumer engagement in electronic word-of-
mouth (eWOM) in social networking sites. International Journal of Advertising, 30(1),
47-75.
Chung, J. (2011). Investigating the roles of online buzz for new product diffusion and its cross-
country dynamics. Journal of Business Research, 64(11), 1183-1189.
Churchill, G. A. (1979). A paradigm for developing better measures of marketing constructs.
Journal of Marketing Research, 16(1), 64-73.
Churchill, G. A., & Iacobucci, D. (2005). Marketing research - Methodological foundations (9
ed.). Mason, OH: Thomson South-Western.
334
Churchill, G. A., & Moschis, G. P. (1979). Television and interpersonal influence on adolescent
consumer learning. Journal of Consumer Research, 6(June), 23-35.
Chylinski, M., & Chu, A. (2010). Consumer cynicism: Antecedents and consequences.
European Journal of Marketing, 44(6), 796-837.
Cialdini, R. B., & Trost, M. R. (1998). Social influence: Social norms, conformity, and
compliance. In D. T. Gilbert, S. T. Fiske & G. Lindzey (Eds.), The handbook of social
psychology (4 ed., Vol. 2, pp. 151-192). New York, NY: McGraw-Hill.
Citera, M., Beauregard, R., & Mitsuya, T. (2005). An experimental study of credibility in e-
negotiations. Psychology & Marketing, 22(2), 163-179.
Clark, B. H. (1999). Marketing performance measures: History and inter-relationships. Journal
of Marketing Management, 15(8), 711-732.
Clark, L. A., & Watson, D. (1995). Constructing validity: Basic issues in scale development.
Psychological Assessment, 7(3), 309-319.
Clark, M. C., & Payne, R. L. (1997). The nature and structure of workers' trust in management.
Journal of Organizational Behavior, 18, 802-824.
Clemons, e., Gao, G., & Hitt, L. (2006). When online reviews meet hyperdifferentiation: A
study of the craft beer industry. Journal of Management Information Systems, 23(2),
149-171.
Cliff, N. (1988). The eigenvalue-greater-than-one rules and reliability of components.
Psychological Bulletin, 103(2), 276-279.
Cohen, J. B. (1992). A power primer. Psychological Bulletin, 112(1), 155-159.
Cohen, J. B., & Golden, E. (1972). Informational social influence and product evaluation.
Journal of Applied Psychology, 56(1), 54-59.
Coleman, J. (2000). Foundations of social theory. Cambridge, MA: The Belknap Press of
Harvard University Press.
Coleman, J. S. (1990). Foundations of social theory. Cambridge, MA: Harvard University
Press.
Colquitt, J. A., Scott, B. A., & LePine, J. A. (2007). Trust, trustworthiness, and trust propensity:
A meta-analytic test of their unqiue relationships with risk taking and job performance.
Journal of Applied Psychology, 92(4), 909-927.
Comrey, A. (1970). Manual of the Comrey personality scales. San Diego, CA: EdiTs.
Comrey, A. L., & Lee, H. B. (2002). A first course in factor analysis. Hillsdale, NJ: Larence
Erlbaum.
ComScore. (2007). comScore releases worldwide rankings of top web properties. Retrieved
06.08.2013, 2013, from [Link]
Top_Worldwide_Web_Properties
Cone. (2011). Game changer: Cone survey finds 4-out-of-5 consumers reverse purchase
decisions based on negative online reviews [Press release]. Retrieved from
[Link]
3bb2/files/2011_cone_online_influence_trend_tracker_release_and_fact_sheet.pdf
Connolly, R., & Bannister, F. (2007). Consumer trust in Internet shopping in Ireland: Towards
the development of a more effective trust measurement instrument. Journal of
Information Technology, 22, 102-118.
Cook, J., & Wall, T. (1980). New work attitude measures of trust, organizational commitment
and personal need non-fulfillment. Journal of Occupational Psychology, 53(1), 39-52.
Cook, T. D., & Campbell, D. T. (1979). Quasi-experimentation: Design and analysis issues for
field settings. Boston, MA: Houghton Mifflin.
Coppola, N., Hiltz, S. R., & Rotter, N. (2004). Building trust in virtual teams. IEEE
Transactions on Professional Communication, 47(2), 95-104.
Corbitt, B. J., Thanasankit, T., & Yi, H. (2003). Trust and e-commerce: A study of consumer
perceptions. Electronic Commerce Reserch and Applications, 2, 203-215.
335
Corritore, C. L., Kracher, B., & Wiedenbeck, S. (2001). Trust in the online enviornment. In M.
J. Smith, G. Salvendry, D. Harris & R. J. Koubek (Eds.), Usability evaluation and
interface design: Cognitive engineering, intelligent agents and virtual reality (pp. 1548-
1552). Mahway, NJ: Erlbaum.
Corritore, C. L., Kracher, B., & Wiedenbeck, S. (2003). On-line trust: Concepts, evolving
themes, a model. International Journal of Human-Computer Studies, 58, 737-758.
Corsini, R. J., & Osaki, B. D. (1984). Encyclopedia of psychology. New York, NY: Wiley.
Cortina, J. M. (1993). What is coefficient alpha? An examination of theory and applications.
Journal of Applied Psychology, 78, 98-104.
Costa, P. T., & McCrae, R. R. (1985). The NEO Personality Inventory: Manual form S and
form R. Odessa, FL: Psychological Assessment Resources.
Costa, P. T., & McCrae, R. R. (1989). The NEO-PI/NEO-EFI manual supplement. In J. P.
Cuieford (Ed.), Fundamental statistics in psychology and education. New York, NY:
McGraw-Hill.
Costa, P. T., & McCrae, R. R. (1992a). The NEO PI-R professional manual. Odessa, FL:
Psychological Assessment Resources.
Costa, P. T., & McCrae, R. R. (1992b). Revised NEO personality inventory and new five-factor
inventory: Pofessional manual. Odessa: Psychological Assessment Resources.
Costa, P. T., & McCrae, R. R. (1995). Domains and facets: Hierarchical personality assessment
using the revised NEO personality inventory. Journal of Personality Assessment, 64,
21-50.
Covello, V. T. (1992). Trust and credibility in risk communication. Health Environment Digest,
6(1), 1-4.
Cox, D. E. (1962). Information & uncertainty: Their effects on consumer product evaluations.
Harvard University, Boston, MA.
Cox, D. F., & Bauer, R. A. (1964). Self-confidence and persuasibility in women. Public
Opinion Quarterly, 28(3), 453-466.
Cox, D. F., & Rich, S. U. (1964). Perceived risk and consumer decision-making - The case of
telephone shopping. Journal of Marketing Research(November), 32-39.
Creed, W. E. D., & Miles, R. E. (1996). Trust in organizations: A conceptual framework linking
organizational forms, managerial philosophies, and the opportunity costs of control. In
R. M. Kramer & T. R. Tyler (Eds.), Trust in organizations: Frontiers of theory and
research (pp. 16-38). London, England: Sage Publications.
Cronbach, L. J. (1951). Coefficient alpha and the internal structure of tests. Psychometrica,
16(297-334).
Cronbach, L. J., & Meehl, P. E. (1955). Construct validity for psychological tests.
Psychological Bulletin, 52, 281-302.
Cronin, J. J., & Menelly, N. E. (1992). Discrimination vs. avoidance: Zipping of television
commercials. Journal of Advertising, 21(Summer), 1-7.
Crosby, L. A., Evans, K. R., & Crowles, D. (1990). Relationship quality in services selling: An
interpersonal influence perspective. Journal of Marketing, 54(3), 77-97.
Crowne, D. P., & Marlowe, D. (1960). A new scale for social desirability independent of
psychopathology. Journal of Consulting Psychology, 24(4), 349-354.
Cummings, L. L., & Bromiley, P. (1996). The organizational trust inventory (OTI):
Development and validation. In K. M. R. Tyler & T. R. Tyler (Eds.), Trust in
organization (pp. 302-330). Thousand Oaks, CA: Sage.
Curtis, R. F., & Jackson, E. F. (1962). Multiple indicators in survey research. American Journal
of Sociology, 68(2), 195-204.
Darley, J. M., & Latané, B. (1968). Bystander intervention in emergencies: Diffusion of
responsibility Journal of Personality and Social Psychology, 8, 377-383.

336
Darley, W. K., & Smith, R. E. (1993). Advertising claim objectivity: Antecedents and effects.
Journal of Marketing, 57, 100-113.
Darley, W. K., & Smith, R. E. (1995). Gender differences in information processing strategies:
An empirical test of the selective model in advertising response. Journal of Advertising,
24(1), 41-57.
Das, T. K., & Teng, B. S. (1998). Between trust and control: Developing confidence in partner
cooperation in alliances. The Academy of Management Review, 23(3), 491-513.
Das, T. K., & Teng, B. S. (2004). The risk-based view of trust: A conceptual framework.
Journal of Business and Psychology, 19(1), 85-116.
Dasgupta, P. (1988). Trust as a commodity. In D. Gambetta (Ed.), Trust making and breaking
cooperative relations (pp. 49-52). New York, NY: Blackwell.
Davenport, T. H., & Prusak, L. (1998). Working knowledge: How organizations manage what
they know. Boston, MA: Harvard Business School Press.
Davis, A., & Khazanchi, D. (2008). An empirical study of online word of mouth as a predictor
for multi-product category e-commerce sales. Electronic Markets, 18(2), 130-141.
Dawes, J. G. (2008). Do data characteristics change according to the number of scale points
used? An experiment using 5 point, 7 point and 10 point scales. International Journal
of Market Research, 51(1), 61-77.
De Bruyn, A., & Lilien, G. L. (2008). A multi-stage model of word-of-mouth influence.
International Journal of Research in Marketing, 25(September), 151-163.
De Valck, K., Van Bruggen, G. H., & Wierenga, B. (2009). Virtual communities: A marketing
perspective. Decision Support Systems, 47(3), 185-203.
Dean, J. W., Brandes, P. M., & Dhwardkar, R. (1998). Organizational cynicism. Academy of
Management Review, 23, 341-352.
Delgado-Ballester, E. (2004). Applicability of a brand trust scale across product categories: A
multigroup invariance analysis. European Journal of Marketing, 38(5-6), 592-593.
Delgado-Ballester, E., & Munuera-Aleman, J. L. (2001). Brand trust in the context of consumer
loyalty. European Journal of Marketing, 35(11-12), 1238-1258.
Delgado, B., E., Munuera-Aleman, J. L., & Yague-Guillen, M. J. (2003). Development and
validation of a brand trust scale. International Journal of Market Research, 45, 35-76.
Dellarocas, C. (2003). The digitalization of word of mouth: Promise and challenges of online
feedback mechanisms. Management Science, 29(10), 1407-1424.
Dellarocas, C., Zhang, X. M., & Awad, N. F. (2007). Exploring the value of online product
reviews in forecasting sales: The case of motion pictures. Journal of Interactive
Marketing, 21(4), 23-48.
Deloitte (Producer). (2007, 2012-08-09). New Deloitte study shows inflection point for
consumer products industry; Companies must learn to compete in a more transparent
age. Retrieved from [Link]
shows-inflection-point-for-consumer-products-industry-companies-must-learn-to-
[Link]
Deutsch, M. (1958). Trust and suspicion. Conflict Resolution, 2(4), 265-279.
Deutsch, M. (1960). The effect of motivational orientation upon trust and suspicion. Human
Relations, 13(2), 123-139.
Deutsch, M. (1962). Cooperation and trust: Some theoretical notes. Nebraska Symposium on
Motivation, 10, 275-318.
Deutsch, M. (1973). The resolution of conflict: Constructive and destructive processes. New
Haven, CN: Yale University Press.
Deutsch, M., & Gerard, H. B. (1955). A study of normative and informational influence upon
individual judgement. Journal of Abnormal Social Psychology, 51(11), 629-636.
DeVellis, R. F. (1991). Scale development: Theory and applications. Newbury Park, CA: Sage.

337
DeVellis, R. F. (2012). Scale development: Theory and applications. Thousand Oaks, CA:
Sage.
Dhar, V., & Chang, E. A. (2009). Does chatter matter? The impact of user-generated content
on music sales. Journal of Interactive Marketing, 23(4), 300-307.
Dholakia, R. R., & Sternthal, B. (1977). Highly credible sources: Persuasive facilitators or
persuasive liabilities. Journal of Consumer Research, 3(March), 223-232.
Diamantopoulos, A. (2011). Incorporating formative measures into covariance-based structural
equation models. MIS Quarterly, 35(2), 335-358.
Diamantopoulos, A., Riefler, P., & Roth, K. P. (2008). Advancing formative measurement
models. Journal of Business Research, 61(12), 1203-1218.
Diamantopoulos, A., & Siguaw, J. A. (2000). Introducing LISREL: Guide for the uninitiated.
London, UK: Sage Publications.
Diamantopoulos, A., & Siguaw, J. A. (2006). Formative versus reflective indicators in
organizational measure development: A comparison and empirical illustration. British
Journal of Management, 17(4), 263-282.
Diamantopoulos, A., & Winklhofer, H. (2001). Index construction with formative indicators:
An alternative to scale development. Journal of Marketing Research, 38(2), 269-277.
Dichter, E. (1966). How word-of-mouth advertising works. Harvard Business Review,
44(11/12), 147-166.
Diehm, R., & Armatas, C. (2004). Surfing: An avenue for socially acceptable risk-taking,
satisfying needs for sensation seeking and experience seeking. Personality and
Individual Differences, 36, 63-677.
Digman, J. M., & Inouye, J. (1986). Further specification of the five robust factors of
personality. Journal of Personality and Social Psychology, 50(1), 116-123.
Dimitriadis, S., Kouremenos, A., & Kyrezis, N. (2011). Trust-based segmentation: Preliminary
evidence from technology-enabled bank channels. International Journal of Bank
Marketing, 29(1), 5-31.
Discoll, J. W. (1978). Trust and participation in organizational decision making as predictors
of satisfaction. Academy of Management Journal, 21, 44-56.
Dodgson, M. (1993). Learning, trust, and technological collaboration. Human Relations, 46,
77-95.
Doll, W. J., Hendrickson, A., & Xiaodong, D. (1998). Using Davis's perceived usefulness and
ease-of-use instruments for decision making: A confirmatory and multigroup invariance
analysis. Decision Sciences, 29(4), 839-869.
Doney, P. M., & Cannon, J. P. (1997). An examination of the nature of trust in buyer-seller
relationships. Journal of Marketing, 61(April), 35-51.
Doney, P. M., Cannon, J. P., & Mullen, M. R. (1998). Understanding the influence of national
culture on the development of trust. Academy of Management Review, 23(3), 601-620.
Donthu, N., Cherian, J., & Bhargava, M. (1993). Factors influencing recall of outdoor
advertising. Journal of Advertising Research, 33(3), 64-72.
Dou, X., Walden, J. A., Lee, S., & Lee, J. Y. (2012). Does source matter? Examining source
effects in online product reviews. Computers in Human Behavior, forthcoming.
Dowling, G. R. (1986). Perceived risk: The concept and its measurement. Psychology &
Marketing, 3(3), 193-210.
Drèze, X., & Hussherr, F. X. (2003). Internet advertising: Is anybody watching? Journal of
Interactive Marketing, 17(4), 8-23.
Duan, W., Gu, B., & Whinston, A. B. (2008). Do online reviews matter? - An empirical
investigation of panel data. Decision Support Systems, 45(4), 1007-1016.
Ducoffe, R. H. (1996). Advertising value and advertising on the web. Journal of Advertising
Research, 36(September/October), 21-35.

338
Duhan, D., Johnson, S., Wilcox, J., & Harrell, G. (1997). Influences on consumer use of word-
of-mouth recommendation sources. Journal of the Academy of Marketing Science,
25(4), 283-295.
Dunn, M. H. (1988). Trust in political agency. In D. Gambetta (Ed.), Trust: Making and
breaking cooperative relations (pp. 73-93). New York, NY: Blackwell.
Dunn, P. (2000). The importance of consistency in establishing cognitive-based trust: A
laboratory experiment. Teaching and Business Ethics, 4(3), 285-306.
Dunning, D. (2007). Self-image motives and consumer behavior: How sacrosanct self-beliefs
sway preferences in the marketplace self image motives. Journal of Consumer
Psychology, 17(4), 237-249.
Durkheim, E. (1951). Suicide: A study in sociology. New York, NY: Free Press.
Durvasula, S., Andrews, C. A., Lysonski, S., & Netemeyer, R. G. (1993). Assessing the cross-
national applicability of consumer behavior models: A model of attitude toward
advertising in general. Journal of Consumer Research, 19(4), 626-636.
Durvasula, S., Lysonski, S., & Andrews, C. J. (1993). Cross-cultural generalizability of a scale
for profiling consumers' decision-making styles. Journal of Consumer Affairs, 27(1),
55-65.
Dwyer, F. R., & Oh, S. (1987). Output sector munificence effects on the internal political
economy of marketing channels. Journal of Marketing Research, 24(4), 347-358.
Dwyer, P. (2007). Measuring the value of electronic word of mouth and its impact in consumer
communities. Journal of International Marketing, 21(2), 63-79.
Dwyer, R. F., Schurr, P. H., & Oh, S. (1987). Developing buyer-seller relationships. Journal of
Marketing, 51(2), 11-27.
Dyer, J. H., & Chu, W. (2003). The role of trustworthiness in reducing transaction costs and
improving performance: Empirical evidence from the United States, Japan, and Korea.
Organization Science, 14(1), 57-68.
Eagly, A. H., & Chaiken, S. (1993). The Psychology of Attitudes. Fort Worth, TX: Harcourt
Brace Jovanovich.
Eagly, A. H., & Chaiken, S. (1998). Attitude structure and function. In D. T. Gilbert, S. T. Fiske
& G. Lindzey (Eds.), Handbook of social psychology (pp. 269-322). New York, NY:
McGraw-Hill.
Eastin, M., & LaRose, R. (2000). Internet self-efficacy and the psychology of the digital divide.
Journal of Computer-mediated Communication, 6(1).
Eayrs, M. A. (1993). Time, trust and hazard: Hairdressers' symbolic roles. Symbolic Interaction,
16, 19-37.
Eccleston, D., & Griseri, L. (2008). How does Web 2.0 stretch traditional influencing patterns?
International Journal of Market Research, 50(5), 575-590.
Edwards, J. R. (2001). Multidimensional constructs in organizational behavior research: An
integrative analytical framework. Organizational Research Methods, 4(2), 141-192.
Ekici, A. (2004). Pradoxes, ambivalences, and consumer coping strategies of food
biotechnologies. Advances in Consumer Research, 31, 237-271.
eMarketer (Producer). (2008, 2012-08-09). The growing influence of online social shoppers.
Retrieved from [Link]
eMarketer. (2011). Social media outlook for 2011. Retrieved 2014-05-29, 2014, from
[Link]
outlook-2011/
Erdogan, B. Z. (1999). Celebrity endorsement: A literature review. Journal of Marketing
Management, 15(3), 291-314.
Erikson, E. H. (1963). Childhood ans society. New York, NY: W. W. Norton.
Erikson, E. H. (1968). Identity: Youth and crisis. New York, NY: Norton.

