Electronic Commerce Research and Applications 14 (2015) 616–630

Contents lists available at ScienceDirect

Electronic Commerce Research and Applications

journal homepage: [Link]/locate/ecra

Understanding why consumers engage in electronic word-of-mouth

communication: Perspectives from theory of planned behavior and
justice theory
Jen-Ruei Fu a,⇑, Pei-Hung Ju b,1, Chiung-Wen Hsu a,2
a
Department of Information Management, National Kaohsiung University of Applied Sciences, 415 Chien Kung Road, Kaohsiung 807, Taiwan, ROC
b
Department of Business Administration, National Kaohsiung University of Applied Sciences, 415 Chien Kung Road, Kaohsiung 807, Taiwan, ROC

a r t i c l e i n f o a b s t r a c t

Article history: Word-of-mouth has been recognized as a dominant factor in shaping consumer behavior. However, what
Received 3 March 2015 drives consumers to post their positive and negative purchasing experiences in the online environment
Received in revised form 17 September 2015 remains an important but largely neglected issue. Building on the theory of planned behavior, the justice
Accepted 17 September 2015
theory and the social psychology literature, this study aims to investigate the antecedents of consumers’
Available online 9 October 2015
intention to engage in eWOM communication. Specifically, we separate negative and positive eWOM into
two distinct concepts since the motivations underlying consumers’ decisions to post positive and nega-
Keywords:
tive eWOM are likely to be different. Through an experience survey, respondents were required to reflect
Electronic word-of-mouth (eWOM)
communication
on recent positive or negative shopping experiences. The findings reveal that intentions to engage in pos-
Theory of planned behavior itive and negative eWOM communication are associated with different antecedents. Consumers who
Justice theory intend to post positive eWOM appear to be more driven by underlying attitudinal factors, whereas those
Satisfaction who consider posting negative eWOM are more driven by social pressure. In addition, consumers’ feel-
ings of satisfaction are largely driven by their perception of distributive justice for negative shopping
experiences. In contrast, satisfaction is significantly influenced by the perceptions of interactional and
procedural justice for positive shopping experiences. Our results provide insights and implications for
scholars and managers.
Ó 2015 Elsevier B.V. All rights reserved.

1. Introduction significantly improve consumers’ perceptions of the trustworthi-

ness of the seller (Lee and Lee 2006).
Word-of-mouth (WOM) has long been recognized as an impor- In light of these developments, the business press has identified
tant factor in shaping consumer purchase behavior (East et al. electronic word-of-mouth (thereafter, eWOM) as a growing oppor-
2007). It is described as ‘‘a dominant force in the marketplace” tunity and a threat that deserves managers’ consideration (Ariely
and the ‘‘ultimate test of the customer’s relationship” (Bendapudi 2007). Yet, despite these warnings, to our surprise relatively few
and Berry 1997). This practice now extends into the online world. studies help practitioners understand the factors that influence
The increasing access to the Internet has transformed people from consumers’ intention to engage in eWOM communication.
passive consumers to active participants. While only a limited Although existing studies have advanced our knowledge of eWOM,
number of consumers may be exposed to WOM in the offline most of them have focused their attention on how consumers’
setting, consumers at present have more convenient access to eWOM evaluations affect their purchase decisions (e.g., Utz et al.
websites where they are free to share positive or horrible 2012). Limited attention has been paid to the investigation of ante-
experiences with total strangers (Cheung and Lee 2012). These cedents of the two distinct types (positive and negative) of eWOM.
posted comments and evaluations will substantially damage or Past research has considered intention to engage in WOM com-
munication as an important dependent variable (e.g. de Matos
et al. 2009). Hennig-Thurau et al. (2004) report that due to the con-
⇑ Corresponding author. Tel.: +886 7 3814526x7512.
ceptual closeness of traditional offline and online WOM communi-
E-mail addresses: fred@[Link] (J.-R. Fu), [Link]@[Link]
(P.-H. Ju), sandrahsu33@[Link] (C.-W. Hsu).
cation, factors which in the literature have been identified as
1
Tel.: +886 7 3814526x7317. relevant in the traditional offline environment can also be antici-
2
Tel.: +886 7 3814526x7510. pated to be of relevance in the online setting. However, Brown

[Link]
1567-4223/Ó 2015 Elsevier B.V. All rights reserved.
 J.-R. Fu et al. / Electronic Commerce Research and Applications 14 (2015) 616–630 617