339
European Commission. (2014). Trust online: Enhancing trust through intermediaries,
trustworthy consumer reviews and beyond? Retrieved 2014-05-30, from
[Link]
Eysenbach, G., & Diepgen, T. L. (1999). Labeling and filtering of medical informatiion on the
internet. Methods of Information in Medicine, 38(2), 80-88.
Fagley, N. S., & Miller, P. M. (1987). The effects of decision framing on choice of risky vs.
certain options. Organizational Behavior and Human Decision Processes, 39, 264-277.
Fahr, R., & Irlenbusch, B. (2008). Identifying personality traits to enhance trust between
organisations: An experimental approach. Managerial and decision economics, 29, 469-
487.
Fallows, D. (Producer). (2005, 08.02.2012). Search engine users. Retrieved from
[Link]
Fang, Y.-H., & Chiu, C.-M. (2010). In justice we trust: Exploring knowledge-sharing
continuance intentions in virtual communities of practice. Computers in Human
Behavior, 26, 235-246.
Farmer, T. A. (1993). Testing the effect of risk attitude on auditor judgments using
multiattribute utility theory. Journal of Accounting, Auditing, and Finance, 8(1), 91-
114.
Farrell, A. M. (2010). Insufficient discriminant validity: A comment on Bove, Pervan, Beatty
and Shiu (2009). Journal of Business Research, 63(3), 324-327.
Farris, G., Senner, F., & Butterfield, D. (1973). Trust, culture, and organizational behavior.
Industrial Relations, 12, 144-157.
Fassnacht, M., & Köse, I. (2008). Consequences of web-based service quality: Uncovering a
multi-faceted chain of effects. Journal of Interactive Marketing, 21(3), 35-54.
Fazio, R. H., & Zanna, M. P. (1981). Direct experience and attitude-behavior consistency. In
L. Berkowitz (Ed.), Experimental Social Psychology. New York, NY: Academic Press.
Feick, L. F., & Higie, R. A. (1992). The effects of preference heterogeneity and source
characteristics on ad processing and judgments about endorsers. Journal of Advertising,
21(June), 9-24.
Ferdous, A. S., & Towfique, B. (2008). Consumer sentiment towards marketing in Bangladesh:
The relationship between attitudes to marketing, satisfaction and regulation. Marketing
Intelligence & Planning, 26(5), 481-495.
Fernandez, C. (2011). Tripadvisor bribes: Hotel owners offer free rooms in reutnr for glowing
reviews. Retrieved 2014-05-15, 2014, from [Link]
2013391/[Link]
Ferres, N., Julia, C., & Travaglione, A. (2004). Co-worker trust as a social catalyst for
constructive employee attitudes. Journal of Managerial Psychology, 19(6), 608-622.
Festinger, L. (1954). A theory of social comparison processes. Human Relations, 7, 117-140.
Festinger, L. (1957). A theory of cognitive dissonance. Stanford: Stanford University Press.
Finn, R. (1972). Effects of some variations in rating scale characteristics on the means and
reliabilities of ratings. Educational and Psychometrical Measurement, 32(2), 255-265.
Fishbein, M., & Ajzen, E. (1975). Belief, attitude, intention and behavior: An introduction to
theory and research. Reading, MA: Addison and Wesley.
Fishburn, P. C. (1977). Mean-risk analysis with risk associated with below-target returns.
American Economic Review, 67(2), 116-126.
Fisher, R., & Chu, S. Z. (2009). Initial online trust formation: The role of company location and
web assurance. Managerial Auditing Journal, 24(6), 542-563.
Flanagin, A. J., & Metzger, M. J. (2000). Perceptions of information credibility. Journal of
Mass Communication Quarterly, 77(3), 515-540.

340
Flavian, C., & Guinaliu, M. (2005). The influence of virtual communities on distribution
strategies in the Internet. International Journal of Retail & Distribution Management,
33(6), 405-425.
Flavian, C., & Guinaliu, M. (2006). Consumer trust, perceived security and privacy policy:
Three basic elements of loyalty to a web site. Industrial Management & Data Systems,
106(5), 601-620.
Floyd, F. J., & Widaman, K. F. (1995). Factor analysis in the development and refinement of
clinical assessment instruments. Psychological Assessment, 7(3), 286-299.
Flynn, L. R. (1993). Do standard scales work in older samples? Marketing Letters, 4(April),
127-138.
Foddy, W. (1993). Constructing questions for interviews and questionnaires. Theory and
practice in social research. Cambridge, UK: Cambridge University Press.
Fogel, J., & Nehmad, E. (2009). Internet social network communities: Risk taking, trust, and
privacy concerns. Computers in Human Behavior, 25, 153-160.
Fogg, B. J., Kameda, T., Boyd, J., Marshall, J. J., Sethi, R., Sockol, M., & Trowbridge, T.
(2002). Stanford-Makovsky web credibility study 2002: Investigating what makes
websites credible today: Stanford University.
Fong, J., & Burton, S. (2006). Online word-of-mouth: A comparison of American and Chinese
discussion boards. Asia Pacific Journal of Marketing and Logistics, 18(2), 146-156.
Ford, G. T., Smith, D. B., & Swasy, J. L. (1988). An empirical test of the search, experience,
and credence attributes framework. Advances in Consumer Research, 15, 239-243.
Ford, G. T., Smith, D. B., & Swasy, J. L. (1990). Consumer skepticism of advertising claims:
Testing hypotheses from economics of information. Journal of Consumer Research, 16,
433-441.
Ford, J. D., & Ellis, E. (1980). A Re-examination of group influence on member brand
performance. Journal of Marketing Research, 17(2), 125-132.
Forehand, M. R., & Grier, S. (2003). When is honesty the best policy? The effect of stated
company intent on consumer skepticism. Journal of Consumer Psychology, 13(3), 349-
356.
Forman, C., Ghose, A., & Wiesenfeld, B. (2008). Examining the relationship between reviews
and sales: The role of reviewer identity disclosure in electronic markets. Information
Systems Research, 19(3), 291.
Fornell, C., & Bookstein, F. L. (1982). A comparative analysis of two structural equation
models: LISREL and PLS applied to market data. In C. Fornell (Ed.), A second
generation of multivariate analysis (pp. 289-324). New York, NY: Praeger.
Fornell, C., & Larcker, D. F. (1981). Evaluating structural equation models with unobservable
variables and measurement error. Journal of Marketing Research, 18(1), 39-50.
Foubert, J. D., & Sholley, B. K. (1996). Effects of gender, gender role, and individualized trust
on self-disclosure. Journal of Social Behavior and Personality, 11(5), 277-288.
Fournier, S. (1998). Consumers and their brands: Developing relationship theory in consumer
research. Journal of Consumer Research, 24, 343-373.
Freeman, K. S., & Spyriadakis, J. H. (2004). An examination of factors that affect the credibility
of online health information. Technical Communication, 51(2), 239-263.
Frenzen, J. K., & Davis, H. L. (1990). Purchasing behavior in embedded markets. Journal of
Consumer Research, 17(1-12).
Friedland, N. (1990). Attribution of control as a determinant of cooperation in exchange
interactions. Journal of Applied Social Psychology, 20, 303-320.
Frost, T., Stimpson, D. V., & Maughan, M. R. C. (1978). Some correlates of trust. Journal of
Psychology, 99, 103-108.
Fu, J. (1993). Increased risk aversion and risky investment. The Journal of Risk and Insurance,
60(3), 494-501.
341
Fukuyama, F. (1995). Trust: The social virtues and the creation of prosperity. New York, NY:
Free Press.
Gabarro, J. J. (1978). The development of trust, influence and expectations. In A. G. Anthos &
J. J. Gabarro (Eds.), Interpersonal behavior: Communication and understanding in
relationships (pp. 290-330). Englewood Cliffs, NJ: Prentice-Hall.
Gaines, J. H. (1980). Upward communication in industry: an experiment Human Relations,
33(12), 929-942.
Gambetta, D. (1988). Trust: Making and braking cooperative relations. New York, NY:
Blackwell.
Ganesan, S. (1994). Determinants of long-term orientation in buyer-seller relationships.
Journal of Marketing, 58, 1-19.
Ganesan, S., & Hess, R. (1997). Dimensions and levels of trust: Implications for commitment
to a relationship. Marketing Letters, 8(4), 439-448.
Ganesh, J., Reynolds, K. E., Luckett, M., & Pomirleanu, N. (2010). Online shopper motivations,
and e-store attributes: An examination of online patronage behavior and shopper
typologies. Journal of Retailing, 1, 106-115.
Garbarino, E., & Johnson, M. S. (1999). The different roles of satisfaction, trust, and
commitment in customer relationships. Journal of Marketing, 63(2), 70-87.
Garfinkel, H. (1967). Studies in ethnomethodology. Englewood Cliffs, NJ: Prentice-Hall.
Gatignon, H., & Robertson, T. (1985). A propositional inventory for new diffusion research.
Journal of Consumer Research, 11(March), 849-867.
Gauzente, C. (2010). The intention to click on sponsored ads - A study of the role of prior
knowledge and of consumer profile. Journal of Retailing and Consumer Services, 17,
457-463.
Gee, S. (2011). Amazon customer reviews - can we trust them? i Programmer. [Link]
[Link]/news/155-social-media/2922-amazon-cu
Gefen, D. (1997). Building users' trust in freeware providers and the effects of this trust on
users' perceptions of usefulness, ease of use and intended use of freeware. (Unpublished
Doctoral Dissertation), Atlanta, GA.
Gefen, D. (2000). E-commerce: The role of familiarity and trust. Omega International Journal
of Management Studies, 28(6), 725-737.
Gefen, D. (2002a). Customer loyalty in e-commerce. Journal of the Association for Information
Systems, 3, 27-51.
Gefen, D. (2002b). Reflections on the dimensions of trust and trustworthiness among online
consumers. Database for Advances in Information Systems, 33(3), 38-53.
Gefen, D., Benbasat, I., & Pavlou, P. A. (2008). A research agenda for trust in online
environments. Journal of Management Information Systems, 24(4), 275-286.
Gefen, D., & Heart, T. (2006). On the need to include national culture as a central issue in e-
commerce trust beliefs. Journal of Global Information Management, 14(4), 1-30.
Gefen, D., Karahanna, E., & Straub, D. W. (2003). Trust and TAM in online shopping: An
integrated model. MIS Quarterly, 27(1), 51-90.
Gefen, D., Rao, V. S., & Tractinsky, N. (2003). The conceptualization of trust, risk, and their
relationship in electronic commerce: The need for clarifications. Proceedings of the
36th Hawaii International Converence on System Sciences.
Gefen, D., & Silver, M. (1999). Lessons learned form the successful adoption of an ERP system.
Paper presented at the Proceedings of the 5th International Conference of the Decision
Sciences Institute, Athens, Greece.
Gefen, D., & Straub, D. (2003). Managing user trust in B2C e-services. E-Service Journal,
[Link]

342
Gefen, D., & Straub, D. W. (2004). Consumer trust in B2C e-commerce and the importance of
social presence: Experiments in e-products and e-services. Omega International
Journal of Management Studies, 32, 407-424.
Gerbing, D. W., & Anderson, J. C. (1988). An update paradigm for scale development
incorporating unidimensionality and its assessment. Journal of Marketing Research,
25(May), 186-192.
Geyskens, I., & Steenkamp, J.-B. (1995). An investigation into the joint effects of trust and
interdependence on relationship commitment. Paper presented at the 24th Annual
Conference of the European Marketing Academy.
Geyskens, I., Steenkamp, J.-B., & Kumar, N. (1997). Generalizations about trust in marketing
channel relationships using meta-analysis. International Journal of Research in
Marketing, 15, 223-248.
Gholamisaman, E. (2012). Social and personal factors influencing individuals participation in
eWOM on social networking sites. Paper presented at the International Congress on
Business and Economic Research, Port Harcourt, Nigeria.
Ghose, A., & Ipeirotis, P. G. (2007). Designing novel review ranking systems: Predicting the
usefulness an impact of reviews. Paper presented at the Ninth International Conference
on Electronic Commerce (ICEC).
Gibbons, F. X., & Buunk, B. P. (1999). Individual differences in social comparison:
Development of a scale of social comparison orientation. Journal of Personality and
Social Psychology, 76(1), 129-142.
Giffin, K. (1967). The contribution of studies of source credibility to a theory of interpersonal
trust in the communication process. Psychological Bulletin, 68(2), 104-120.
Gilbert, D. T., Giesler, R. B., & Morris, K. A. (1995). When comparisons arise. Journal of
Personality and Social Psychology, 69, 227-236.
Gilly, M. C., Graham, J. L., Wolfinbarger, M., & Yale, L. J. (1998). A dyadic study of
interpersonal information search. Journal of the Academy of Marketing Science, 2, 83-
100.
Godes, D., & Mayzlin, D. (2004). Using online conversations to study word of mouth
communication. Marketing Science, 23(4), 445-460.
Goldberg, L. R. (1993). The structure of phenotypic personality traits. American Psychologist,
48(1), 26-34.
Goldsmith, R. E. (2006). Electronic word-of-mouth. Hershey, PA: Idea Group Publishing.
Goldsmith, R. E., & Emmert, J. (1991). Measuring product category involvement: A multitrait-
multimethod study. Journal of Business Research, 23, 363-371.
Goldsmith, R. E., & Lafferty, B. A. (2002). Consumer response to web sites and their influence
on advertising effectiveness. Internet Research, 12(4), 318-328.
Golembiewski, R. T., & McConkie, M. (1975). The centrality of interpersonal trust in group
processes. In C. L. Cooper (Ed.), Theories of group processes (pp. 131-185). New York,
NY: Wiley.
Good, P. (1988). Individuals, interpersonal relations, and trust. In D. G. Gambetta (Ed.), Trust:
Making and breaking cooperative relations (pp. 31-48). New York, NY: Basil
Blackwell.
Gorsuch, R. L. (1983). Factor analysis. Hillsdale, NJ: Lawrence Erlbaum.
Gosh, D., & Ray, M. R. (1992). Risk attitude, ambiguity intolerance and decision making: An
exploratory investigation. Decision Sciences, 23(2), 431-444.
Grabner-Kraeuter, S. (2002). The role of consumers' trust in online-shopping. Journal of
Business Ethics, 39, 43-50.
Grabner-Kraeuter, S., & Faullant, R. (2008). Consumer acceptance of Internet banking: The
influence of Internet trust. International Journal of Bank Marketing, 26(7), 483-504.

343
Grabner-Kraeuter, S., & Kaluscha, E. A. (2003). Empirical research in online trust: A review
and critical assessment. International Journal of Human-Computer Studies, 58, 783-
812.
Granovetter, M. (1985). Economic action and social structure: The problem of embeddedness.
The American Journal of Sociology, 91(3), 481-450.
Granovetter, M. S. (1973). The strength of weak ties. The American Journal of Sociology, 78(6),
1361-1381.
Greenwald, A. G., & Leavitt, C. (1984). Audience involvement in advertising: Four levels.
Journal of Consumer Research, 11(6), 581-592.
Gretzel, U. (2006). Consumer generated content - trends and implications for branding. Review
of Tourism Research, 4(3), 9-11.
Grimes, A. J. (1978). Authority, power, influence, and social control: A theoretical synthesis.
Academy of Management Review.
Grosuch, R. L. (1983). Factor analysis. Hillsdale, NJ: Lawrence Erlbaum.
Guenzi, P. (2002). Sales force acitivies and customer trust. Journal of Marketing Management,
18, 749-778.
Gulati, R. (1995). Does familiarity breed trust? The implications of repeated ties for contractual
choice in alliances. Academy of Management Journal, 38(1), 85-112.
Gulati, R., & Sytch, M. (2008). The dynamics of trust. Academy of Management Review, 33(1),
276-278.
Gundlach, G. T., & Murphy, P. E. (1993). Ethical and legal foundations of relational marketing
exchanges. Journal of Marketing, 57, 35-46.
Gunther, A. C. (1992). Biased press or biased public attitudes toward media coverage of social
groups. Public Opinion Quarterly, 56(2), 147-167.
Gupta, P., Yadav, M. S., & Varadarajan, R. (2009). How task-facilitative interactive tools foster
buyer's trust in online retailers: A process view of trust development in the electronic
marketplace. Journal of Retailing, 85(2), 159-176.
Gurtman, M. B. (1992). Trust, distrust, and interpersonal problems: A circumplex analysis.
Journal of Personality and Social Psychology, 62, 989-1002.
Hair, J. F., Black, W. C., Babin, B. J., & Anderson, R. E. (2010). Multivariate data analysis -
A global perspective. Upper Saddle River, NJ: Pearson.
Haley, E. (1996). Exploring the construct of organizuation as source: Consumers' understanding
of organizational sponsorship of advocacy adverstising. Journal of Advertising, 25(2),
19-35.
Hallen, L., & Sandstrom, M. (1991). Relationship atmosphere in international business. New
Perspectives on International Marketing, 108-125.
Hancock, J. T., Curry, L. E., Goorha, S., & Woodworth, M. (2007). On lying and being lied to:
A linguistic analysis of deception in computer-mediated communication. Discourse
Processes, 45(1), 1-23.
Hansen, T. (2005). Consumer adoption of online grocery buying: A discriminant analysis
International Journal of Retail & Distribution Management, 33, 101-121.
Harkins, S. G., & Petty, R. E. (1987). Informatiion utility and the multiple source effect. Journal
of Personality and Social Psychology, 52(2), 260-268.
Harris Interactive. (2010). Speak now or forever hold your tweets.
Harris, L. C., & Goode, M. M. (2004). The four levels of loyalty and the pivotal role of trust: a
study of online service dynamics. Journal of Retailing, 80(2), 139-158.
Hart, P., & Saunders, C. (1997). Power and trust: Critical factors in the adoption and use of
electronic data interchange. Organizational Science, 8(1), 23-42.
Hass, R. G. (1981). Effects of source characteristics on cognitive repsones and persuasion. In
R. E. Petty, T. M. Ostrom & T. C. Brock (Eds.), Cognitive responses in persuasion (pp.
141-172). Hillsdale, NJ: Lawrence Erlbaum Associates.
344
Hattie, J. (1985). Methodology review: Assessing unidimensionality of tests and items. Applied
Psychological Measurement, 9(June), 139-164.
Haugtvedt, C. P., & Petty, R. E. (1992). Personality and persuasion: Need for cognition
moderates the persistence and resistance of attitude change. Journal of Personality and
Social Psychology, 63, 308-319.
Hawes, J. M., Mast, K. e., & Swan, J. E. (1989). Trust earning perceptions of sellers and buyers.
Journal of Personal Selling and Sales Management, 9(Spring), 1-8.
Haynes, S., Neslon, N. K., & Baline, D. (1999). Psychometric issues in assessment research. In
P. C. Kendall, J. N. Butcher & G. Holmbeck (Eds.), Handbook of research methods in
clinical psychology. New York, NY: John Wiley & Sons.
Haynes, S. N., Richard, D. C. S., & Kubany, E. S. (1995). Content validity in psychological
assessment: A functional approach to concepts and methods. Psychological Assessment,
7(3), 238-247.
He, S. X., & Bond, S. D. (2013). Word-of-mouth and the forecasting of consumption
enjoyment. Journal of Consumer Psychology, 23(4), 464-482.
Heeler, R. M., & Ray, M. L. (1972). Measure validation in marketing. Journal of Marketing
Research, 9(November), 361-370.
Heider, F. (1946). Attitudes and cognitive organization. The Journal of Psychology, 21, 107-
112.
Heimberg, R. G., & Holaway, R. M. (2007). Examination of the known-groups validity of the
Liebowitz social anxiety scale. Depression and Anxiety, 24(7), 447-454.
Helm, A. (2004). Cynics and skeptics: Consumer dispositional trust. Advances in Consumer
Research, 31, 345-351.
Hemphill, K. J., & Lehman, D. R. (1991). Social comparisons and their affective consequences:
The importance of comparison dimension and individual difference variables. Journal
of Social and Clinical Psychology, 10, 372-394.
Hennig-Thurau, T., Gwinner, K. P., Walsh, G., & Gremler, D. D. (2004). Electronic word of
mouth via consumer-opinion platforms: What motivates consumers to articulate
themselves on the Internet? Journal of Interactive Marketing, 18(1), 38-52.
Heretick, D. M. L. (1981). Gender-specific relationships between trust-suspicion, locus of
control, and psychological distress. Journal of Psychology, 108, 267-274.
Hernandez, B., Jimenez, J., & Martin, M. J. (2009). Adoption vs acceptance of e-commerce:
Two different decisions. European Journal of Marketing, 43 (9/10), 1232-1245.
Herr, P. M., Kardes, F. R., & Kim, J. (1991). Effects of word-of-mouth and product attribute
information on persuation: An accessibility-diagnosticity perspective. Journal of
Consumer Research, 17(March), 454-462.
Hicks, A., Comp, S., Horovitz, J., Hovarter, M., Miki, M., & Bevan, J. L. (2012). Why people
use [Link]: An exploration of uses and gratifications. Computers in Human
Behavior, 28(6), 2274-2279.
Hildebrandt, L., & Temme, D. (2006). Probleme der Validierung mit
Strukturgleichungsmodellen. Die Betriebswirtschaft, 66(6), 618-639.
Hilgard, E. R. (1980). The trilogy of the mind: Cognition, affection, and conation. Journal of
the History of Behavioral Sciences, 16, 107-117.
Hill, C. W. L. (1990). Cooperation, opportunism, and the invisible hand: Implications for
transaction cost theory. Academy of Management Review, 15, 500-513.
Himmelfarb, S., & Eagly, A. H. (1974). Readings in attitude change. New York, NY: Wiley.
Hinkin, T. R., & Tracey, J. B. (1999). An analysis of variance approach to conent validation.
Organizational Research Methods, 2(2), 175-186.
Hinnant, L. C. (2007). The trust experience from the truster's perspective: A theoretical
discussion and experiment. Syracuse University.

345
Ho, J. Y. C., & Dempsey, M. (2010). Viral marketing: Motivations to forward online content.
Journal of Business Research, 63(9-10), 1000-1006.
Hofstede, G. (1980). Motivation, leadership, and organization: Do American theories apply
abroad? Organizational Dynamics, 9(1), 42-63.
Hogg, M., & Cooper, J. (2003). The Sage handbook of social psychology. London, UK: Sage.
Hogg, M. A., & Vaughan, G. M. (2011). Social psychology. Harlow, UK: Prentice Hall.
Hollebeek, L. D. (2011). Demystifying customer brand engagement: Exploring the loyalty
nexus. Journal of Marketing Management, 27(7/8), 785-807.
Holmes, J. G. (1991). Trust and the appraisal process in close relationships. In W. H. Jones &
D. Perlman (Eds.), Advances in Personal Relationships (pp. 57-104). London, UK:
Jessica Kingsley.
Homans, G. (1950). The human group. New York, NY: Routledge & K. Paul.
Homans, G. (1961). Social behavior: Its elementary forms. New York, NY: Harcourt, Brace &
World.
Homburg, C., Droll, M., & Totzek, D. (2008). Customer priorization: Does it pay off, and how
should it be implemented? Journal of Marketing, 72(5), 110-130.
Homburg, C., & Giering, A. (1996). Konzeptualisierung und Operationalisierung komplexer
Konstrukte. Marketing - Zeitschrift für Forschung und Praxis, 18(1), 5-24.
Hong, S., & Park, H. S. (2012). Computer-mediated persuasion in online reviews: Statistical
versus narrative evidence. Computers in Human Behavior, 28, 906-919.
Horn, J. L. (1965). A rationale and test for the number of factors in factor analysis.
Psychometrika, 30, 179-186.
Horrigan, J. B. (2008). The internet and consumer choice: Online Americans use different
search and purchase strategies for different goods. Pew Internet & American Life
Project, May 18, 1-41.
Hosmer, L. T. (1995). Trust: The connecting link between organizational theory and
philosophical ethics. The Academy of Management Review, 20(2), 379-403.
Hovland, C. I., Janis, I. L., & Kelley, H. H. (1953). Communication and persuasion New Haven,
CT: Yale University Press.
Hoyle, R. (1995). Structural equation modeling: Issues and applications. Newbury Park, CA:
Sage.
Hsiao, K.-L., Lin, J. C.-C., Wang, X.-Y., Lu, H.-P., & Yu, H. (2010). Antecedents and
consequences of trust in online product recommendations. Online Information Review,
34(6), 935-953.
Hsu, M.-H., Ju, T. L., Yen, C.-H., & Chang, C.-M. (2007). Knowledge sharing behavior in
virtual communities: The relationship between trust, self-efficacy, and outcome
expectations. International Journal of Human-Computer Studies, 65, 153-169.
Hu, L., & Bentler, P. M. (1999). Cutoff criteria for fit indices in covariance structure analysis:
Conventional criteria versus new alternatives. Structural Equation Modeling, 6(1), 1-
55.
Hu, N., Bose, I., Gao, Y., & Liu, L. (2011). Manipulatioin in digital word-of-mouth: A reality
check for book reviews. Decision Support Systems, 50(3), 627-635.
Huang, J.-H., & Chen, Y.-F. (2006). Herding in online product choice. Psychology &
Marketing, 23(5), 413.
Huang, Y. K., Yang, W. I., Lin, T. M. Y., & Shih, T. Y. (2012). Judgment criteria for the
authenticity of internet book reviews. Library & Information Science Research, 34(2),
150-156.
Hudson, B. (2004). Trust: Towards conceptual clarification. Australian Journal of Political
Science, 39(1), 75-87.
Huff, L. C. (2000). An integrated model of consumer trust formation. Paper presented at the
American Marketing Association, Chicago.
346
Hultén, B. (2007). Customer segmentation: The concepts of trust, commitment and
relationships. Journal of Targeting, Measurement and Analysis for Marketing, 15(4),
256-269.
Hung, K., Li, S. Y., & Tse, D. K. (2011). Interpersonal trust and platform credibiltiy in a
Chinese multibrand online community. Journal of Advertising, 40(3), 99-112.
Hupfer, N., & Gardner, D. (1971). Differential involvement with products and issues: An
exploratory study. In D. M. Gardner (Ed.), Proceedings (pp. 262-269). College Park,
MD: Association for Consumer Research.
Husted, B. W. (1990). Trust in business relations: Directions for empirical research Business
Professional Ethics Journal, 8(2), 23-40.
Husted, B. W. (1998). The ethical limits of trust in business relations. Business Ethics
Quarterly, 8(2), 233-248.
Insko, C. A., Kirchner, J. L., Pinter, B., & Wildschut, T. (2005). Interindividual-intergroup
discontinuity as a function of trust and categorization: The paradox of expected
cooperation. Journal of Personality and Social Psychology, 88(2), 365-385.
Iyer, G. R., Soberman, D., & Villas-Boas, J. M. (2005). The targeting of advertising. Marketing
Science, 24(3), 461-476.
Izard, C. E. (1991). The psychology of emotions. New York, NY: Plenum Press.
Izard, C. E. (1992). Basic emotions, relations among emotions, and emotion-cognition
relations. Psychological Review, 99(3), 561-565.
Jacoby, J., & Kaplan, L. B. (1972). The components of perceived risk. Paper presented at the
Proceedings of the Third Annual Conference of the Association for Consumer Research,
Chicago, IL.
Jamal, A., & Naser, K. (2002). Customer satisfaction and retail banking: an assessment of some
of the key antecedents of customer satisfaction in retail banking. International Journal
of Bank Marketing, 20(4), 146-160.
James, L. R., & Jones, A. P. (1974). Organizational climate: A review of theory and research.
Psychological Bulletin, 81, 1096-1112.
James, W. I., & Kover, A. (1992). Do overall attitudes toward advertising affect involvement
with specific advertisements? Journal of Advertising Research, 32(5), 78-83.
Jarvenpaa, S. L., Knoll, K., & Leidner, D. E. (1998). Is anybody out there? Antecedents of trust
in global virtual teams. Journal of Management Information Systems, 14(4), 29-64.
Jarvenpaa, S. L., & Leidner, D. E. (1998). Communication and trust in global virtual teams.
Journal of Computer-mediated Communication, 3(4).
Jarvenpaa, S. L., & Tractinsky, N. (1999). Consumer trust in an Internet store: A cross-cultural
validation. Journal of Computer-mediated Communication, 5(2), 1-37.
Jarvenpaa, S. L., Tractinsky, N., & Vitale, N. (2000). Consumer trust in an internet store.
Information Technology and Management, 1(1-2), 45-71.
Jarvis, C. B., MacKenzie, S. B., & Podsakoff, P. M. (2003). A critical review of construct
indicators and measurement model misspecification in marketing and consumer
research. Journal of Consumer Research, 30(9), 199-218.
Jarvis, W. B. G., & Petty, R. E. (1995). The need to evaluate. Journal of Personality and Social
Psychology, 70, 172-192.
Jeong, E., & Jang, S. (2011). Restaurant experiences triggering positive electronic word-of-
mouth (eWOM) motivations. International Journal of Hospitality Management, 30,
356-366.
Jepsen, A. L. (2007). Factors affecting consumer use of the Internet for information search.
Journal of Interactive Marketing, 21(3), 21-34.
Jevson, C., & Gabbott, M. (2000). Trust, brand equity and brand reality in internet business
relationships: An interdisciplinary approach. Journal of Marketing Management, 16,
619-634.
347
Jian, J., Bisantz, A. M., & Drury, C. G. (2000). Foundations for an empirically determined scale
of trust in automated systems. International Journal of Cognitive Ergonomics, 4(1), 53-
71.
John, G., & Reve, T. (1982). The reliability and validity of key informant data from dyadic
relationships in markting channels. Journal of Marketing Research, 19(11), 517-524.
John, O. P., & Srivastava, S. (1999). The big-five trait taxonomy: History, measurement, and
theoretical perspectives. In L. Pervin & O. P. John (Eds.), Handbook of personality:
Theory and research. New York, NY: Guilford.
Johnson-George, C., & Swap, W. C. (1982). Measurement of specific interpersonal trust:
Construction and validation of a scale to assess trust in a specific other. Journal of
Personality and Social Psychology, 43, 1306-1317.
Johnson, D. S., & Grayson, K. (2005). Cognitive and affective trust in service relationships.
Journal of Business Research, 58(4), 500-507.
Johnson, E. B. (1996). Cognitive age: Understanding consumer alienation in the mature market.
Review of Business, 17(3), 35-40.
Johnson, T. J., & Kaye, B. K. (2002). Webelievability: A path model examining how
convenience and reliance predict online credibility. Journalism and Mass
Communication Quarterly, 79, 619-642.
Johnson, T. J., & Kaye, B. K. (2009). In blog we trust? Deciphering credibility of components
of the internet among politically interested internet users. Computers in Human
Behavior, 25(1), 175-182.
Jones, G. R., & George, J. M. (1998). The experience and evolution of trust: Implications for
cooperation and teamwork. The Academy of Management Review, 23(3), 531-546.
Jones, K., & Leonard, L. N. K. (2008). Trust in consumer-to-consumer electronic commerce.
Information & Management, 45(2), 88-95.
Jöreskog, K. G. (1971). Statistical analysis of sets of congeneric tests. Psychometrika, 36, 109-
133.
Jöreskog, K. G., & Sörbom, D. (1981). LISREL V: Analysis of linear structural relations by the
method of maximum likelihood. Chicago, IL: International Educational Services.
Jöreskog, K. G., & Sörbom, D. (1984). LISREL 6: A guide to the program and applications.
Chicago, IL: SPSS.
Jöreskog, K. G., & Sörbom, D. (1989). LISREL 7: A guide to the program and applications.
Chicago, IL: SPSS.
Jöreskog, K. G., & Sörbom, D. (1993). LISREL VIII: Users' reference guide. Chicago, IL:
Scientific Software International.
Jung, C. G. (1971). Psychological types (Vol. 7). Princeton, NJ: Princeton University Press.
Kahnemann, D., Slovic, P., & Tversky, A. (1982). Judgment under uncertainty: Heuristics and
biases. Cambridge: Cambridge University Press.
Kaiser, H. F. (1960). The application of electronic computers to factor analysis. Educational
and Psychometrical Measurement, 20, 141-151.
Kamplan, A. M., & Haenlein, M. (2010). Users of the world, unite! The challenges and
opportunities of social media. Business Horizons, 53(1), 59-68.
Kankanhalli, A., Tan, B. C. Y., & Wei, K. K. (2005). Contributing knowledge to electronic
knowledge repositories: An empirical investigation. MIS Quarterly, 29(1), 113-143.
Kanter, D. L., & Mirvis, P. H. (1989). The cynical Americans: Living and working in an age of
discontent and disillusion. San Francisco, CA: Jossey-Bass.
Kanter, D. L., & Wortzel, L. (1985). Cynicism and alienation as marketing considerations:
Some new ways to approach the female consumer. Journal of Consumer Marketing, 2,
5-15.
Kantsperger, R., & Kunz, W. H. (2010). Consumer trust in service companies: A multiple
mediating analysis. Managing Service Quality, 20(1), 4-25.
348
Kaplan, A. (1964). The conduct of inquiry. New York, NY: Chandler.
Kaplan, L. B., Szybillo, G. J., & Jacoby, J. (1974). Components of perceived risk in product
purchase: A cross-validation. Journal of Applied Psychology, 59(3), 287-291.
Kaplan, R. M. (1973). Components of trust: Note on use of Rotter's scale. Psychological
Reports, 63, 143-148.
Kaplan, R. M., & Saccuzzo, D. P. (1997). Psychological testing: Principles, applications, and
issues. Pacific Grove, CA: Brooks/Cole.
Kaplan, S., & Nieschwietz, R. (2003). An examination of the effects of web trust and company
type on consumers' purchase intentions. International Journal of Auditing, 7, 155-168.
Kasperson, R. E., Golding, D., & Tuler, S. (1992). Social distrusts as a factor in siting hazardous
facilities and communicating risks. Journal of Social Issues, 48(4), 161-187.
Kassarjian, W. M. (1962). A study of Riesman's theory of social character. Sociometry, 25, 213-
230.
Katz, D. (1960). The functional approach to the study of attitudes. The Public Opinion
Quarterly, 24(2), 163-204.
Katz, D., & Stotland, E. (1959). A preliminary statement to a theory of attitude structure and
change. In S. Koch (Ed.), Psychology: A study of a science (Vol. 3) (pp. 423-475). New
York, NY: McGraw-Hill.
Katz, E., & Lazarsfeld, P. F. (1955). The part played by people in the flow of mass
communication. Clencoe, IL: Free Press.
Kautonen, T. (2008). Trust and new technologies: Marketing and management on the internet
and mobile data. Cheltenham, UK: Edward Elgar
Kazdin, A. E. (2000). Encyclopedia of psychology. Cambridge, MA: Oxford University Press.
Kee, H. W., & Knox, R. E. (1970). Conceptual and methodological consideration in the study
of trust and suspicion. Journal of Conflict Resolution, 14(3), 357-366.
Kee, T. (2008). Majority of online shoppers check at least four reviews before buying.
Retrieved 06.08.2013, 2013, from [Link]
article/76727/majority-of-online-shoppers-check-at-least-four-
[Link]#axzz2bAozVr8Q
Keen, P. G. W., Balance, C., Chan, S., & Schrump, S. (1999). Electronic commerce
relationships: Trust by design. Englewoods Cliffs, NJ: Prentice-Hall.
Keil, M., Wallace, L., Turk, D., Dixon-Randall, G., & Nulden, U. (2000). An investigation of
risk perception and risk propensity on the decision to continue a software development
project. The Journal of Systems and Software, 53, 145-157.
Kelley, H. H., & Thibaut, J. W. (1978). Interpersonal relations: A theory of interdependence.
New York, NY: John Wiley & Sons.
Kelman, H. C. (1961). Processes of opinion change. Public Opinion Quarterly, 25(Spring), 57-
78.
Kelman, H. C., & Hovland, C. I. (1953). Reinstatement of the communicator in delayed
measurement of opinion change. Journal of Abnormal Social Psychology, 48, 327-335.
Kennedy, M. S., Ferrell, L. K., & LeClair, D. T. (2001). Consumer's trust of salesperson and
manufacturer: An empirical study. Journal of Business Research, 51, 73-86.
Kerling, F. N. (1973). Foundations of behavioral research. New York, NY: Holt McDougal.
Kiel, G., & Layton, R. A. (1981). Dimensions of consumer information seeking behavior.
Journal of Marketing Research, 18(May), 233-239.
Kim, D., & Benbasat, I. (2003). Trust-related arguments in internet stores: A framework for
evaluatiion. Journal of Electronic Commerce Research, 4, 49-64.
Kim, D. C. (1992). Risk preferences in participative budgeting. The Accounting Review, 67(2),
303-318.