et al. (2007) disagreed and insisted that existing offline research suggests that most research on eWOM intention remains explora-
may not be appropriate to describe eWOM and its influence on tory in nature. Thus, we contribute conceptually to enriching the
consumer behavior. We concur with Brown et al. (2007) since current eWOM literature with a more comprehensive model of
there are several incidents and issues unique to the online environ- the simultaneous effects of several key antecedents. Finally, East
ment. First, in the offline setting, exit can be seen as an economic et al. (2007) suggest that the motivations behind a consumer’s
action, whereas voice is more of a political statement (Fornell and decision to engage in positive vs. negative WOM communication
Wernerfelt 1987). However, in the online environment, voice turns are likely to differ. Therefore, we specifically separate negative
into not only a political movement, but also a significant economic and positive eWOM communication into two different concepts.
action, which can influence the decisions of other consumers (Lee Thus, the theoretical understanding of the eWOM intention is
and Lee 2006). Past research unanimously suggests that service extended by demonstrating how positive and negative eWOM
providers should encourage their customers to voice their com- are triggered by different antecedents.
plaints rather than exit silently (e.g., East 2000) because sellers
can benefit from direct complaints by receiving valuable feedback.
However, when it comes to online environments, negative eWOM 2. Literature review of eWOM communication
is likely to dissuade potential buyers from considering a particular
product or brand, thus damaging the company’s reputation and The topic of eWOM communication is attracting increasing
financial position (Pfeffer et al. 2014). Compared with offline attention in many research disciplines, such as marketing, con-
word-of-mouth, negative eWOM is mass-public oriented and is sumer behavior, computer-mediated communication, and eco-
likely to entail an explicit intent to ‘‘get the firm in trouble”. nomics. Several different definitions have been proposed for
Although this threat has been considered, what drives consumers behaviors related to eWOM communication. Sun et al. (2006)
to post eWOM online has seldom been investigated. defined eWOM communication as ‘‘online opinion leadership”
Second, eWOM communication is an observable, social activity which stressed the process by which people (opinion leaders)
(Mattila and Patterson 2004). Thus, the social and cultural effects influence the attitudes or behaviors of friends. It implies that
may keep consumers silent even if they are willing to put extra eWOM communication is an act by which individuals exert power-
effort into reporting their satisfactory/poor performing counter ful influences upon others. Hsiao (2011) defined ‘‘public online
parties (Hsiao 2011). Spreading eWOM is thwarted by the fact that consumer complaining behaviors” as seeking redress from the
the content is bound to, or is likely to be traced back to the name of company or resorting to the third party via the Internet. Negative
the person distributing it (Riivits-Arkonsuo and Leppiman 2013). eWOM communication was viewed as an action to recover losses
Thus, consumers’ concerns about their own reputations are impor- caused by unsatisfactory products or services. These definitions,
tant (Dellarocas and Wood 2008). however, imply certain motivations underlying a consumer’s
Third, substantial evidence in the literature suggests that social eWOM communication behaviors. A more holistic conceptualiza-
behavior is often motivated by the ‘‘need to help others” or by a tion, which was also adopted by the current study, was provided
desire to improve social welfare (Chelminski and Coulter 2011). by Hennig-Thurau et al. (2004), who defined ‘‘eWOM communica-
Consumers in communal relationship norms (especially on the tion” as any positive or negative statement made by potential, actual,
Internet) are likely to do something for other people without or former customers about a product or company, which is made
expecting anything in return. Thus, consumers may engage in available to a multitude of people and institutions via the Internet.
eWOM communication in anticipation of other people becoming To provide an overview of the eWOM communication research
aware of a company’s service failures to prevent them from expe- field, a systematic literature review was conducted. Key words
riencing similar situations. Most past research in offline settings such as ‘‘online word-of-mouth”, ‘‘ewom”, ‘‘owom” were searched
has neglected consumers’ social characteristic and assumes they in a number of academic databases (ABI/INFORM, ScienceDirect,
are rational; thus, there are limitations to our understanding of Emerald, EBSCOhost and Proquest). A few hundred journal articles
why consumers engage in eWOM communication with the intent were found. In the literature, we found that how eWOM influences
of low benefits but high costs, which is usually found in online consumers’ decisions is far more extensively studied than why
settings. consumers engage in eWOM communication. Studies unrelated
The objective of this study is to provide a better understanding to intention to engage in eWOM communication were excluded
of the motivations behind a consumer’s decision to engage in from the list. This resulted in a total of 11 relevant papers.
eWOM communication. Given its assigned importance, there is Drawing upon Table 1, two observations can be made. First,
clearly a gap in understanding the antecedents of consumers’ most past research has focused on the determinants of negative
intention to engage in eWOM communication, since several eWOM. These studies demonstrated that individuals are likely to
researchers have found that consumers increasingly place more disclose negative eWOM when they want to influence others and
weight on eWOM when making evaluations (e.g. Lee et al. 2008). gain revenge by exerting their power (Grégoire et al. 2010; Ward
In particular, since the motivations for individuals to engage in and Ostrom 2006), when they are motivated by altruism (e.g.,
positive vs. negative eWOM communication are likely to be differ- warning other consumers) (Sparks and Browning 2010), when con-
ent, we separate negative and positive eWOM into two different sumers are part of an individualistic culture (Koh et al. 2010), and
concepts. This approach allows us to understand the relative when they are driven by negative emotions (Verhagen et al. 2013).
importance of different possible determinants. It will also be effec- It represents the efforts made by customers to cause inconvenience
tive for practitioners to identify which factors should be stressed in to or even punish the accused firm for the damages it caused them.
certain situations, and then to develop effective programs. The intention to engage in eWOM communication is motivated by
There are three reasons why we believe this study is important. a desire to ‘‘bring down” a firm in some fashion, or by an intention
First, it addresses an important topic yet largely ignored in the lit- to take possible means to ‘‘get even”.
erature. While eWOM today is an influential source of consumer Second, another stream of research focuses on motivations to
judgments and decisions, much of the research examines the recei- contribute to or participate in online communities. Desire for social
ver’s perspective rather than the senders’ views. What remains interaction (Hennig-Thurau et al. 2004), concern for other con-
unknown is why consumers are inclined to spend their own time sumers (Cheung and Lee 2012), the potential to enhance their
posting eWOM (Yap et al. 2013). The present study is an attempt own self-worth (Hennig-Thurau et al. 2004; Yoo and Gretzel
to address this research gap. Second, a review of the past literature 2008), and a sense of belonging (Cheung and Lee 2012) are distinct
618 J.-R. Fu et al. / Electronic Commerce Research and Applications 14 (2015) 616–630

Table 1
Overview of relevant research about eWOM communication.

Author(s) Theoretical eWOM Method Data collection Key finding(s)

base(s) investigated
Hennig-Thurau et al. – WOM in virtual Experience Convenience sample Desire for social interaction, desire for economic
(2004) community survey (online questionnaire) incentives, concern for other consumers, and the
potential to enhance their own self-worth are the
primary factors leading to eWOM behavior
Sun et al. (2006) Diffusion Opinion Experience University Student eWOM is conceptualized as ‘‘online opinion leadership”.
theory of leadership survey Innovativeness, Internet usage, and Internet social
innovations connection are significant factors for eWOM behavior
Ward and Ostrom (2006) Protest-framing Negative WOM Content Identify tactics (injustice, identity, and agency framing)
theory Analysis consumers used to mobilize mass audiences against a
firm. Consumers ‘‘frame” their corporate betrayal to the
public to demonstrate their power to influence others
and gain revenge
Yoo and Gretzel (2008) – WOM in virtual Experience TripAdvisor users Online travel review writers are mostly motivated by
community Survey helping a travel service provider, concerns for other
consumers, and needs for enjoyment/positive self-
enhancement
Grégoire et al. (2010) Customer Negative WOM Experience [Link] Perceived greed predict customer desire to revenge. And
revenge model Survey users desire to revenge predicts direct and indirect revenge
behaviors
Sparks and Browning Negative WOM Qualitative Complainers tell highly persuasive and credible stories
(2010) often motivated by altruism and revenge
Koh et al. (2010) Hofstede’s Negative Online data Online Dataset & Consumers in individualistic countries are more prone to
theory, movie review analysis and Student sample engage in negative eWOM than consumers in
Survey collectivistic countries
Moldovan et al. (2011) Amount and Experiment University Student Product originality increases buzz (eWOM). And
Valence of originality enhances the effect of usefulness such that
WOM consumers spread relatively more, and more positively
valenced WOM about original and useful products
Cheung and Lee (2012) Public good WOM in virtual Experience Convenience sample To serve the public good, individuals engage in eWOM
community Survey communication in order to benefit oneself (egoism), a
group (collectivism), one or more others (altruism), or to
uphold a principle (principlism)
Yap et al. (2013) Characteristics Experience online consumer panel Cognitive and affective characteristics of messages were
of WOM Survey linked to different motivations to engage in electronic
word-of-mouth, which further differed across positive
and negative messages
Verhagen et al. (2013) Negative WOM Experience User in the consumer Negative eWOM is directly driven by positive and
Survey discussion forum negative emotions and is strongly predictive for the
sender’s intended conduct

reasons for posting eWOM online. These studies are largely of negative WOM. Therefore, a more comprehensive model would
exploratory in nature and tend to view eWOM communication as help to account for the variance beyond satisfaction. The theory of
a monolithic behavior. However, both positive and negative planned behavior is adopted to predict eWOM intention. Figs. 1
WOM communications have different implications for theory and and 2 present our proposed model.
practice (Sundaram et al. 1998). East et al. (2007) investigated
the relative incidence of positive and negative word-of-mouth to 3.1. Consumer satisfaction and justice perception
identify patterns across categories. Their results shed some light
on the issue that the motivation for individuals to engage in posi- Consumer satisfaction, defined as ‘‘the favorability of the indi-
tive word-of-mouth communication is likely to be different from vidual’s subjective evaluations of outcomes and experiences asso-
that of negative actions. ciated with his or her consumption activities” (Brown et al. 2005),
In sum, despite the considerable volume of studies on eWOM, it is a post-purchase response. Satisfaction with products or services
is important to acknowledge that intention to engage in eWOM is usually related to outcomes such as loyalty and positive WOM
communication still remains a very under-researched area. In (e.g., de Matos et al. 2009; Wangenheim and Bayón 2007). In the
particular, what drives individuals to engage in different types shopping context, consumers perceive an emotional imbalance
(positive vs. negative) of eWOM remains less explored. when their expectations are either surpassed (satisfaction) or not
met (dissatisfaction). They will strive to restore homeostasis by
3. Conceptual framework and hypotheses engaging in some post-purchase behaviors (e.g., word-of-mouth)
(Hennig-Thurau et al. 2004).
In order to develop our research framework, we begin by A consumer with a satisfactory service experience is more likely
examining the relationships between satisfaction and eWOM that to respond to the seller in a positive way. Past researchers provide
typically appear in the literature. We then discuss the role of empirical evidence of the significant impact of satisfaction on
justice perceptions which has been established as an important WOM in the offline settings (e.g., Wangenheim and Bayón 2007).
antecedent for satisfaction and its indirect effect on WOM inten- We propose that this reasoning can be extended to the online set-
tion. East (2000) found that although satisfaction may be a neces- ting, where the greater the satisfaction among customers, the
sary condition for WOM to occur, limited association has been greater the possibility that they will intend to engage in positive
found between the degree of dissatisfaction and the probability eWOM communication. Therefore,
 J.-R. Fu et al. / Electronic Commerce Research and Applications 14 (2015) 616–630 619