349
Kim, D. J., Ferrin, D. L., & R., R. H. (2008). A trust-based consumer decision-making model
in electronic commerce: The role of trust, perceived risk, and their antecedents. Decision
Support Systems, 44, 544-564.
Kim, E., & Tadisina, S. (2007). A model of customer's trust in e-businesses: Micro-level inter-
party trust formation. Journal of Computer Information Systems, 48(1), 88-104.
Kim, J., Jin, B., & Swinney, J. L. (2009). The role of etail quality, e-satisfaction and e-trust in
online loyalty development process. Journal of Retailing and Consumer Services, 16,
239-247.
Kim, P. H., Ferrin, D. L., Cooper, C. D., & Dirks, K. T. (2004). Removing the shadow of
suspicion: The effects of apology versus denial for reparing competence- versus
integrity-based trust violations. Journal of Applied Psychology, 89(1), 104-118.
Kim, Y. A., & Ahmad, M. A. (2013). Trust, distrust and lack of confidence of users in online
social media-sharing communities. Knowledge-Based Systems, 37(1), 438-450.
Knoke, D., & Kuklinski, J. H. (1982). Network analysis. Beverly Hills, CA: Sage.
Koehn, D. (1996). Should we trust in trust? American Business Law Journal, 34(2), 183-203.
Kogan, N., & Wallach, M. A. (1964). Risk taking: A study in cognition and personality. New
York, NY: Hold, Rinehart & Winston.
Kohring, M., & Matthes, J. (2007). Trust in news media: Development and validation of a
multidimensional scale. Communication Research, 34, 231-252.
Koller, M. (1988). Risk as a determinant of trust. Basic and Applied Social Psychology, 9(4),
265-276.
Komiak, S. X., & Benbasat, I. (2006). The effects of personalization and familiarity on trust
and adoption of recommendation agents. MIS Quarterly, 30(4), 941-960.
Konovsky, M. A., & Pugh, S. D. (1994). Citizenship behavior and social exchange. Academy
of Management Journal, 37(3), 656-669.
Korfiatis, N., Garcia-Bariocanal, E., & Sanchez-Alonso, S. (2012). Evaluating content quality
and helpfulness of online product reviews: The interplay of review helpfulness vs.
review content. Electronic Commerce Research and Applications, 11, 205-217.
Korsgaard, M. A., Schweiger, D. M., & Speienza, H. J. (1995). Building commitment,
attachment, and trust in strategic decision-making teams: The role of procedural justice.
Academy of Management Journal, 38(1), 60-84.
Koufaris, M., & Hampton-Sosa, W. (2004). The development of inital trust in an online
company by new customers. Information & Management, 41, 377-397.
Kozinets, R. V. (2002). The field behind the screen: Using netography for marketing research
in online communities. Journal of Marketing Research, 39(1), 61-72.
Krackhardt, D., & Stern, R. N. (1988). Informal networks and organizational crises: An
experimental simulation. Social Psychology Quarterly, 51(2), 123-140.
Kramer, R. M. (1999). Trust and distrust in organisations: Emerging perspectives, enduring
questions. Annual Review of Psychology, 50, 569-598.
Kraut, R., Patterson, M., Lundmark, V., Kiesler, S., Mukopadhyay, T., & Scherlis, W. (1998).
Internet paradox: A social technology that reduces social involvement and
psychological well-being? American Psychologist, 53, 1017-1031.
Krech, D., Crutchfield, R., & Ballachey, R. (1962). Individual in society. New York, NY:
McGraw-Hill.
Krishnan, P. (2008). Consumer alienation by brands: Examining the roles of powerlessness
and relationship types. University of Manitoba, Winnipeg.
Krishnan, P., Bhatnagar, N., & Manchanda, R. (2009). Consumer alientation by brands:
Examining the roles of powerlessness and relationship. Advances in Consumer
Research, 36, 955-956.

350
Kropp, F., Lavack, A. M., & Silvera, D. H. (2005). Values and collective self-esteem as
predictors of consumer susceptibility to interpersonal influence among university
students. International Marketing Review, 22(1), 7-33.
Kruglanski, A., Webster, D. W., & Klem, A. (1993). Motivated reistance and openness to
persuasion in the presence or absence of prior information. Journal of Personality and
Social Psychology, 65, 871-876.
Ku, Y.-C., Wei, C.-P., & Hsiao, H.-W. (2012). To whom should I listen? Finding reputable
reviewers in opinion-sharing communities. Decision Support Systems, 53, 534-542.
Kucuk, S. U. (2009). Consumer empowerment model: From unspeakable to undeniable. Direct
Marketing: An International Journal, 3(4), 327-342.
Kuhn, T. S. (1962). The structure of scientific revolutions. Chicago, IL: University of Chicago
Press.
Kumar, N. (1996). The power of trust in manufacturer-retailer relationships. Harvard Business
Review, 92-106.
Kumar, N., & Benbasat, I. (2006). The influence of recommendations and consumer reviews
on evaluations of websites. Information System Research, 17(4), 426-439.
Kumar, N., Scheer, L. K., & Steenkamp, J.-B. E. M. (1995). The effects of perceived
interdependence on dealers attitudes. Journal of Marketing Research, 32, 348-356.
Kumar, R., Novak, J., & Tomkins, A. (2004). Structure and evolution of blogspace.
Communication of the ACM, 47(12), 35-39.
Kumra, R., & Mittal, R. K. (2004). Trust and its determinants in Internet banking: A study of
private sector banks in India. Decision, 31(1), 76-96.
Kwok, J. S. H., & Gao, S. (2004). Knowledge sharing community in P2P network: A study of
motivational perspective. Journal of Knowledge Management, 8, 94-102.
Laer, T. v., & Ruyter, K. (2010). In stories we trust: How narrative apologies provide cover for
competitive vulnerability after integrity-violating blog posts. International Journal of
Research in Marketing, 27, 164-174.
Lafferty, B. A., & Goldsmith, R. E. (1999). Corporate credibility's role in consumers' attitudes
and purchase intentions: When a high versus a low credibility endorser is used in the
Ad. Journal of Business Research, 44, 109-116.
Lambert, Z. V. (1980). Consumer alienation, general dissatisfaction, and consumerism issues:
Conceptual and managerial perspectives. Journal of Retailing, 56(2), 3-24.
Lander, M. C., Purvis, R. L., McCray, G. E., & Leigh, W. (2004). Trust-building mechanisms
utilized in outsourced IS development project: A case study. Information &
Management, 41(4), 509-558.
Lane, C., & Bachmann, R. (1996). The social constitution of trust: Supplier relations in Britain
and Germany. Organization Studies, 17(3), 365-395.
Larzerele, R. E., & Huston, T. L. (1980). The dyadic trust scale: Toward understanding
interpersonal trust in close relationships. Journal of Marriage and Family, 42, 595-604.
LaTour, S. A., & Manrai, A. J. (1989). Interactive impact of informational and normative
influence on donations. Journal of Marketing Research, 26(8), 327-335.
Lau, G. T., & Ng, S. (2001). Individual and situational factors influencing negative word-of-
mouth behavior. Canadian Journal of Administrative Sciences, 18, 63-178.
Law, K. S., Wong, C.-S., & Mobley, H. (1998). Toward a taxonomy of multidimensional
constructs. The Academy of Management Review, 23(4), 741-755.
Lazarsfeld, P. F., & Merton, R. K. (1954). Friendship as a social process: A substantive and
methodological analysis. In M. Berger, T. Abel & C. H. Page (Eds.), Freedom and
contol in modern society (pp. 18-66). New York, NY: Van Nostrand.
Lazarus, R. S. (1984). On the primacy of cognition. American Psychologist, 39, 124-129.

351
Lee, D.-H., Yoon, S. N., Choi, H. Y., & Park, Y. (2011). Electronic word of mouth systems and
acceptane of user-generated contents. International Journal of Information and
Decision Sciences, 3(1), 54-69.
Lee, E.-J. (2002). Factors that enhance consumer trust in human-computer interactions: An
examination of interface factors and the moderating influence. University of Tennessee,
Knoxville.
Lee, E.-J., & Shin, S. Y. (2014). When do consumers buy online product reviews? Effects of
review quality, product type, and reviewer's photo. Computers in Human Behavior,
31(2), 356-366.
Lee, J., Park, D.-H., & Han, I. (2008). The effect of negative online consumer reviews on
product attitude: An information processing view. Electronic Commerce and
Applications, 7(3), 341.
Lee, J., & Soberon-Ferrer, H. (1997). Consumer vulnerability to fraud: Influencing factors.
Journal of Consumer Affairs, 31(1), 70-89.
Lee, J. D., & Neville, M. (1994). Trust, self-confidence, and operators' adaption to automation.
International Journal of Human-Computer Studies, 40, 153-184.
Lee, K.-T., & Koo, D.-M. (2012). Effects of attribute and valence of e-WOM on message
adoption: Moderating roles of subjective knowledge and regulatory focus. Computers
in Human Behavior, 28, 1974-1984.
Lee, L., & Ariely, D. (2006). Shopping goals, goal concreteness, and conditional promotions.
Journal of Consumer Research, 33(June), 60-70.
Lee, M. K., & Turban, E. (2001). A trust model for consumer Internet shopping. International
Journal of Electronic Commerce, 6(1), 75-91.
Lee, S., & Lumking, J. R. (1992). Differences in attitude toward tv advertising: VCR usage as
a moderator. International Journal of Advertising, 11(4), 333-342.
Lennox, R. D., & Wolfe, R. N. (1984). Revision of the self-monitoring scale. Journal of
Personality and Social Psychology, 46, 1349-1369.
Leventhal, H., & Schwerer, K. (1987). The relationship of emotion to cognition: A functional
approach to a semantic controversy. Cognition and Emotion, 1(1), 3-28.
Levin, D. Z., & Cross, R. (2004). The strength of weak ties you can trust: The mediating role
of trust in effective knowledge transfer. Management Science, 50(11), 1477-1490.
Levy, K. (2012). Consumer decision making and word of mouth communication. University of
Waterloo, Waterloo, Ontario, Canada.
Lewicki, R., & Bunker, B. (1996). Developing and maintaining trust in work relationships. In
R. M. Kramer & T. R. Tyler (Eds.), Trust in organizations: Frontiers of theory and
research (pp. 114-139). Thousand Oaks, CA: Sage Publications.
Lewicki, R., McAllister, D. J., & Bies, R. J. (1998). Trust and distrust: New relationships and
realities. Academy of Management Review, 23(3), 438-458.
Lewicki, R. J., & Bunker, B. B. (1995). Trust in relationships: A model of trust development
and decline. In B. B. Bunker & J. Z. Rubing (Eds.), Cooperation and justice. San
Francisco, CA: Jossey-Bass.
Lewis, D. J., & Weigert, A. (1985). Trust as social reality. Social Forces, 63(4), 967-985.
Lewis, J. D., & Weigert, A. J. (1985). Social atomism, holism, and trust. The Sociological
Quarterly, 26(4), 455-471.
Li, F., Zhou, J. N., Kashyap, R., & Yang, Z. (2008). Brand trust as a second-order factor: An
alternative measurement model. International Journal of Market Research, 50(6), 1-23.
Li, X., & Hitt, L. M. (2008). Self-selection and information role of online product reviews.
Information Systems Research, 19(4), 456-474.
Lichtenstein, D. R., Ridgway, N. M., & Netemeyer, R. G. (1993). Price perceptions and
consumer shopping behavior: A field study. Journal of Marketing Research, 30(5), 234-
245.
352
Liebermann, J. K. (1981). The litigious society. New York, NY: Basic Books.
Lietz, P. (2010). Research into questionnaire design: A summary of the literature. International
Journal of Market Research, 52(2), 249-272.
Lightspeed Research. (2011). Consumers rely on online reviews and price comparisons to make
purchase decisions.
Lin, C. H., Sher, P. J., & Shih, H. Y. (2005). Past progress and future directions in
conceptualizing customer perceived value. International Journal of Service Industry
Management, 16(4), 318-336.
Lin, J.-S. C., & Hsieh, P.-L. (2011). Assessing the self-service technology encounters:
Development and validation of SSTQUAL scale. Journal of Retailing, 87(2), 194-206.
Lindskold, S. (1978). Trust development, the GRIT proposal, and the effects of conciliatory
acts on conflict and cooperation. Psychological Bulletin, 85(4), 772-793.
Ling, L.-L., & Liu, Y.-F. (2008). Comprehension and persuasion on advertising message:
Heuristic-systematic model approach. Journal of Management, 25(5), 487-503.
Lis, B. (2013). In eWOM we trust: A framework of factors that determine the eWOM
credibility. Business & Information Systems Engineering, 3, 129-140.
Litvin, S. W., Goldsmith, R. E., & Pan, B. (2008). Electronic word-of-mouth in hospitality and
ourism management. Tourism Management, 29, 458-468.
Liu, R. R., & Zhang, W. (2010). Informational influence of online customer feedback: An
empirical study. Database Marketing & Customer Strategy Management, 17(2), 120-
131.
Liu, Y. (2003). Developing a scale to measure the interactivity of websites. Journal of
Advertising Research, 43(2), 207-216.
Liu, Y. (2006). Word of mouth for movies: Its dynamics and impact on box office revenue.
Journal of Marketing, 70(July), 74-89.
Locander, W. B., & Herman, P. W. (1979). The effect of self-confidence and anxiety on
information seeking in consumer risk reduction. Journal of Marketing Research, 16(2),
268-274.
Loevinger, J. (1957). Objective tests as instruments of psychological theory. Psychological
Reports, 3, 635-694.
Lord, F. M., & Novick, M. R. (1968). Staticstical theories of mental test scores. Reading, MA:
Addison-Wesely.
Lord, K. R., Lee, M.-S., & Choong, P. (2001). Differences in normative and informational
social influence. In J. Meyers-Levy & M. C. Gilly (Eds.), Advances in Consumer
Research (pp. 280-285). Salt Lake City, UT: Association for Consumer Research.
Lowry, P. B., Vance, A., Moody, G., Beckman, B., & Read, A. (2008). Explaining and
predicting the impact of branding alliances and web site quality on initial consumer trust
of e-commerce websites. Journal of Management Information Systems, 24(4), 199-224.
Lu, Y., Zhao, L., & Wang, B. (2010). From virtual community members to C2C e-commerce
buyers: Trust in virtual communities and its effect on consumers' purchase intention.
Electronic Commerce and Applications, 9, 346-360.
Lu, Y., & Zhou, T. A. (2007). A research of consumer's initial trust in online stores in China.
Journal of Research and Practice in Information Technology, 39(3), 167-180.
Luhmann, N. (1979). Trust and power. Chichester, UK: John Wiley and Sons.
Luhmann, N. (1988). Familiarity, confidence, trust: Problems and alternatives. In D. G.
Gambetta (Ed.), Trust. New York, NY: Basil Blackwell.
Luo, C., Luo, X., & Schatzberg, L. (2013). Impact of informational factors on online
recommendation credibility: The moderating role of source credibility. Decision
Support Systems, Forthcoming.

353
Lutz, R. (1985). Affective and cognitive antecedents of attitude towards the ad: A conceptual
framewrok. In L. Allowed & A. Mitchell (Eds.), Psychological Processes and
Advertising Effects: Theory, Research, and Applications. Hillsdale, NJ.
Lynch, J., & Ariely, D. (2000). Wine online: search costs affect competition on price, quality
and distribution. Marketing Science, 19, 83-103.
Lysonski, S., Durvasula, S., & Zotos, Y. (1996). Consumer decision-making styles: a multi-
country investigation. European Journal of Marketing, 30(12), 10-21.
MacCallum, R. C., Widaman, K. F., Zhang, S., & Hong, S. (1999). Sample size in factor
analysis. Psychological Methods, 4(1), 84-99.
MacCrimmon, K. R., & Wehrung, D. A. (1986). Taking risks: The management of uncertainty.
New York, NY: Free Press.
MacCrimmon, K. R., & Wehrung, D. A. (1990). Characteristics of risk taking executives.
Management Science, 36(4), 422-435.
MacDonald, K. E., & Uncles, M. D. (2007). Consumer savvy: Conceptualisation and
measurement. Journal of Marketing Management, 23(5/6), 497-517.
Mack, R. W., Blose, J. E., & Pan, B. (2008). Believe it or not: Credibility of blogs in tourism.
Journal of Vacation Marketing, 14(2), 133-143.
MacKenzie, S. B., & Lutz, R. J. (1989). An empirical examination of the structural antecedents
of attitude toward the ad in an advertising pretesting context. Journal of Marketing,
53(April), 48-65.
MacKenzie, S. B., Podsakoff, P. M., & Jarvis, C. B. (2005). The problem of measurement
model misspecification in behavioral and organizational research and some
recommended solutions. Journal of Applied Psychology, 90(4), 710-730.
MacKenzie, S. B., Podsakoff, P. M., & Podsakoff, N. P. (2011). Construct measurement and
validation procedures in MIS and behavioral research: Integrating new and existing
techniques. MIS Quarterly, 35(2), 293-334.
Mackiewicz, J. (2007). Reviewer bias and credibility in online reviews. Proceedings of the 2007
Association for Business Communication Annual Convention.
MacQueen, J. B. (1967). Some methods for classification and analysis of multivariate
observations. Paper presented at the Proceedings of the 5th Berkeley symposium of
mathematical statistics and probability, Berkeley, CA.
Macy, M. W., & Skvoretz, J. (1998). The evolution of trust and cooperation between strangers:
a computational model. American Sociological Review, 63(10), 638-660.
Mady, T. T. (2011). Sentiment toward marketing: Should we care about consumer alienation
and readiness to use technology? Journal of Consumer Behaviour, 10, 192-204.
Malhotra, N., & Peterson, M. (2006). Basic marketing research: A decision-making approach.
New Jersey: Prentice Hall.
Mangleburg, T. F., & Bristol, T. (1998). Socialization and adolescents' skepticism toward
advertising. Journal of Advertising, 27(3), 11-21.
Mangleburg, T. F., Doney, P. M., & Bristol, T. (2004). Shopping with friends and teens'
susceptibility to peer influence. Journal of Retailing, 80, 101-116.
March, J. G., & Shapira, Z. (1987). Managerial perspectives on risk and risk taking.
Management Science, 33(11), 1404-1418.
Markus, H. R., & Kitayama, S. (1991). Culture and the self: Implications for cognition,
emotion, and motivation. Psychological Review, 98(2), 224-253.
Marsh, H. (1995). Confirmatory factor analysis models of facotrial invariance: A multifaceted
approach. Structural Equation Modeling, 1, 5-34.
Marsh, H., & Hocevar, D. P. (1985). The application of confirmatory factor analysis to the
study of self-concept: First and higher order factor structures and their invariance across
age groups. Psychological Bulletin, 97, 562-582.

354
Martens, M. P. (2005). The use of structural equation modeling in counseling psychology
research The Counseling Psychologist, 33, 269-298.
Martens, M. P., & Hasse, R. F. (2006). Advanced applications of structural equation modeling
in counseling psychology research. The Counseling Psychologist, 34, 878-911.
Martin, W. C., & Lueg, J. E. (2013). Modeling word-of-mouth usage. Journal of Business
Research, 66(7), 801-808.
Matzler, K. W., Würtele, A., & Renzl, B. (2006). Dimensions of price satisfaction: A study in
the retail banking industry. International Journal of Bank Marketing, 24(4), 216-231.
Maute, M. F., & Dubé, L. (1999). Patterns of emotional response and behavioural consequences
of dissatisfaction. Applied Psychology: An International Review, 48(3), 349-356.
Mayer, R. C., & Davis, J. H. (1999). The effect of the performance appraisal system on trust
for management: A field quasi-experiment. Journal of Applied Psychology, 84(1), 123-
136.
Mayer, R. C., Davis, J. H., & Schoorman, F. D. (1995). An integrative model of organizational
trust. Academy of Management Review, 20(3), 709-734.
Mayzlin, D. (2006). Promotional chat on the Internet. Marketing Science, 25(2), 155-163.
McAlister, L., & Pessemier, E. (1982). Variety seeking behavior: An interdisciplinary review.
Journal of Consumer Research, 9(3), 311.
McAllister, D. J. (1995). Affect- and cognition-based trust as foundations for interpersonal
cooperation in organizations. Academy of Management Journal, 38(1), 24-59.
McCormick, M., & Martinko, M. (2004). Identifying leadersocial cognitions: Integrating the
causal reasoning perspetive into social cognitive theory. Journal of Leadership &
Oganizational Studies, 10(4), 2-9.
McCracken, G. (1989). Who is the celebrity endorser? Cultural foundations of the endorsement
process. Journal of Consumer Research, 16, 310-321.
McCroskey, L. L., McCroskey, J. C., & Richmond, V. P. (2006). Analysis and improvement of
the measurement of interpersonal attraction and homophily. Communication Quarterly,
54(1), 1-31.
McDonald, M. (2010). A brief review of marketing accountability and a research agenda. The
Journal of Business and Industrial Marketing, 25(5), 383-394.
McEvily, B., Perrone, V., & Zaheer, A. (2003). Trust as an organization principle. Organization
Science, 14(1), 91-103.
McGinnies, E., & Ward, C. D. (1980). Effects of source expertness, phyiscal attractiveness and
supporting arguments on persuation: A case of brains over beauty. Journal of
Personality and Social Psychology, 39(2), 235-244.
McGregor, D. (1967). The professional manger. New York, NY: McGraw-Hill.
McGuigan, D. (2008). 70% of online shoppers read multiple product reviews. Retrieved 2013-
08-06, 2013, from [Link]
read-multiple-product-reviews/
McGuire, W. J. (1968). Personality and susceptibility to social influence. In E. F. Borgatta &
W. W. Lambert (Eds.), Handbook of personality theory and research (pp. 1130-1187):
Rand McNally.
McKnight, D. H., & Chervany, N. L. (2001). Trust and distrut definitions: One bite at a time.
In R. Falcone, M. Singh & Y.-H. Tan (Eds.), Trust in Cyber-societies (pp. 27-54).
Berling-Heidelberg: Springer.
McKnight, D. H., & Chervany, N. L. (2002). What trust means in e-commerce customer
relationships: An interdisciplinary conceptual typology. International Journal of
Electronic Commerce, 6(2), 35-59.
McKnight, D. H., & Chervany, N. L. (2006). Reflections on an initial trust-building model. In
R. Bachmann & A. Zaheer (Eds.), Handbook of trust research (pp. 29-51): Edward
Edgar Publishing.
355
McKnight, D. H., & Chevany, N. L. (1996). The meaning of trust. Working paper. University
of Minnesota.
McKnight, D. H., Choudhury, V., & Kacmar, C. (2000). Trust in e-commerce vendors: A two-
stage model. Proceedings of the Twenty First International Conference on Information
Systems, 532-536.
McKnight, D. H., Choudhury, V., & Kacmar, C. (2002a). Developing and validating trust
measures for e-commerce: An integrative typology. Information Systems Research,
13(3), 334-359.
McKnight, D. H., Choudhury, V., & Kacmar, C. (2002b). The impact of initial consumer trust
with a web site: A trust building model. Journal of Strategic Information Systems, 11,
297-323.
McKnight, D. H., Cummings, L. L., & Chervany, N. L. (1998). Initial trust formation in new
organization relationships. The Academy of Management Review, 23(3), 473-490.
McKnight, D. H., Kacmar, C., & Choudhury, V. (2004). Dispositional trust and distrust
distinctions in predicting high- and low-risk Internet expert advice site perceptions. E-
Service Journal, 3(2), 35-58.
McKnight, D. H., & Kacmar, C. J. (2007). Factors and effects of information credibility.
Communication of the ACM, August, 423-432.
McLain, D. L., & Hackman, Z. K. (1999). Trust, risk, and decision-making in organizational
change. Public Administration Quarterly, 23(2), 152-176.
Mediapost. (2010). Online product reviews gain credibility.
Mehta, A. (2000). Advertising attitudes and advertising effectiveness. Journal of Advertising,
5, 67-72.
Mehta, A., & Purvis, S. C. (1995). When attitudes toward advertising in general influence
advertising success. Paper presented at the Conference of the American Academy of
Advertising, Norfolk, VA.
Menon, A. M., Deshpande, A. D., Perri III, M., & Zinkhan, G. M. (2002). Trust in online
prescription drug information among Internet users: The impact on information search
behavior after exposure to direct-to-consumer advertising. Health Marketing Quarterly,
20(1), 17-35.
Menon, N. M., Konana, N. M., Browne, G. J., & Balasubramanian, S. (1999). Understanding
trustworthiness beliefs in electronic brokerage usage. In P. De & J. I. DeGross (Eds.),
Proceedings of the Twentieth International conference on Information Systems (pp.
552-555). Charlotte, NC: Omnipress.
Merriam-Webster's New Encyclopedic Dictionary. (Ed.) (1996). New York, NY: Black Dog &
Leventhal Publishers.
Metzger, M. J., Flanagin, A. J., Eyal, K., & Lemus, D. R. (2003). Credibility for the 21st
century: Integrating perspectives on source, message, and media credibility in the
contemporary media environment. Communication Yearbook, 27, 293-335.
Meyer, P. (1988). Defining and measuring credibility of newspapers: Developing an index.
Journalism Quarterly, 65, 567-574.
Meyers-Levy, J., & Maheswaran, D. (1991). Exploring differences in males' and females'
processing strategy. Journal of Consumer Research, 18(1), 63-70.
Meyers-Levy, J., & Sternthal, B. (1991). Gender differences in the use of message cues and
judgments. Journal of Marketing Research, 28(1), 84-96.
Miniard, P. W., & Joel, J. B. (1983). Modeling personal and normative influences on behavior.
Journal of Consumer Research, 10(September ), 169-180.
Mish, F. C. E. (1993). Merriam Webster's Collegiate Dictionary. Springfield, MA.
Mishra, A. (1996). Organizational responses to crises: The centrality of trust. International
Journal of Management, 20(2), 125-138.

356
Misztal, B. A. (1996). Trust in modern societies: The search for the bases of social order. New
York, NY: Polity Press.
Mitchell, A. A., & Olson, J. C. (1981). Are product attribute beliefs the only mediator of
advertising effects on brand attitude? Journal of Marketing Research, 18(August), 318-
322.
Mitchell, V. W. (1999). Consumer perceived risk: Conceptualizations and models. European
Journal of Marketing, 33(1/2), 163-195.
Mittal, V., Huppertz, J. W., & Khare, A. (2008). Customer complaining: The role of tie strength
and information control. Journal of Retailing, 84(2), 195-204.
Moe, J. (2012). Can you really trust online customer reviews, Marketplace. Retrieved from
[Link]
Mohr, J. I., & Spekman, r. E. (1994). Characteristics of partnership success: Parntership
attributes, communication behavior and conflict resolution technique. Strategic
Management Journal, 57(January), 135-152.
Mohr, L. A., Eroglu, D., & Ellen, P. S. (1998). The development and testing of a measure of
skepticism toward environment claims in marketers' communication. Journal of
Consumer Affairs, 32, 30-55.
Möllering, G., Bachmann, r., & Lee, S. H. (2004). Understanding organizational trust -
foundations, constellations, and issues of operationalisation. Journal of Managerial
Psychology, 19(6), 556-570.
Money, R. B., Gilly, M. C., & Graham, J. L. (1998). Explorations of national culture and word-
of-mouth referral behavior in the purchase of industrial services in the United States and
Japan. Journal of Marketing, 62, 76-87.
Moorman, C., Deshpande, R., & Zaltman, G. (1993). Factors affecting trust in market research
relationships. Journal of Marketing, 57, 81-101.
Moorman, C., Zaltman, G., & Deshpande, R. (1992). Relationships between providers and users
of market research: The dynamics of trust within and between organizations. Journal of
Marketing Research, 29(August), 314-328.
Morgan, N. A., Slotegraff, R. J., & Vorhies, D. W. (2009). Linking marketing capabilities with
profit growth. International Journal of Research in Marketing, 26(4), 284-293.
Morgan, R. M., & Hunt, S. D. (1994). The commitment-trust theory of relationship marketing.
Journal of Marketing, 58(3), 20-38.
Morrow, J. L., Hansen, M. H., & Pearson, A. W. (2004). The cognitive and affective
antecedents of general trust within cooperative organizations. Journal of Managerial
Issues, 16(1), 48-64.
Morse, B. J., Gullekson, N. L., Morris, S. A., & Popovich, P. M. (2011). The development of
an Internet attitude scale. Computers in Human Behavior, 27(1), 480-489.
Mourali, M., Laroche, M., & Pons, F. (2005). Antecedents of consumer relative preference for
interpersonal information sources in pre-purchase search. Journal of Consumer
Behaviour, 4(5), 307-318.
Mudambi, S. M., & Schuff, D. (2010). What makes a helpful online review? A study of
customer reviews on [Link]. MIS Quarterly, 34(1), 185-200.
Muehling, D. D. (1987). An investigation of factors underlying attitudes-towards-advertising-
in-general. Journal of Advertising, 16(1), 32-40.
Muir, B. M. (1994). Trust in automation: Part I, theoretical issues in the study of trust and
human intervention in automated systems. Ergonomics, 39(3), 429-460.
Murphy, G. B. (2003). Propensity to trust, purchase experience, and trusting beliefs of
unfamiliar e-commerce ventures. New England Journal of Entrepreneurship, 6(2), 53-
64.

357
Murphy, G. B., & Tocher, N. (2011). Gender differences in the effectiveness of online trust
building information cues: An empirical examination. Journal of High Technology
Management Research, 22, 26-35.
Murtha, T. C., Kanfer, R., & Ackerman, P. L. (1996). Toward an interactionist taxonomy of
personality and situations: An integrative situational-dispositional representation of
personality traits. Journal of Personality and Social Psychology, 71, 193-207.
Mynatt, C., Dohety, M., & Tweeny, R. (1977). Confirmation bias in a simulated research
environment: An experimental study of scientific inference. Quarterly Journal of
Experimental Psychology, 29.
Nahapiet, J., & Ghoshal, S. (1998). Social capital, intellectual capital, and the organizational
advantage. Academy of Management Review, 23(2), 242-266.
Nan, X. (2009). The influence of source credibility on attitude certainty: Exploring the
moderating effects of timing of source identification and individual need for cognition.
Psychology & Marketing, 26(4), 321-332.
Nardi, B. A., Schiano, D. J., Gumbrecht, M., & Swartz, L. (2004). Why we blog.
Communications of the ACM, 47(12), 41-46.
Nederhof, A. (1985). Methods of copyin with social desirability bias: A review. European
Journal of Social Psychology, 15(July/September), 263-280.
Nelson, P. (1974). Advertising as information. Journal of Political Economy, 82, 729-754.
Netemeyer, R., Bearden, W., & Sharma, S. (2003). Scaling procedures. Thousand Oaks: Sage
Publications.
Netemeyer, R., Boles, J. S., & McMurrian, R. (1996). Development and validation of work-
family conflict and family-work conflict scales. Journal of Applied Psychology, 81(4),
400-410.
Netemeyer, R., Burton, S., & Lichtenstein, D. R. (1995). Trait aspects of vanity: Measurement
and relevance to consumer behavior. Journal of Consumer Research, 21(4), 612-626.
Nevo, B. (1985). Face validity revisited. Journal of Educational Measurement, 22, 287-293.
Newell, S. J., & Goldsmith, R. E. (2001). The development of a scale to measure perceived
corporate credibility. Journal of Business Research, 52, 235-247.
Nielsen. (2012). Global trust in advertising and brand messages.
Nissenbaum, H. (2001). Securing trust online: Wisdom, or oxymoron? Boston University Law
Review, 81, 101-131.
Nooteboom, B. (1996). Trust, opportunism and governance: A process and control model.
Organization Studies, 17(6), 985-1010.
Nord, W. P., & Peter, J. P. (1980). A behavior modification perspective on marketing. Journal
of Marketing Research, 44(Spring), 36-47.
Norman, W. T. (1963). Toward an adequate taxonomy of personality attributes: Replicated
factor structure in peer nomination personality ratings. Journal of Abnormal Social
Psychology, 66, 574-583.
North, D. C. (1980). Institutions, institutional change, and economic performance. New York,
NY: Cambridge University Press.
Nöteberg, A., Christiaanse, E., & Wallage, P. (2003). Consumer trust in electronic channels.
eService Jorunal, 2(2), 46-67.
Nunnally, J. C. (1978). Psychometric theory. New York, NY: McGraw-Hill.
Nunnally, J. C., & Berstein, I. H. (1994). Psychometric theory. New York, NY: McGraw-Hill.
O'Cass, A. (2002). Political advertising believability and information source value during
election. Journal of Advertising, 26(1), 63-74.
Obermiller, C., & Spangenberg, E. R. (1998). Development of a scale to measure consumer
skepticism toward advertising. Journal of Consumer Psychology, 7(2), 159-186.
Obermiller, C., & Spangenberg, E. R. (2000). On the origin and distinctiveness of skepticism
toward advertising. Marketing Letters, 11, 311-322.
358
Obermiller, C., Spangenberg, E. R., & MacLachlan. (2005). Ad skepticism: The consequences
of disbelief. Journal of Advertising, 34(3), 7-17.
Ogonowski, A., Montandon, A., Botha, E., & Reyneke, M. (2014). Should new online stores
invest in social presence elements? The effect of social presence on initial trust
formation. Journal of Retailing and Consumer Services, 21(4), 482-491.
Ohanian, R. (1990). Construction and validation of a scale to measure celebrity endorsers'
üerceived expertise, trustworthiness, and attractiveness. Journal of Advertising, 19(3),
39-52.
Ohanian, R. (1991). The impact of celebrity spokespersons' perceived image on consumers'
intention to purchase. Journal of Advertising Research, 31, 46-54.
Olney, T. J., Holbrook, M. B., & Batra, R. (1991). Consumer responses to advertising: The
effects of ad content, emotions, and attitude toward the ad on viewing time. Journal of
Consumer Research, 17(4), 440-453.
Olshavsky, R. W., & Granbois, D. B. (1979). Consumer decision making - fact or fiction?
Journal of Consumer Research, 6, 93-100.
Omodei, M. M., & McLennan, J. (2000). Conceptualizing and measuring global interpersonal
mistrust-trust. Journal of Social Psychology, 140(3), 279-294.
Ostrom, T. M. (1968). The relationship between the affective, behavioural, and cognitive
components of attitude. Journal of Experimental Social Psychology, 5, 12-30.
Ou, W.-M., Abratt, R., & Dion, P. (2006). The influence of retailer reputation on store
patronage. Journal of Retailing and Consumer Services, 13(2), 221-230.
Ouchi, W. G. (1981). Theory Z: How American business can meet the Japanese challenge.
Reading, MA: Addison-Wesley.
Paliszkiewicz, J., & Klepacki, B. (2013). Tools of building customer trust. Warsaw, PL.
Palmer, A., & Bejou, D. (1994). Buyer-seller relationships: A conceptual model and empirical
investigation. Journal of Marketing Management, 10, 495-512.
Palmer, C. T. (1991). Kin-selection, reciprocal altruism, and information sharing among Maine
lobstermen. Ethology and Sociobiology, 12, 221-235.
Palvia, P. (2009). The role of trust in e-commerce relational exchange: A unified model.
Information & Management, 46, 213-220.
Pan, L.-Y., & Chiou, J.-S. (2011). How much can you trust online information? Cues for
perceived trustworthiness of consumer-generated online information. Journal of
Interactive Marketing, 25, 67-74.
Panteli, N., & Sockalingam, S. (2005). Trust and conflict within virtual inter-organizational
alliances: A framework for facilitating knowledge sharing. Decision Support Systems,
39(4), 599-617.
Park, C., & Lee, T. M. (2009). Antecedents of online reviews' usage and purchase influence:
An empirical comparison of U.S. and Korean consumers. Journal of Interactive
Marketing, 23(4), 332-340.
Park, C., & Lessig, V. P. (1977). Students and housewives: Differences in susceptibility to
reference group influence. Journal of Consumer Research, 4(2), 102-110.
Park, C., Wang, Y., Yao, Y., & Kang, Y. R. (2011). Factors influencing eWOM effects: Using
experience,credibility, and susceptibility. International Journal of Social Science and
Humanity, 1(1), 74-79.
Park, C. W., Mothersbaugh, D. L., & Feick, L. F. (1994). Consumer knowledge assessement.
Journal of Consumer Research, 21(June), 71-82.
Park, D. H., & Kim, S. (2008). The effects of consumer knowledge on message processing of
electronic word-of-mouth via online consumer reviews. Electronic Research and
Applications, 7(4), 399.