Beliefs Theory of Planned Behavior

Helping other H9 Attitude
Consumers
H5
H10 H7

Helping the H6 Positive

Subjective eWOM
Company Norm
H11 Intention
H8

Self Perceived
Enhancement Behavioral
Control H1a

Justice Theory
Distributive
Justice
H2
Past
Procedural H3 Experience
Satisfaction
Justice

H4
Interactional Product
Justice Involvement
.
Fig. 1. Research model (positive eWOM).

Beliefs Theory of Planned Behavior

Warning H12 Attitude
other
Consumers H5
H13 H7

Vengeance H6 Negative
Subjective eWOM
Norm Intention
H14
H8
Anxiety
Perceived
Reduction
Behavioral
Control H1b

Justice Theory
Distributive
Justice
H2 Past
Experience
H3 Satisfaction
Procedural
Justice

H4 Advice
Interactional Seeking
Justice

Fig. 2. Research model (negative eWOM).

H1a. Consumer satisfaction has a positive effect on positive that the less satisfied consumers are with the service, the more
eWOM intention. likely their intention to engage in negative eWOM communication.
Hence, it is hypothesized:
Prior research has confirmed that satisfaction is negatively
associated with complaining behavior (Zeelenberg and Pieters H1b. Consumer satisfaction has a negative effect on negative
2004), which implies that consumers will be more likely to mani- eWOM intention.
fest a negative eWOM communication if they are less satisfied with
the service. Thogersen et al. (2009) argued that consumers’ dissat- Researchers have investigated the formation of satisfaction
isfaction with the products or services is the basic antecedent of perceptions. While several theories exist regarding the process
consumers’ complaining intention. Similarly, Voorhees and Brady through which consumers reach satisfaction/dissatisfaction
(2005) found that dissatisfaction would make future complaint judgments, the perspective of justice theory seems particularly of
intentions occur. Extending previous research, therefore, we argue relevance in a service context (Maxham III 2001). Moreover, ample
620 J.-R. Fu et al. / Electronic Commerce Research and Applications 14 (2015) 616–630