359
Park, D. H., Lee, J., & Han, I. (2007). The effect of on-line consumer reviews on consumer
purchasing intention: The moderating role of involvement. International Journal of
Electronic Commerce, 11(4), 125-148.
Park, J.-K., Chung, H.-E., & Yoo, W. S. (2009). Is the internet a primary source of consumer
information search? Group comparison for channel choices. Journal of Retailing and
Consumer Services, 16, 92-99.
Park, J. H., Gu, B., & Lee, H. Y. (2012). The relationship between retailer-hosted and third-
party hosted WOM sources and their influence on retailer sales. Electronic Commerce
Research and Applications, 11(3), 253-261.
Parson, T. (1970). Research with human subjects and the professional complex. In P. Freund
(Ed.), Experimentation with human subjects: Braziller.
Pavlou, P. A. (2003). Consumer acceptance of electronic commerce: Integrating trust and risk
with the technology acceptance model. International Journal of Electronic Commerce,
7(3), 101-134.
Pavlou, P. A., & Chai, L. (2002). What drives electronic commerce across cultures? A cross-
cultural emprical investigation of the theory of planned behavior. Journal of Electronic
Commerce Research, 3(4), 240-253.
Pavlou, P. A., & Fygenson, M. (2006). Understanding and predicting electronic commerce
adoption: An extension of the theory of planned behavior. MIS Quarterly, 30(1), 115-
143.
Pavlou, P. A., & Gefen, D. (2004). Building effective online marketplaces with institution-
based trust. Information Systems Research, 15(1), 37-59.
Pearce, W. B. (1974). Trust in interpersonal relationships. Speech Monographs, 41(3), 236-244.
Pervin, L. A., Cervone, D., & John, O. P. (2004). Personality: Theory and research. Chichester,
UK: John Wiley and Sons.
Peter, J. P. (1981). Construct validity: A review of basic issues and marketing practices. Journal
of Marketing Research, 18(May), 133-145.
Petermann, F. (1996). Psychologie des Vertrauens (3rd Edition ed.). Göttingen: Hogrefe.
Peters, R. G., Covello, V. T., & McCallum, D. B. (1997). The determinants of trust and
credibility in environmental risk communication: An empirical study. Risk Analysis,
17(1), 43-54.
Peters, T. (1992). Liberation management: Necessary disorganizational for the nanosecond
ninties. New York, NY: Alfred A. Knopf.
Peterson, A., & Merino, M. (2003). Consumer information search behavior and the Internet.
Psychology and Marketing, 20(2), 99-121.
Peterson, R. A. (1997). Electronic Marketing: Visions, Definitions, and Implications Electronic
Marketing and the Consumer (pp. 1-16). Thousand Oaks, CA: Sage.
Peterson, R. A. (2000). Constructing effective questionnaires. Thousand Oaks, CA: Sage
Publications.
Petrovici, D., & Marinov, M. (2007). Determinants and antecedents of general attitudes towards
advertising: A study of two EU accession countries. European Journal of Marketing,
41(3/4), 307-326.
Petter, S., Straub, D. W., & Rai, A. (2007). Specification and validation of formative constructs
in IS research. MIS Quarterly, 31(4), 623-656.
Petty, R. E., & Cacioppo, J. T. (1981). Attitude and persuasion: classic and contemporary
approaches. Dubuque, MA: C. Brown.
Petty, R. E., & Cacioppo, J. T. (1986). Communication and persuasion: central and peripheral
routes of attitude change. New York, NY: Springer.
Phillips, L., & Sternthal, B. (1977). Age differences in information processing: A perspective
on the aged consumer. Journal of Marketing Research, 14(November), 444-457.

360
Picazo-Vela, S., Chou, S. Y., Melcher, A. J., & Pearson, J. M. (2010). Why provide an online
review? An extended theory of planning behaviorand the role of big-five personality
traits. Computers in Human Behavior, 26, 685-696.
Piccoli, G., & Ives, B. (2003). Trust and the unintended effects of behavior control in virtual
teams. MIS Quarterly, 27(3), 365-395.
Pigg, K. E., & Crank, L. D. (2004). Building communicty social capital: The potential and
promise of information and communications technologies. Journal of Community
Informatics, 1(1), 58-73.
Pilon, M. (2009). Tracking down fake Amazon reviews. The Wall Street Journal.
[Link]
Pirson, M. (2007). Facing the trust gap: Measuring and building trust in organizations.
University of St. Gallen, St. Gallen, CH.
Podsakoff, P. M., MacKenzie, S. B., Lee, J. Y., & Podsakoff, N. P. (2003). Common method
biases in behavioral research: A critical review of the literature and recommended
remedies. Journal of Applied Psychology, 88(5), 879-903.
Podsakoff, P. M., MacKenzie, S. B., Moorman, R. H., & Fetter, R. (1990). Transformational
leader behaviors and their effects on followers' trust in leader, satisfaction, and
organizational citizenship behaviors. Leadership Quarterly, 1, 107-142.
Podsakoff, P. M., MacKenzie, S. B., & Podsakoff, N. P. (2003). The mismeasure of man
(agement) and its implications for leadership research. Leadership Quarter, 14, 615-
656.
Pollay, R. W., & Banwari, M. (1993). Here's the beef: Factors, determinants, and segments in
consumer criticism of advertising. Journal of Marketing, 57(3), 99-114.
Pollay, R. W., & Mittal, B. (1993). Here's the beef: Factors, determinants and segments in
consumer criticism of advertising. Journal of Marketing, 57(3), 99-114.
Pookulangara, S., Hawley, J., & Xiao, G. (2011). Explaining consumers' channel-switching
behavior using the theory of planned behavior. Journal of Retailing and Consumer
Services, 18, 311-321.
Pornpitakpan, C. (2004a). Factors associated with opinion seeking: A cross-national study.
Journal of Global Marketing, 17(2/3), 91-113.
Pornpitakpan, C. (2004b). The persuasiveness of source credibility: A critical review of five
decades' evidence. Journal of Applied Social Psychology, 34, 243-281.
Power, J., Whelan, S., & Davies, G. (2008). The attractiveness and connectedness of ruthless
brands: The role of trust. European Journal of Marketing, 42(5/6), 586-602.
PowerReviews. (2010). 5 social shopping trends.
Price, L. L., & Feick, L. F. (1984). The role of interpersonal sources in external search: An
informational perspective. Advances in Consumer Research, 11, 250-255.
Price, L. L., Feick, L. F., & Guskey, A. (1995). Everyday market helping behavior. Journal of
Public Policy & Marketing, 14, 225-266.
Price, L. L., Feick, L. F., & Higie, R. A. (1987). Preference heterogeneity and co-orientatiion
as determinants of referent influence in the choice of service providers. Working Paper.
University of Pittsburgh. Pittsburgh.
Pruden, H. O., Shuptrine, F. K., & Longman, D. S. (1974). A measure of alienation from the
marketplace. Journal of the Academy of Marketing Science, 2(4), 610-619.
Pruitt, D. G. (1981). Negotation behavior. New York, NY: Academic Press.
Punj, G., & Stewart, D. W. (1983). Cluster analysis in marketing research: Review and
suggestions for applications. Journal of Marketing Research, 20(2), 134-148.
Putnam, R. D. (1993). Makind democracy work. Civic traditions in modern Italy. Princeton,
NJ: Princeton University Press.

361
Qiu, L., Pang, J., & Lim, K. H. (2012). Effects of conflicting aggregated rating on eWOM
review credibility and diagnosticity: The moderating role of review valence. Decision
Support Systems, 54, 631-643.
Rabjohn, N., Cheung, C. M. K., & Lee, M. K. O. (2008). Examining the perceived credibility
of online opinions: Information adoption in the online environment. Proceedings of the
41th Hawaii International Conference on System Sciences.
Racherla, P., Mandviwalla, M., & Connolly, D. J. (2012). Factors affecting consumers' trust in
online product reviews. Journal of Consumer Behaviour, 11, 94-104.
Rackman, N., Friedman, L. g., & Ruff, R. (1996). Getting partnering right: How market leaders
are creating long term competitive advantage. New York, NY: McGraw-Hill.
Radecki, C. M., & Jaccard, J. (1995). Perceptions of knowledge, actual knowledge, and
information search behavior. Journal of Experimental Social Psychology, 31, 107-138.
Raden, D. (1985). Strength-related attitude dimensions. Social Psychology Quarterly, 48, 312-
330.
Rafaeli, S., & Raban, D. R. (2005). Information sharing online: A research challenge.
International Journal of Knowledge and Learning, 1(2), 62-79.
Rahn, W. M., & Transue, J. E. (1998). Social trust and value change: The decline of social
capital in American youth, 1976-1995. Political Psychology, 19(3), 545-565.
Ramaswami, S. N., Srinivasan, S. S., & Gorton, A. S. (1997). Information asymmetry between
salesperson and supervisor: Postulates from agency and social exchange theories.
Journal of Personal Selling and Sales Management, 17(3), 29-50.
Ranaweera, C., & Prabhu, J. (2003). On the relative importance of customer satisfaction and
trust as determinants of customer retention and positive word of mouth. Journal of
Targeting, Measurement and Analysis for Marketing, 12(1), 82-90.
Ranjbarian, B., & Kia, M. (2010). The influence of personality traitson consideration set size.
European Journal of Social Sciences, 15(2), 263-269.
Rappaport, S. D. (2007). Lessons from online practice: New advertising models. Journal of
Advertising Research, 47(2), 135-141.
Ratchford, B. T., Lee, M.-S., & Talukdar, D. (2003). The impact of the Internet on information
search for automobiles. Journal of Marketing Research, 40(May), 193-209.
Ratchford, B. T., Talukdar, D., & Lee, M.-S. (2001). A model of consumer choice of the
Internet as an information source. International Journal of Electronic Commerce,
5(Spring), 7-21.
Ratnasingham, P., & Pavlou, P. A. (2003). Technology trust in Internet-based
interogrganizational electronic commerce. Journal of Electronic Commerce in
Organizations, 1(1), 17-41.
Regan, P. C., & Berscheid, E. (1999). Lust: What we know about human sexual desire.
Thousand Oaks, CA: Sage.
Reichheld, F. F., & Schefter, P. (2000). E-loyalty: Your secret weapon on hte web. Harvard
Business Review, 78(4), 103-113.
Reinstein, D. A., & Snyder, C. M. (2005). The influence of expert reviews on consumer demand
for experience goods: A case study of movie critics. Journal of Industry Economics,
53(1), 27-51.
Rempel, J. K., Holmes, J. G., & Zanna, M. P. (1985). Trust in close relationships. Journal of
Personality and Social Psychology, 49(1), 95-112.
Reutterer, T., Mild, A., Natter, M., & Taudes, A. (2006). A dynamic segmentation approach for
targeting and customizing direct marketing campaigns. Journal of Interactive
Marketing, 20(3/4), 43-57.
Reynolds, F. D., & Darden, W. R. (1971). Mutually adaptive effects of interpersonal
communication. Journal of Marketing Research, 8(4), 449-454.

362
Rheingold, H. (1993). The virtual community: Homestaeding on the electronic frontier. New
York, NY: Harper Perennial.
Rhodes, N., & Wood, W. (1992). Self-esteem and intelligence affect influenceability: The
mediating role of message reception. Psychological Bulletin, 111, 156-171.
Richins, M. L., & Root-Shaffer, T. (1988). The role of involvement and opinion leadership in
consumer word of mouth: An implicit model made explicit. Advances in Consumer
Research, 15(1), 32-36.
Ridings, C. M., Gefen, D., & Arinze, B. (2002). Some antecedents and effects of trust in virtual
communities. Journal of Strategic Information Systems, 11, 271-295.
Riefler, P., Diamantopoulos, A., & Siguaw, J. A. (2012). Cosmopolitan consumers as a target
group for segmentation. Journal of International Business Studies, 43, 285-305.
Riegelsberger, J., Sasse, M. A., & McCarthy, J. D. (2003). The researcher's dilemma:
Evaluating trut in computer-mediated communication. International Journal of Human-
Computer Studies, 58(6), 759-781.
Rieh, S. Y., & Danielson, D. R. (2007). Credibility: A multidisciplinary framework. Annual
Review of Informatiion Science and Technology, 41(1), 307-364.
Ring, P. S. (1996). Fragile and resilient trust and their roles in economic exchange. Business &
Society, 35, 148-175.
Ring, P. S., & Van de Ven, A. H. (1994). Developmental process of cooperative
interorganizational relationships. Academy of Management Review, 19(1), 90-118.
Ring, P. S., & Ven, A. v. d. (1992). Structuring cooperative relationships between
organizations. Strategic Management Journal, 13, 483-498.
Ripperger, T. (1998). Ökonomik des Vertrauens. Analyse eines Organisationsprinzips.
Tübingen: Mohr Siebeck.
Robinson, J. P., Shaver, P. R., & Wrightsman, L. S. (1991). Criteria for scale selection and
evaluation. In J. P. Robingson, P. R. Shaver & L. S. Wrightsman (Eds.), Measures of
personality and social psychological attitudes. San Diego, CA: Academic Press.
Rodgers, W., Andrews, F., & Herzog, R. (1992). Quality of survey measures: A structural
modeling approach. Journal of Official Statistics, 8(3), 251-275.
Rogers, E. M. (1983). Diffusion of innovations. New York, NY: Fee Press.
Rogers, E. M., & Bhowmik, D. K. (1970). Homophily-heterophily: Rational concepts for
communication research. Public Opinion Quarterly, 34 (4), 523-538.
Rohm, A. J., & Swaminathan, V. (2004). A typology of online shoppers based on shopping
motivations. Journal of Business Research, 57, 748-757.
Rosenberg, M. (1956). Misanthropy and political ideology. American Sociological Review, 21,
690-695.
Rosenberg, M. (1965). Society and the adolescent self-image. Princeton, NJ: Princeton
University Press.
Rosenberg, M. J., & Hovland, C. I. (1960). Cognitive, affective and behavioural components
of attitude. In C. I. Hovland & M. J. Rosenberg (Eds.), Attitude Organization and
Change (pp. 1-14). New Haven, CT: Yale University Press.
Rosseau, D. M., Sitkin, S. B., Burt, R. S., & Camerer, C. (1998). Introduction to special topic
forum: Not so different after all: A cross-discipline view of trust. The Academy of
Management Review, 23(3), 393-404.
Rossiter, J. R. (2002). The C-OAR-SE procedure for scale development in marketing.
International Journal of Research in Marketing, 19, 305-335.
Rothschild, M. L. (1984). Perspectives in involvement: Current problems and future directions.
In T. Kinnear (Ed.), Advances in Consumer Research (Vol. 11, pp. 216-217). Ann
Arbor, MI: Association for Consumer Research.
Rotter, J. B. (1966). Generalized expectancies for internal vs. external control. Psychological
Monography, 80, 1-28.
363
Rotter, J. B. (1967). A new scale of the measurement of interpersonal trust. Journal of
Personality, 35(4), 651-656.
Rotter, J. B. (1971). Generalized expectancies for interpersonal trust. American Psychologist,
26, 443-452.
Rotter, J. B. (1980). Interpersonal trust, trustworthiness, and gullibility. American Psychologist,
35(1), 1-7.
Ruble, D. N., & Frey, K. S. (1991). Changing patterns of comparative behavior as skills are
acquired: A functional model of self-evaluation. Hillsdale, NJ: Erlbaum.
Ruef, M., Aldrich, H. E., & Carter, N. M. (2003). The structure of founding teams: Homophily,
strong ties and isolation among U.S. entrepreneurs. American Sociological Review, 68,
195-222.
Rui, H., Liu, Y., & Whinston, A. (2013). Whose and what chatter matters? The effect of tweets
on movie sales. Decision Support Systems, 55(4), 863-870.
Rust, R. T., Lemon, K. N., & Zeithaml, V. A. (2004). Return on Marketing: Using Customer
Equity to Focus Marketing Strategy. Journal of Marketing, 68, 109-127.
Sabel, C. F. (1993). Studies trust: Building new forms of cooperation in a volatile economy.
Human Relations, 46(9), 1133-1170.
Sandage, C. H., & Leckenby, J. D. (1980). Students' attitudes towards advertising: Institution
vs instrument. Journal of Advertising, 9(2), 29-32.
Saris, W. E., & Gallhofer, I. (2007). Estimation of the effects of measurement characteristics
on the quality of survey questions. Survey Research Methods, 1(1), 29-43.
Sato, K. (1988). Trust and group size in social dilemma. Japanese Psychological Research,
30(2), 88-93.
Scarpello, L. (2012). Top 250 e-commerce retailers using social media.
Schacter, S. (1959). The psychology of affiliation: Experimental studies of the source of
gregariuosness. Stanford, CA: Standord University Press.
Scheer, L. K., & Stern, L. W. (1992). The effect of influence type and performance outcomes
on attitude toward the influencer Journal of Marketing Research, 29(February), 128-
142.
Schiffman, L. G., Sherman, E., & Kirpalani, N. (2002). Trusting souls: A segmentation of
voting public. Psychology & Marketing, 19(12), 993-1007.
Schlenker, B. R., Helm, B., & Tedschi, J. R. (1973). The effects of personaity and situational
variables on behavioral trust. Journal of Personality and Social Psychology, 25(3), 419-
427.
Schlosser, A. E. (2011). Can including pros and cons increase the helpfulness and
persuasiveness of online reviews? The interactive effects of ratings and arguments.
Journal of Consumer Psychology, 21, 226-239.
Schlosser, A. E., Shavitt, S., & Kanfer, A. (1999). Survey of internet users' attitudes toward
internet advertising. Journal of Interactive Marketing, 13(3), 34-54.
Schlosser, A. E., White, T. B., & Lloyd, S. (2006). Converting web site visitors into buyers:
How web site investment increases consumer trusting beliefs and online purchase
intentions. Journal of Marketing, 70(April), 133-148.
Schmalhausen, S. D. (1928). Why we misbehave. Garden City, NY: Garden City Pub. Col.
Schmitt, N., & Stults, D. M. (1986). Methodology review: Analysis of multitrait-multimethod
matrices. Applied Psychological Measurement, 10(1), 1-22.
Schneider, S. L., & Lopes, L. L. (1986). Reflections in preferences under risk: Who and when
may suggest why. Journal of Experimental Psychology: Human Perception and
Performance, 12(4), 535-548.
Schoefer, K., & Diamantopoulos, A. (2009). A typology of consumers' emotional response
styles during service recovery encounters. British Journal of Management, 20(3), 292-
308.
364
Schoorman, D., Mayer, R. C., & Davis, J. (2007). An integrative model of organizational trust:
Past, present, and future. Academy of Management Review, 32(2), 344-354.
Schoorman, F. D., Mayer, R. C., & Davis, J. H. (1996). Empowerment in veterinary clinics:
The role of trust in delegation. Paper presented at the 11th Annual Meeting of the
Society for Industrial and Organizational Psychology.
Schriesheim, C. A., Powers, K. J., Scandura, T. A., Gardiner, C. C., & Lankau, M. J. (1993a).
Improving construct measurement in management research: Comments and a
quanitative approach for assessing the theoretical content adeqacy of paper-and-pencil
survey-type instruments. Journal of Management, 19, 384-417.
Schriesheim, C. A., Powers, K. J., Scandura, T. A., Gardiner, C. C., & Lankau, M. J. (1993b).
Improving construct measurement in management research: Comments and a
quantitative approach for assessing the theoretical adequacy of paper-and-pencil survey-
type instruments. Journal of Management, 19(2), 385-417.
Schröder, J. E. (1996). An analysis of the consumer susceptibility to interpersonal influence
scale. Journal of Social Behavior and Personality, 11, 585-600.
Schumacker, R. E., & Lomax, R. G. (2004). A beginner's guide to structural equation modeling.
Mahaw, NJ: Lawrence Erlbaum Associates.
Schurr, P. H., & Ozanne, J. L. (1985). Influences on exchange processes: Buyers'
perconceptions of a seller's trustworthiness and bargaining toughness. Journal of
Consumer Research, 11(4), 939-953.
Schwab, D. P. (1980). Construct validity in organizational behavior. In B. M. Staw & L. L.
Cummings (Eds.), Research in organizational behvior (pp. 3-43).
Searle, J. R. (1995). The construction of social reality. New York, NY: The Free Press.
Seeman, M. (1959). On the meaning of alienation. American Sociological Review, 24, 783-791.
Selnes, F. (1998). Antecedents and consequences of trust and satisfaction in buyer-seller
relationships. European Journal of Marketing, 32(3-4), 305-322.
Sen, S. (2007). Investigating gender differences in attitude towards negative eWOM consumer
reviews: An attribution mediation model. Psychology & Marketing.
Sen, S. (2008). Determinants of consumer trust in virtual word-of-mouth: An observation study
froma retail website. Journal of American Academy of Business, 14(1), 30.
Sen, S., & Lerman, D. (2007). Why are you telling me this? An examination into negative
consumer reviews on the web. Journal of Interactive Marketing, 21(4), 76.
Senecal, S., & Nantel, J. (2004). The influence of online product recommendations on
consumers' online choices. Journal of Retailing, 80(2), 159-169.
Sensis. (2013). Aussies are getting socially mobile while businesses miss the mark! [Press
release]. Retrieved from [Link]
Releases/?ItemID=1225&count=1
Serva, M. A., Benamati, J., & Fuller, M. A. (2005). Trustworthiness in B2C e-commerce: An
examination of alternative models. Database for Advances in Information Systems,
36(3), 89-107.
Shamdasani, P. N., Stanaland, A. J. S., & Tan, J. (2001). Location, location, location: Insights
for advertising placement on the web. Journal of Advertising Research, 41(4), 7-21.
Shankar, V., Urban, G. L., & Sultan, F. (2002). Online trust: A stakeholder perspectiv, concepts,
implications, and future directions. Journal of Strategic Information Systems, 11, 325-
344.
Shapiro, D. L., Sheppard, B. H., & Cheraskin, L. (1992). Business on a handshake. Negotiation
Journal(October), 365-377.
Shapiro, S. P. (1987). The social control of impersonal trust. American Journal of Sociology,
93(3), 623-658.

365
Sharif, K. J., Kalafatis, S. P., & Samouel, P. (2005). Cognitive and behavioural determinants of
trust in small and medium-sized enterprises. Journal of Small Business and Enterprise
Development, 12(3), 409-421.
Sharma, A., & Stafford, T. F. (2002). The effect of retail atmospherics on customers'
perceptions of salespeople and customer persuasion: An empirical investigation.
Journal of Business Research, 49, 183-191.
Shavitt, S., Lowrey, P., & Haefner, J. (1998). Public attitude toward advertising: More favorable
than you might think. Journal of Advertising Research, 38(July/August), 7-22.
Sheppard, B. H., & Sherman, D. M. (1998). The grammars of trust: a model and general
implications. The Academy of Management Review, 23(3), 422-437.
Sher, P., & Lee, S. (2009). Consumer skepticism and online reviews: An elaboration likelihood
model perspective. Social Behavior and Personality, 37(1), 137.
Shimp, T. A., & Sharma, S. (1987). Consumer ethnocentrism: Construction and validation of
the CETSCALE. Journal of Marketing Research, 24(August), 280-289.
Shin, D.-H. (2010). The effects of trust, security and privacy in social networking: A security-
based approach to understand the pattern of adoption. Interacting with Computers, 22,
428-438.
[Link]. (2009). Message boards: State of retailing online.
Shorstrom, E. L. (1975). Personal orientation dimension. San Diego, CA: EdITs.
Shuptrine, F. K., Pruden, H. O., & Longman, D. S. (1977). Alienation from the marketplace.
Journal of the Academy of Marketing Science, 5(3), 233-248.
Sichtmann, C. (2007). An analysis of antecedents and consequences of trust in a corporate
brand. European Journal of Marketing, 41(9-10), 999-1015.
Sillence, E., Briggs, P., Peter, H., & Fishwick, L. (2006). A framework for understanding trust
factors in web-based health advice. Information Journal of Human-Computer Studies,
64(8), 697-713.
Simmel, G. (1964). The sociology of Georg Simmel. New York, NY: The Free Press.
Simons, T. (2002). Behavioral integrity: The perceived alignment between managers' words
and deeds as a research focus. Organizational Science, 13(1), 18-38.
Singh, J., & Sirdeshmukh, D. (2000). Agency and trust mechanisms in consumer satisfaction
and loyalty judgments. Journal of the Academy of Marketing Science, 28(1), 150-167.
Sirdeshmukh, D., Singh, J., & Sabol, B. (2002). Consumer trust, value, and loyalty in relational
exchanges. Journal of Marketing, 66(January), 15-37.
Sitkin, S. B., & Pablo, A. L. (1992). Reconceptualising the determinants of risk behavior.
Academy of Management Review, 17(1), 9-38.
Sitkin, S. B., & Roth, N. L. (1993). Explaining the limited effectiveness of legalistic "remedies"
for trust/distrust. Organizational Science, 4, 367-392.
Sitkin, S. B., & Weingart, L. R. (1995). Determinants of risky decision-making behavior: A test
of the mediating role of risk perceptions and propensity. The Academy of Management
Journal, 38(6), 1573-1592.
Smith, D., Menon, S., & Sivakumar, K. (2005). Online peer and editiorial recommendations,
trust, and choice in virtual markets. Journal of Interactive Marketing, 19(3), 15-37.
Smith, D. N. (2002). Trust me, would I steer you wrong? The influence of peer
recommendatiions within virtual communities. (PhD), University of Illinois.
Smith, H. J., Milberg, S. J., & Burke, S. J. (1996). Information privacy: Measuring individuals'
concerns about organizational practices. MIS Quarterly, 20(2), 167-196.
Smith, J. B. (1997). Selling alliances: Issues and insights. Industrial Marketing Management,
26(2), 146-161.
Smith, M. B. (1947). The personal setting of public opinions: A study of attitudes towards
Russia. Public Opinion Quarterly, 11, 507-523.

366
Smith, N. C., & Martin, E. C. (1997). Ethics and target marketing: The role of product harm
and consumer vulnerability. Journal of Marketing, 61(3), 1-20.
Smith, T., Coyle, J. R., Lightfoot, E., & Scott, A. (2007). Reconsidering models of influence:
The relationship between consumer social networks and word-of-mouth effectiveness.
Journal of Advertising Research, 47(4), 387-397.
Soh, H. (2007). Measuring trust in advertising: Development and validation of the ADTRUST
scale. (Unpublished Doctorial Dissertation), Athens, GA.
Soh, H., Reid, L. N., & Whitehill, K. (2009). Measuring trust in advertising: Developing and
validation of the ADTRUST scale. Journal of Advertising, 38(2), 83-102.
Solomon, D. (1963). Influneces on the decisions of adolescents. Human Relations, 16, 45-60.
Solomon, L. (1960). The influence of some types of power relationships and game strategies
upon the development of interpersonal trust. Journal of Abnormal Social Psychology,
61(2), 223-230.
Solomon, M., Barnossy, G., Askegaard, S., & Hogg, M. (2006). Consumer behavior: A
European perspective (Vol. 3rd Edition). Harlow, UK: Prentice Hall Europe.
Solomon, R. C., & Flores, F. (2001). Building trust in business, politics, relationships, and life.
New York, NY: Oxford University Press.
Speck, P. S., & Elliott, M. T. (1997). Predictors of advertising avoidance in print and broadcast
media. Journal of Advertising, 26(Fall), 61-76.
Spector, P. E. (1992). Summated rating scale construction: An introductiion. Newbury Park,
CA: Sage Publications, Inc.
Srinivasan, N., & Ratchford, B. T. (1991). An empirical test of a model of external search for
automobiles. Journal of Consumer Research, 18(September), 233-242.
Stack, L. C. (1978). Trust. In H. London & J. E. Exner (Eds.), Dimensionality of personality
(pp. 561-599). New York, NY: Wiley.
Stanley, D. J., Meyer, J. P., & Topolnytsky, L. (2005). Employee cynicism and resistance to
organizational change. Journal of Business and Psychology, 19(4), 429-459.
Statistik Austria. (2012). Online shopping.
Statistik Austria. (2013). 3,5 Mio. Personen gehen online shoppen [Press release]. Retrieved
from [Link]
Statistik Austria. (2014). Internetnutzerinnen und Internetnutzer 2013. Retrieved 2014-05-29,
from [Link]
einsatz_in_haushalten/[Link]
Statistisches Bundesamt. (2014). IT Nutzung. Retrieved 2014-05-29, from
[Link]
nsbedingungen/ITNutzung/[Link]
Steenkamp, J.-B., & Baumgartner, H. (1998). Assessing measurement invariance in cross-
national consumer research. Journal of Consumer Research, 25(June), 78-90.
Steenkamp, J.-B., De Jong, M. J., & Baumgartner, H. (2010). Social desirability response
tendencies in survey research. Journal of Marketing Research, 47(2), 199-214.
Steenkamp, J.-B., & Trijp, H. C. M. (1997). Attribute elicitation techniques in marketing
research: A comparison of three methods. Marketing Letters, 8(2), 153-165.
Steiger, J. H., & Lind, J. C. (1980). Statistically-based tests for the number of common factors.
Paper presented at the Meeting of the Psychometric Society, Iowa City, IA.
Steil, J. M., & Hay, J. L. (1997). Social comparison in the workplace: A study of 60 dual-career
couples. Personality and Social Psychology Bulletin, 23, 427-438.
Sternthal, B., Phillips, L. W., & Dholakia, R. (1978). The persuasive effect of scarce credibility:
A situational analysis. Public Opinion Quarterly, 42(3), 285-314.
Stewart, K. J. (2003). Trust transfer on the world wide web. Organization Science, 14(1), 5-17.
Stickland, L. H. (1958). Surveillance and trust Journal of Personality, 26, 200-215.