empirical evidence reveals that consumers’ justice perceptions are through an assessment of one’s beliefs with respect to the
critical to their satisfaction with a firm’s efforts (e.g., Tax et al. outcomes of a behavior and an evaluation of the desirability of
1998). these outcomes. Before posting their opinion or comments online,
The literature in social psychology indicates that consumers in a consumers are likely to be concerned about whether or not such
service encounter base their justice perceptions on three dimen- behavior is appropriate. Consumers may feel that they should dis-
sions: the perceived justice of the tangible outcome (distributive close their opinion online because airing one’s discontent/content
justice), the perceived justice of the procedures leading to the out- with products or services is one’s moral obligation to benefit the
come (procedural justice), and the perceived justice of the manners public. Phau and Sari (2004) suggest that people are likely to
(e.g., courtesy or rudely) in which consumers were treated complain if they believe it is appropriate. That is, consumers who
throughout the process (interactional justice). These dimensions have a more favorable attitude towards eWOM communication
have been hypothesized and demonstrated to act as significant are more likely to express their opinion and comments on the
cognitive antecedents to satisfaction (e.g., de Matos et al. 2009; Internet. Conversely, a consumer who has a negative attitude
Tax et al. 1998; Wu 2013). Consumers will make judgments about towards eWOM communication would tend to do nothing, or
the degree to which the products/services were just, and in turn may engage in a private action. Accordingly, we posit:
form their satisfaction perceptions. That is, consumers will
perceive more satisfaction if firms provide their service in a just H5. eWOM Attitude has a positive effect on eWOM intention.
manner. Hence, it is hypothesized that:
Subjective norm is a social pressure which represents an
H2. Distributive justice has a positive effect on consumer individual’s perception of whether significant referents think the
satisfaction. behavior should be performed. Chu and Kim (2011) identified
normative interpersonal influence as an important driver of
eWOM behavior. Riivits-Arkonsuo and Leppiman (2013) found that
H3. Procedural justice has a positive effect on consumer consumers do not want to share their negative consumption
satisfaction. experiences on Facebook since negative comments are bound to
or are likely to be traced back to the name of the person who
posted them; thus they prefer to avoid negative posts since these
H4. Interactional justice has a positive effect on consumer are public – instead they would rather share their comments via
satisfaction. offline WOM. Consumers may hesitate if they perceive others’ neg-
ative reactions to such behavior. Stated differently, the possibility
3.2. Theory of planned behavior (TPB) of losing face is conceivably emotionally undesirable. Richins
(1982) noted that some feel they do not like to be seen or to see
Singh and Wilkes (1996) suggested that complaining behavior themselves as nuisances/troublemakers or recommenders/opinion
is goal directed. Extending this reasoning, we propose that individ- leaders. This would inhibit them from posting their opinions or
uals would consider the consequences and implications of the comments about their satisfactory/unsatisfactory experiences.
eWOM communication to decide whether or not to do it. Thus, Therefore,
eWOM communication is the product of an individual’s conse-
quent actions and therefore can be treated as a planned behavior H6. Subjective norms have a positive effect on eWOM intention.
in order to maximize its utility by fulfilling the obligation and dis- Fishbein and Ajzen (1975) have admitted that attitudinal and
posing of any sanctions or consequences (Merchant and Van der normative influences might be dependent on each other. Drawn
Stede 2007). from persuasion theory and cognitive dissonance theory, Lim and
Based on the TPB, intention signifies the motivational compo- Dubinsky (2005) provide explanations of the relationship between
nents of a behavior. It represents the conscious effort that a person subjective norms and attitudes. In order to associate oneself with
is willing to invest in a behavior. Human action is guided by three significant others, cognitive dissonance theory predicts that an
kinds of readily accessible beliefs: behavioral beliefs are those about individual’s attitude toward a specific behavior may change. In
the likely consequences of the behavior, normative beliefs are those contrast, a person’s attitude toward carrying out a behavior can
about the normative expectations and actions of important refer- be influenced through the internalization process of recommenda-
ents, and control beliefs are those about the presence of factors that tions and arguments received from important others, as suggested
may facilitate or impede performance of the behavior (East 2000). by persuasion theory. Thus, positive attitudes may arise directly
In their respective aggregates, behavioral beliefs bring on a favor- out of advice or information from others.
able or unfavorable attitude (ATT) toward the behavior; normative Evidence in past research indicates that attitudes are influenced
beliefs give rise to subjective norms (SN) or perceived social by subjective norms. Chang (1998) found that the model fit was
pressure (which also contribute to the forming of attitudes); and significantly improved when the causal path from subjective
control beliefs result in perceived behavioral control (PBC). norms to attitudes was added. For these reasons, we conclude that
Drawing from the complaint literature in offline settings, consumers’ attitudes toward eWOM communication are influ-
several studies provide empirical support for the impact of the enced by subjective norms.
three variables on consumers’ complaint intention. For example,
consumers’ attitude towards complaining significantly influences H7. Subjective norms have a positive effect on eWOM attitude.
complaining intentions (East 2000; Hansen et al. 2011). It has been
shown to significantly influence customers’ decisions as to Perceived behavioral control is the perception of the ease or dif-
whether or not to voice a complaint. Likewise, subjective norms ficulty of performing a particular behavior. The more opportunities
were found to be significant in predicting consumers’ complaining and requisite resources (e.g., information technology, knowledge,
intentions (East 2000), and perceived control was positively ability) consumers believe they have, the greater their perceived
related to voicing (Bodey and Grace 2006). behavioral control over eWOM communication should be. Empiri-
Attitudes toward eWOM communication is defined as the over- cal evidence suggests that the inclusion of perceived behavioral
all affect of ‘goodness’ or ‘badness’ of eWOM communication and is control in the TPB model has resulted in significant improvements
not specific to a certain episode of satisfaction. It is determined in the prediction of intentions (Ajzen 1991). Accordingly, we posit:
 J.-R. Fu et al. / Electronic Commerce Research and Applications 14 (2015) 616–630 621

H8. Perceived behavioral control has a positive effect on eWOM work and separate this altruistic belief into ‘‘helping other
intention. consumers” and ‘‘warning other consumers”. Thus, consumers
with a favorable experience are likely to display concern by helping
others make the right decision. Conversely, consumers with an
3.3. Antecedents of attitudes toward eWOM communication unfavorable experience are concerned about warning other
consumers to prevent them from experiencing the same problems.
Ajzen (Ajzen 1991) suggested that the TPB is open to expansion. Therefore, a consumer who believes that voluntarily helping or
To improve its explanatory power, many new predictors have been warning others with his/her experience is appropriate will tend
integrated into the original TPB by researchers. In this study, sev- to have a more favorable attitude towards eWOM communication.
eral additional variables were incorporated into the model in order
to improve its predictive power and to enhance the relevance of H9. Helping other consumers has a positive effect on eWOM
the TPB to the particular context under examination. Past studies attitude.
have suggested a range of factors relating to consumers’ engage-
ment in WOM (e.g., Dichter 1966; Hennig-Thurau et al. 2004;
Sundaram et al. 1998; Yap et al. 2013). Based on these studies, H12. Warning other consumers has a positive effect on eWOM
we examined variables which are relevant to consumers’ attitudes attitude.
toward eWOM communication. Three of the variables were identi-
fied to explain positive eWOM communication (i.e., helping the 3.3.2. Reciprocity
company, helping other consumers, and self-enhancement), while Reciprocity is the belief or norm that we should do to others as
the other three variables give reasons for negative eWOM commu- they do to us, calling for positive responses to favorable treatment
nication (i.e., warning other consumers, vengeance, and anxiety and negative responses to unfavorable treatment (Tedeschi et al.
reduction). Furthermore, based on the social psychology literature, 1985). In reciprocity, the more one receives, the more one is obli-
three perspectives were identified to explain consumers’ attitudes gated to give to the other party. We separate this reciprocity belief
towards eWOM communication: altruism, reciprocity, and egoism. into ‘‘helping the company” and ‘‘vengeance”.
A very positive consumption experience creates a consumer’s
3.3.1. Altruism true wish to help and promote the company (Riivits-Arkonsuo
Numerous studies testing TPB have included a measure of and Leppiman 2013). Therefore, we posit that consumers who
moral obligation. A moral obligation is a duty or responsibility that believe that ‘‘a good social institution is worthy of support” tend
someone feels bound to perform because of personal beliefs and to hold a favorable attitude toward eWOM communication in order
values. Altruism is a belief which holds that individuals have a to give the firm ‘‘something in return” and to reward the company
moral obligation to help, serve or benefit others, if necessary at by referring it to others.
the sacrifice of self-interest (Shechter and Freeman 1994). The
inclusion of moral obligation as an antecedent of attitude has been H10. Helping the company has a positive effect on eWOM attitude.
justified on the basis that some behaviors are likely to be regarded
as inherently right or wrong by individuals (Manstead 2000). On the contrary, people who are unhappy with a consumption
Yap et al. (2013) suggested that altruists recognize that the best experience may want to convince others to boycott the offending
way to be helpful in WOM communication is to be informative and organizations, intending to seek vengeance and punish the organi-
functional. Hennig-Thurau et al. (2004) conceptualized ‘‘concern zation (Sundaram et al. 1998; Ward and Ostrom 2006). There is
for others” in terms of both help and warning other consumers. evidence to suggest someone looking to punish an organization
‘‘Concern for others” implies an element of altruism, which is a or vent negative feelings is likely to word his/her eWOM message
voluntary act benefiting others without expecting something in with strong emotional language (Yap et al. 2013). This suggests
return. However, Yap et al. (2013) posited that ‘‘concern for others” that consumers who believe that a bad company should get what
is not likely to be uni-dimensional. We follow Yap et al. (2013) it deserves would have a more favorable attitude toward posting

Table 2
Sample demographics.

Positive eWOM, n = 404 Negative eWOM, n = 365

N % N %
Gender Male 195 48.3 171 46.8
Female 209 51.7 194 53.2
Internet Experience Less than 1 year 2 0.5 2 0.5
1–2 years 4 1.0 4 1.1
2–4 years 15 3.7 15 4.1
4–6 years 40 9.9 36 9.9
More than 6 years 343 84.9 308 84.4
Past experience of Posting WOM online Never 101 25.0 101 27.67
Very few (one time) 128 31.7 110 30.14
A few (2–3 times) 104 25.7 87 23.84
Frequently (4–5 times) 44 10.9 35 9.59
Often (more than 5 times) 27 6.7 32 8.77
Where to Post WOM online Never 56 13.9 51 14.0
Blog 101 25.0 96 26.3
Online Community 116 28.7 104 28.5
Message board 20 5.0 18 4.9
BBS 103 25.5 88 24.1
Others 8 2.0 8 2.2
622 J.-R. Fu et al. / Electronic Commerce Research and Applications 14 (2015) 616–630

Appendix A
Questionnaire of this study.