367
Stikin, S. B. (1995). On the positive effect of legalization on trust. In R. J. Bies, R. J. Lewicki
& B. J. Sheppard (Eds.), Research on negotiation in organizations (pp. 195-217).
Greenwich, CT: JAI Press.
Stone, R. N., & Gronhaug, K. (1993). Perceived risk: Further considerations for the marketing
discipline. European Journal of Marketing, 27(3), 39-50.
Straub, D. W. (1989). Validating instruments in MIS research. MIS Quarterly, 13(2), 147-165.
Straub, D. W., Boudreau, M.-C., & Gefen, D. (2004). Validation guidelines for IS positivist
research. Communications of the AIS, 13, 380-427.
Streitfeld, D. (2011). In a race to out-rave, 5-star web reviews go for $5. New York Times.
[Link]
[Link]?_r=0
Strickland, L. H. (1958). Surveillance and trust. Journal of Personality, 26, 200-215.
Sundaram, D. S., Mitra, K., & Webster, C. (1998). Word-of-mouth communications: A
motivational analysis. Advances in Consumer Research, 25, 527-531.
Swan, J. E., Bowers, M. R., & Richardson, L. D. (1999). Customer trust in the salesperson: An
integrative review and meta-analysis of the empirical literature. Journal of Business
Research, 44, 93-107.
Swan, J. E., Trawick, F. I., & Silva, D. (1985). How industrial salespeople gain customer trust.
Industrial Marketing Management, 14, 203-211.
Swan, J. E., Trawick, I. F. J., Rink, D. R., & Roberts, J. J. (1988). Measuring dimensions of
purchaser trust of industrial salespeople. Journal of Personal Selling and Sales
Management, 8(May), 1-9.
Sweeney, J. C., & Soutar, G. N. (2001). Consumer perceived value: The development of a
multiple item scale Journal of Retailing, 77(2), 203-220.
Sweeney, J. C., Soutar, G. N., & Mazzarol, T. (2008). Influencing word of mouth effectiveness:
Receiver perspectives. European Journal of Marketing, 42(2/3), 344-364.
Szulanski, G., Cappetta, R., & Jensen, R. J. (2004). When and how trustworthiness matters:
Knowledge transfer and the moderating effect of causal ambiguity. Organization
Science, 15(5), 600-613.
Tan, F. B., & Sutherland, P. (2004). Online consumer trust: A multi-dimensional model.
Journal of Electronic Commerce in Organizations, 2(3), 40-58.
Tan, S. J., & Chia, L. (2007). Are we measuring the same attitude? Understanding media effects
on attitude towards advertising. Marketing Theory, 7(4), 353-377.
Tan, Y., & Theon, W. (2001). Toward a generic model of trust for electronic commerce.
International Journal of Electronic Commerce, 5(2), 61-74.
Taylor, C. R., & Lee, D.-H. (2008). Introduction: New developments in e-commerce research.
Psychology & Marketing, 25, 565-567.
Taylor, R. G. (1989). The role of trust in labor-management relations. Organization
Development Journal, Summer, 85-89.
Taylor, R. N., & Dunnette, M. D. (1974). Relative contributioin of decision-maker attributes to
decision processes. Organisational Behavior and Human Performance, 12, 286-298.
Taylor, S. E., Buunk, B. P., & Aspinwall, L. G. (1990). Social comparison, stress, and coping.
Personality and Social Psychology Bulletin, 16, 74-89.
Tellis, G. J., Chandy, R. K., MacInnis, D., & Thaivanich, P. (2005). Modeling the microeffects
of television advertising: Which ad works, when, where, for how long, and why?
Marketing Science, 24(3), 359-366.
The American Heritage Dictionary of the Englisch Language. (Ed.) (1992). New York, NY:
Houghton Mifflin Comp.
The Concise Oxford Dictionary of Current English. (Ed.) (1995). Oxford: Clarendon Press.
Thelwall, M. (2009). Homophily in MySpace. Journal of the American Society for Information
Science and Technology, 60(2), 219-231.
368
Thibaut, J. W., & Kelley, H. H. (1959). The social psychology of groups. New York, NY:
Wiley.
Thomson, N. (2003). More companies pay heed to their word of mouse reputation, The New
York Times. Retrieved from [Link]
more-companies-pay-heed-to-their-word-of-mouse-
[Link]?pagewanted=all&src=pm
Thorslund, C. (1976). Interpersonal trust: A review and examination of the concept. Goteborg
Psychological Report, 6(6), 1-21.
Thorson, K., Vraga, E., & Ekdale, B. (2010). Credibility in context: How uncivil online
commentary affects news credibility. Mass Communication and Society, 13(3), 289-
313.
Thurstone, L. L. (1931). The measurement of social attitudes. Journal of Abnormal and Social
Psychology, 26, 249-269.
Tian, K. T., Bearden, W. O., & Hunter, G. L. (2001). Consumers' need for uniqueness: Scale
development and validation. Journal of Consumer Research, 28(June), 50-66.
Tormala, Z. L., & Petty, R. E. (2007). Multiple roles for source credibility under high
elaboration: It's all in the timing. Social Cognition, 25(4), 536-552.
Trifts, V., & Häubl, G. (2003). Information availability and consumer preference: Can online
retailers benefit from providing access to competitor price information? Journal of
Consumer Psychology, 13(1-2), 149-159.
Trochim, A. (2012). Construct validity. Retrieved 2012-11-01, 2012, from
[Link]
Trusov, M., Bucklin, R. E., & Paulwels, K. (2009). Effects of word-of-mouth versus traditional
marketing: Findings from an internet social networking site. Journal of Marketing,
73(5), 90-102.
Tsai, W., & Ghoshal, S. (1998). Social capital and value creation: The role of intrafirm
networks. Academy of Management Journal, 41(4), 464-476.
Tsao, W.-C., & Chang, H.-R. (2010). Exploring the impact of personality traits on online
shopping behavior. African Journal of Business Management, 4(9), 1800-1812.
Tseng, S., & Fogg, B. J. (1999). Credibility and computing technology. Communications of the
ACM, 42(5), 39-44.
Tucker, L. R., & Lewis, C. (1973). A reliability coefficient for maximum likelihood factor
analysis. Psychometrika, 38, 1-10.
Tung, L. L., Tan, P. L. J., Chia, P. J. T., Koh, Y. L., & Yeo, H. L. (2001). An empirical
investigation of virtual communities and trust. Paper presented at the Twenty-Second
International Conference on Information Systems.
Tuttle, B. (2012). 9 reasons why you shouldn't trust online reviews, Time Magazine. Retrieved
from [Link]
online-reviews/
Tyler, T. R., & Kramer, R. M. (1996). Wither trust? In R. M. Kramer & T. R. Tyler (Eds.),
Trust in organizations: Frontiers of theory and research (pp. 1-15). Thousand Oaks,
CA: Sage Publications, Inc.
Tyler, T. R., & Schuller, R. A. (1991). Aging and attitude change. Journal of Personality and
Social Psychology, 61, 689-697.
Ullman, J. B. (2001). Structural equation modeling. In B. G. Tabachnick & S. Fidell (Eds.),
Using multivariate statistics. Needham Heights, MA: Allyn & Bacon.
Urban, G. L., Amyx, C., & Lorenzon, A. (2009). Online trust: State of the art, new frontiers,
and research potential. Journal of Interactive Marketing, 23(2), 179-190.
Urban, G. L., Sultan, F., & Qualls, W. J. (2000). Placing trust at the center of your Internet
strategy. Sloan Management Review, 42(1), 39-48.
US Census Bureau. (2011). Computer and Internet use [Link]
369
Utz, S., Kerkhof, P., & Van den Bos, J. (2012). Consumers rule: How consumer reviews
influence perceived trustworthiness of online stores. Electronic Commerce Research
and Applications, 11, 49-58.
Valck, K. d., Bruggen, G. H. v., & Wierenga, B. (2009). Virtual communities: A marketing
perspective. Decision Support Systems, 47, 185-203.
Van den Berge, B., Soley, L., & Reid, L. (1981). Factor study of dimensions of advertiser
credibility. Journalism Quarterly, 58(4), 629-631.
Van der Heijden, H., Verhagen, T., & Creemers, M. (2003). Understanding online purchase
intentions: Contributions from technology and trust perspectives. European Journal of
Information Systems, 12(1), 41-48.
VanBoskirk, S. (2009). US interactive marketing forecast by industry, 2009 to 2014: Forrester.
Vanhala, M., Puumalainen, K., & Blomqvist, K. (2011). Impersonal trust: The development of
the construct and the scale. Personnel Review, 40(4), 485-513.
Velicer, W. F., & Jackson, D. N. (1990). Component analysis versus common factor analysis:
Some issues in selecting an appropriate procedure. Multivariate Behavioral Rresearch,
25, 1-28.
Velicer, W. F., Peacock, A. C., & Jackson, D. N. (1982). A comparison of component and factor
patterns: A Monte Carlo approach. Multivariate Behavioral Rresearch, 17, 371-388.
Venkatesh, V., Moris, M. G., Davis, G. B., & Davis, F. D. (2003). User acceptance of
information technology: Toward a unified view. MIS Quarterly, 27(3), 425-478.
Verhoef, P. C., & Donkers, B. (2005). The effect of acquisition channels on customer loyalty
and cross-buying. Journal of Interactive Marketing, 19(2), 31-43.
Vijayasarathy, L. R. (2004). Predicting consumer intentions to use on-line shopping: The case
for an augmented technology acceptance model. Information & Management, 41, 747-
762.
Visser, P. S., & Krosnick, J. A. (1998). Development of attitude strength over the life cycle:
Surge and decline. Journal of Personality and Social Psychology, 75, 1389-1410.
Vlek, C., & Stallen, P. J. (1980). Rational and personal aspects of risk. Acta Psychologica, 45,
273-300.
Volken, T. (2002). Elements of trust: The cultural dimension of Internet diffusion revisited.
Electronic Journal of Sociology. [Link]
Walczuch, R., & Lundgren, H. (2004). Psychological antecedents of institution-based consumer
trust in e-retailing. Information & Management, 42, 159-177.
Walker, J., Wassermann, S., & Wellman, B. (1994). Statistical models for social support
networks. In S. Wassermann & J. Galaskiewicz (Eds.), Advances in social network
analysis. Thousand Oaks, CA: Sage.
Walsh, G., & Beatty, S. E. (2007). Customer-based corporate reputation of a service firm: Scale
development and validation. Journal of the Academy of Marketing Science, 35(1), 127-
143.
Walsh, G., Hassan, L. M., Shiu, E., Andrews, J. G., & Hastings, G. (2010). Segmentation in
social marketing: Insights from the European Union's multi-country, antismoking
campaign. European Journal of Marketing, 44(7-8), 1140-1164.
Walsh, G., & Mitchell, V. W. (2010). The effect of consumer confusion proneness on word of
mouth, turst, and customer satisfaction. European Journal of Marketing, 44(6), 838-
859.
Wang, L. (2010). The individual's disposition to trust as a moderatorof the relationship between
electronic word-of-mouth and consumer brand attitude. Paper presented at the
Proceedings of the 5th International Conference on Cooperation and Promotiion of
Information Resources in Science and Technology.
Wang, W., & Benbasat, I. (2005). Trust in and adoption of online recommendation agents.
Journal of Association for Information Systems, 6(3), 72-101.
370
Wang, X. T., Kruger, D. J., & Wilke, A. (2009). Life history variables and risk-taking
propensity. Evolution and Human Behavior, 30, 77-84.
Wang, Y., Sun, S., & Lei, W. (2009). Examining beliefs and attitudes toward online advertising
among Chinese consumers. Direct Marketing: An International Journal, 3(1), 52-66.
Wang, Y. D., & Emurian, H. H. (2005). An overview of online trust: Concepts, elements, and
implications. Computers in Human Behavior, 21, 105-125.
Ward, J. C., & Ostrom, A. L. (2003). The Internet as information minefield: An analysis of the
source and content of brand information yieled by net searches. Journal of Business
Research, 56, 907-914.
Wasko, M. M., & Faraj, S. (2000). It is what one does: Why people participate and help others
in electronic communities of practice. Journal of Strategic Information Systems, 9, 155-
173.
Wasko, M. M., & Faraj, S. (2005). Why should I share? Examining social capital and
knowledge contribution in electronic networks of practice. MIS Quarterly, 29, 35-57.
Wassermann, T. (2006). Consumers don't trust blogs. Brandweek, 47(31), 10.
Wathen, C. N., & Burkell, J. (2002). Believe it or not: Factors influencing credibility on the
web. Journal of American Society for Information Science and Technology, 53(2), 134-
144.
Watson, M. L. (2005). Can there be just one trust? A cross-disciplinary identification of trust
definitions and measurement. Miami, FL.
Waymnet, H. A., & Taylor, S. E. (1995). Self-evaluation processes: Motives, information use,
and self-esteem. Journal of Personality, 63, 729-757.
Web, D. J., Mohr, L. A., & Harris, K. E. (2008). A re-examination of socially responsible
consumption and its measurement. Journal of Business Research, 61(2), 91-98.
Webb, W. M., & Worchel, P. (1986). Trust and distrust. In P. Worchel & W. G. Austin (Eds.),
Psychology of intergroup relations (pp. 213-228). Chicago, IL: Nelson-Hall.
Weber, E. U., & Milliman, R. A. (1997). Perceives risk attitudes: Relating risk perception to
risky choice. Management Science, 43(2), 123-144.
Webster, F. E. (1968). Interpersonal communication and salesman effectiveness. Journal of
Marketing, 32, 7-13.
Weiner, B. (2000). Attributional thoughts about consumer behavior. Journal of Consumer
Research, 27, 382-387.
Wicker, A. W. (1969). Attitudes versus actions: The rleationship of verbal and overt behavioral
responses to attitude objects. Journal of Social Issues, 25, 41-78.
Wicks, A. C., Berman, S. L., & Jones, T. M. (1999). The structure of optimal trust: Moral and
strategic implications. Academy of Management Review, 24(1), 99-116.
Widing, R. E., & Talarzyk, W. W. (1993). Electronic information systems for consumers: An
evaluation of computer-assisted formats in multiple decision environments. Journal of
Marketing Research, 30, 125-141.
Wiggins, J. S. (1968). Personality structure. In P. R. Farnsworth (Ed.), Annual Review of
Psychology (Vol. 19, pp. 320-322). Palo Alto, CA: Annual Reviews.
Wiggins, J. S. (1994). The five-factor model of personality: Theoretical perspectives. New
York, NY: Guilford Press.
Wilcox, J. B., Howell, R. D., & Breivik, E. (2008). Questions about formative measurement.
Journal of Business Research, 61, 1219-1228.
Williams, L. J., Vandenberg, R. J., & Edwards, J. R. (2009). 12 structural equation modeling in
management research: A guide for improved analysis. The Academy of Management
Annals, 3(1), 543-604.
Williams, R. B., Barefoot, J. C., & Shekelle, R. B. (1985). The health consequences of hostility.
In M. A. Chesney & R. H. Rosenham (Eds.), Anger and hostility in cardiovascular and
behavioral disorders (pp. 173-185). Washington, DC: Hemisphere.
371
Williamson, O. E. (1974). Markets and hierarchies. New York, NY: Free Press.
Williamson, O. E. (1985). The economic institutions of capitalism. New York, NY: The Free
Press.
Williamson, O. E. (1993). Calculativeness, trust and economic organization. Journal of Law
and Economics, 36(April), 487-500.
Witt, L. A., Burke, L. A., Barrick, M. A., & Mount, M. K. (2002). The interactive effects of
conscientiousness and agreeableness on job performance. Journal of Applied
Psychology, 87(1), 164-169.
Witt, R. E., & Bruce, G. D. (1972). Group influence and brand choice congruence. Journal of
Marketing Research, 9(11), 440-443.
Wolin, L. D., Korgaonkar, P., & Lund, D. (2002). Beliefs, attitudes, and behavior towards web
advertising. International Journal of Advertising, 21(1), 87-113.
Wong, C.-S., Law, K. S., & Huang, G.-H. (2008). On the importance of conducting construct-
level analysis for multidimensional constructs in theory development and testing.
Journal of Management, 34(4), 744-764.
Wood, R., & Bandura, A. (1989). Impact of conceptions of ability on self-regulatory
mechanisms and complex decision making. Journal of Personality and Social
Psychology, 46, 407-415.
Wood, W., & Stanger, B. (1994). Why are some people easier to influence than others. In S.
Shavitt & T. C. Brock (Eds.), Persuasion: Psychological insights and perspectives (pp.
149-174): Allyn & Bacon.
Woodside, A. G., & Davenport, W. (1974). The effect of salesman similaity and expetise on
consumer purchase behavior. Journal of Marketing Research, 11(May), 198-202.
Worchel, P. (1979). Trust and distrust. In W. G. Austin & S. Worchel (Eds.), The social
psychology of intergroup relations. Belmont, CA: Wadsworth.
Worthington, R. L., & Whittaker, T. A. (2006). Scale development research: A content analysis
and recommendations for best practices. The Counseling Psychologist, 34, 806-838.
Wrightsman, L. S. (1974). Assumptions about human nature: A social-psychological approach.
Monterey, CA: Brooks/Cole.
Wu, J.-J., Chen, Y.-H., & Chung, Y.-S. (2010). Trust factors influencing virtual community
members: A study of transaction communities. Journal of Business Research, 63, 1025-
1032.
Wu, P. C. S., & Wang, Y.-C. (2011). The influences of electronic word-of-mouth message
appeal and message source credbility on brand attitude. Asia Pacific Journal of
Marketing and Logistics, 23(4), 448-472.
Xie, H., Miao, L., Kuo, P.-J., & Lee, B.-Y. (2011). Consumers' response to ambivalent online
hotel reviews: The role of perceived source credibility and pre-decisional disposition.
International Journal of Hospitality Management, 30, 178-183.
Yamagishi, T., & Yamagishi, M. (1994). Trust and commitment in the United States and Japan.
Motivation and Emotion, 18(2), 129-166.
Yang, K. C. C. (2003). Internet users' attitude toward and beliefs about Internet advertising: An
exploratory research from Taiwan. Journal of International Consumer Marketing,
15(4), 43-65.
Yang, K. C. C. (2007). Factors influencing Internet users' peceived credibility of news-related
blogs in Taiwan. Telematics and Informatics, 24, 69-85.
Yang, Y., & Chou, M. J. (2000). Effects of word-of-mouth communication on consumer's
attitude change. The Korean Journal of Advertising, 11, 7-34.
Yavas, U. (1994). Research note: Students as subjects in advertising and marketing research.
International Marketing Review, 11(4), 35-43.
Young, L., & Albaum, G. (2003). Measurement of trust in salesperson-customer relationship
in direct selling. Journal of Personal Selling and Sales Management, 23(3), 253-269.
372
Yu, C., & Tang, X. (2010). The construct of influence of word of mouth: Receiver perspectives.
Paper presented at the Proceedings of 2010 Summit International Marketing Science
and Management Technology Conference
Zaffane, R., Tipu, S. A., & Ryan, J. C. (2011). Communication, commitment and trust:
Exploring the triad. International Journal of Business and Management, 6(6), 77-87.
Zaheer, A., McEvily, B., & Perrone, V. (1998). Does trust matter? Exploring the effects of
inter-organizational and interpersonal trust on performance. Organization Science, 9(2),
141-159.
Zaichkowsky, J. L. (1985). Measuring the involvement construct. Journal of Consumer
Research, 12(12), 341-352.
Zaichkowsky, J. L. (1986). Conceptualizing involvement. Journal of Advertising, 15(2), 4-14.
Zaichkowsky, J. L. (1994). The personal involvement inventory: Reduction, revision, and
application to advertising. Journal of Advertising, 23(4), 59-70.
Zajonc, R. B. (1984). On the primacy of affect. American Psychologist, 39, 117-121.
Zaltman, G., & Moorman, C. (1988). The importance of personal trust in the use of research.
Journal of Advertising Research, 28(October/November), 16-24.
Zand, D. E. (1972). Trust and managerial problem solving. Administrative Science Quarterly,
17, 229-239.
Zanna, M. P., & Rempel, J. K. (1988). Attitudes: A new look at an old concept. In D. Bar-Tal
& A. Kruglanski (Eds.), The social psychology of knowledge New York, NY:
Cambridge University Press.
Zeithaml, V., A., & Bitner, M. J. (2000). Services Marketing: Integrating Customer Focus
across the Firm. Boston, MA.
Zeithaml, V. A., Rust, R. T., & Lemon, K. N. (2001). The customer pyramid: Creating and
serving profitable customers. California Management Review, 43(4), 118-142.
ZenithOptimedia. (2012). ZenithOptimedia forecasts 4.1% growth in global adspend in 2013
[Press release]. Retrieved from [Link]
forecasts-4-1-growth-in-global-adspend-in-2013/
Zhou, L., Zhang, P., & Zimmermann, H.-D. (2013). Social commerce research: An integrated
view. Electronic Commerce and Applications, 12(2), 61-68.
Zhu, F., & Zhang, X. (2010). Impact of online consumer reviews on sales: The moderating role
of product and consumer characteristics. Journal of Marketing, 74(2), 133-148.
Ziegler, C. N., & Goldbeck, J. (2007). Investigating interactions of trust and interest similarity.
Decision Support Systems, 43(2), 460-475.
Zou, P., Yu, B., & Hao, Y. H. (2011). Does the valence of online consumer reviews matter for
consumer decision-making? The moderating role of consumer expertise. Journal of
Computers, 6(3), 484-488.
Zucker, L. G. (1986). Production of trust: Institutional sources of economic structure. In B. M.
Straw & L. L. Cummings (Eds.), Research organizational behavior (pp. 53-111).
Greenwich, CT: JAI Press.

373
Appendices

Available at [Link]:

A) Main study’s questionnaire

B) Generalizability study’s questionnaire

375
© Springer Fachmedien Wiesbaden 2017
W. Weitzl, Measuring Electronic Word-of-Mouth Effectiveness,
DOI 10.1007/978-3-658-15889-7