Distributive Justice (adapted from Grégoire and Fisher (2008)) Strongly disagree Strongly agree
dj01. Overall, the outcomes I received from the seller were fair 1–2–3–4–5–6–7
dj02. Given the time, money and hassle, I got fair outcomes 1–2–3–4–5–6–7
dj03. I got what I deserved 1–2–3–4–5–6–7
Procedural Justice (adapted from Grégoire and Fisher (2008))
pj01. The seller gave me an opportunity to have a say in the handling of the problem 1–2–3–4–5–6–7
pj02. In the handling of my purchase, the seller gave me accurate information 1–2–3–4–5–6–7
pj03. In the handling of my purchase, the seller answered my request in a timely manner 1–2–3–4–5–6–7
pj04. In the handling of my purchase, the seller was flexible in the way it responded to my concerns 1–2–3–4–5–6–7
Interactional Justice (adapted from Grégoire and Fisher (2008))
ij01. The seller who interacted with me gave me detailed explanations and relevant advice 1–2–3–4–5–6–7
ij02. The seller who interacted with me treated me with respect 1–2–3–4–5–6–7
ij03. The seller who interacted with me treated me with empathy 1–2–3–4–5–6–7
Satisfaction with the service experience (Adapted from de Matos et al. (2009))
Based on this experience, how do you evaluate your decision to choose this seller?
sa01. a terrible choice 1–2–3–4–5–6–7 A wonderful
choice
sa02. I am very unsatisfied 1–2–3–4–5–6–7 I am very
satisfied
sa03. Service is awful 1–2–3–4–5–6–7 Service is great
sa04. I feel very unhappy with its service 1–2–3–4–5–6–7 I feel very happy
with its service
Attitude Toward Positive (Negative) eWOM communication (adapted from Kim et al. (2003))
att01. Not posting positive (negative) comments online about this satisfactory (unsatisfactory) experience makes me uneasy 1–2–3–4–5–6–7
att02. It is my duty to post positive (negative) comments online about this satisfactory (unsatisfactory) experience 1–2–3–4–5–6–7
att03. I rarely post positive (negative) comments online when products or services are satisfactory (unsatisfactory) (R) 1–2–3–4–5–6–7
att04. I feel uncomfortable when I tried to post positive (negative) comments online about this satisfactory (unsatisfactory) 1–2–3–4–5–6–7
experience (R)
Subjective Norm (adapted from Liao et al.(2007))
sn01. People important to me supported my posting this positive (negative) experience online 1–2–3–4–5–6–7
sn02. People who influence my behavior would want me to post this positive (negative) experience online 1–2–3–4–5–6–7
sn03. People whose opinions I valued preferred that I post my positive (negative) experiences online 1–2–3–4–5–6–7
Perceived Behavioral Control (adapted from Liao et al.(2007))
pbc01. Posting comments online is entirely within my control 1–2–3–4–5–6–7
pbc02. I would be able to post comments online well 1–2–3–4–5–6–7
pbc03. I have the resources, knowledge, and ability to post comments online 1–2–3–4–5–6–7
Help other Consumers (Adapted from Hennig-Thurau et al. (2004))
hoc01. It my duty to help other consumers with my own positive experiences 1–2–3–4–5–6–7
hoc02. I want to give others the opportunity to buy the right product 1–2–3–4–5–6–7
hoc03. I feel that companies deserve free positive publicity for good service 1–2–3–4–5–6–7
Help The Company (Adapted from Hennig-Thurau et al. (2004))
htc01. If I had positive consumption experience with the company, I should help the company to be successful 1–2–3–4–5–6–7
htc02. In my own opinion, good companies should be supported 1–2–3–4–5–6–7
htc03. It is right to support the company and the product that I like 1–2–3–4–5–6–7
Self Enhancement (Adapted from Hennig-Thurau et al. (2004))
se01. I feel good when I can tell others about my buying successes 1–2–3–4–5–6–7
se02. With my contributions, I can show others that I am a clever customer 1–2–3–4–5–6–7
se03. This way I can express my joy about a good buy 1–2–3–4–5–6–7
Warning other Consumers (Adapted from Hennig-Thurau et al. (2004))
wa01. It’s my duty to warn others of a bad seller 1–2–3–4–5–6–7
wa02. I want to save others from having the same negative experiences as me 1–2–3–4–5–6–7
wa03. I want to warn others of bad product and services 1–2–3–4–5–6–7
Vengeance (Adapted from Hennig-Thurau et al. (2004))
ven01. It is righteous to take actions to get a bad seller in trouble 1–2–3–4–5–6–7
ven02. It is proper to punish a bad seller in some way 1–2–3–4–5–6–7
ven03. I think it is legitimate to cause inconvenience to a bad seller 1–2–3–4–5–6–7
ven04. It is a right thing to do to get even with a bad seller 1–2–3–4–5–6–7
ven05. I want to make a bad seller get what it deserved 1–2–3–4–5–6–7
Anxiety Reduction (Adapted from Hennig-Thurau et al. (2004))
ar01. My contributions help me to shake off frustration about bad buys if I post negative comments online about this 1–2–3–4–5–6–7
experience
ar02. It will get anger off my chest if I post negative comments online about this experience 1–2–3–4–5–6–7
ar03. It helps me reduce frustration after the bad experience If I post negative comments online about this experience 1–2–3–4–5–6–7
Product Involvement (Adapted from Hennig-Thurau et al. (2004))
pi01. I have a strong interest in the product I bought 1–2–3–4–5–6–7
pi02. The product is very important to me 1–2–3–4–5–6–7
pi03. For me, the product does not matter. (R) 1–2–3–4–5–6–7
Advice Seeking (Adapted from Hennig-Thurau et al. (2004))
adv01. I expect to receive tips or support from other users 1–2–3–4–5–6–7
adv02. I hope to receive advice from others that helps me solve my problems 1–2–3–4–5–6–7
 J.-R. Fu et al. / Electronic Commerce Research and Applications 14 (2015) 616–630 623

Appendix A (continued)

adv03. I heard about something new and I want to find out more about it 1–2–3–4–5–6–7
eWOM intention (positive/negative) (adapted from Ajzen (2006))
pw01. I intend to post this positive (negative) buying experience online 1–2–3–4–5–6–7
pw02. I will try to post this positive (negative) buying experience online 1–2–3–4–5–6–7
pw03. I plan to post this positive (negative) buying experience online 1–2–3–4–5–6–7

their eWOM to express their anger, frustration, and irritation to or relevant tend to create excitement and employ eWOM to vent
strike back at the company associated with a negative shopping their positive feelings (Sundaram et al. 1998).
experience. Therefore,
4. Method
H13. Vengeance has a positive effect on eWOM attitude.
4.1. Data collection

3.3.3. Egoism
A cross-section survey design was used to assess the reaction of
Egoism is defined as ‘‘one’s self is, or should be, the motivation
the respondents to a self-reported consumption experience in the
and the goal of one’s own action” (Brandt 1972). A belief is
near-past (within the past 3 months). Data were collected from
regarded as egoistic if the ultimate goal is to enhance one’s own
students in the computer classrooms in a University. After obtain-
welfare or interest (Cheung and Lee 2012). As rational people, indi-
ing the consent of the instructors, three to four trained research
viduals try to maximize their benefits and minimize their cost for
assistants administered the surveys to students in the computer
returns (e.g., recognition, reputation, and pay). While a person can
classrooms. Student participation was voluntary. To encourage
have multiple ultimate goals, the beliefs of altruism and egoism
response, participants who complete the survey will have a chance
can co-occur (Batson 2010).
to win a lottery for supermarket vouchers by entering their email
Stern (2000) asserted that the self-enhancement dimension
address. Using a student sample to investigate consumer behavior
includes egoistic value orientations. Individuals with self-
is not uncommon in the literature and is in line with many of the
enhancement value desire positive recognition from others.
studies which address consumer word of mouth behaviors (e.g.,
Wojnicki and Godes (2008) found that consumers’ propensities
Alexandrov et al. 2013; Davidow 2003; Mdakane et al. 2012; Sun
to generate positive word-of-mouth (WOM) are affected by their
et al. 2006) because college-aged students comprise an important
motivation to self-enhance, that is, to seek experiences that bolster
segment to marketers (Howard et al. 2001; Sun et al. 2006). Basil
the self-concept and seek opportunities to enhance their own self-
et al. (2002) suggested that correlation research is accurate enough
worth. A desired outcome commonly associated with positive self-
with convenience samples. Since our study is not of the propor-
enhancement is to present oneself as an intelligent or discerning
tions of the target audience but of the relationship between vari-
shopper (Hennig-Thurau et al. 2004; Yap et al. 2013). Therefore,
ables, we used convenience sampling to collect our data. In the
we posit that consumers with higher self-enhancement value are
survey, the students were required to recall a recent positive or
inclined to have a more favorable attitude toward posting their
negative consumption experience. The data were collected online
positive eWOM.
by assigning respondents randomly to one of the two scenarios.
As an example, in the negative scenario group, respondents were
H11. Self-Enhancement has a positive effect on eWOM attitude.
instructed to recall a recent negative consumption experience
Alicke et al. (1992) suggested that a desire for catharsis is con- before answering the questions. Respondents wrote the product
sidered as a primary driving force for the relief of negative emo- or service they had consumed on the survey and were asked to
tion. Venting negative feelings through articulation and sharing answer all questions with that specific consumption experience
of negative consumption experiences online can help consumers in mind. East and Uncles (2008) have argued that retrospective
reduce their anxiety and lessen the frustration associated with surveys are a useful way of studying WOM because it is difficult
the events (Sundaram et al. 1998). Thus, consumers who consider to study WOM communication when it occurs. Similar recall
negative eWOM as an effective avenue to ease their anger, anxiety, approaches have been used by Wu (2013), Yap et al. (2013) and
and frustration are likely to have a more favorable attitude toward Thogersen et al. (2009).
negative eWOM communication. All contacted students agreed to participate. Respondents were
assured of anonymity without disclosing their identity by inform-
H14. Anxiety reduction has a positive effect on eWOM attitude. ing them that any information provided would be used only for
academic purposes. The initial number of responses was 842
(417 for the positive version, and 425 for the negative version).
3.4. Control variables Due to missing data, 23 samples were eliminated from the dataset.
To make sure that respondent answers were reliable and authentic,
We used several control variables which in theory are influen- respondents who did not have positive/negative shopping experi-
tial in consumers’ eWOM intention. The first is the past experience ences and those who had not shopped within the past three
of posting WOM online. The uncertainty involved in eWOM com- months were also excluded. In sum, a total of 404 subject
munication would be lowered if consumers were knowledgeable responses with positive service experiences, and 365 subject
or experienced about eWOM communication. The second variable responses with negative consumption experiences, were deemed
is the post-purchase advice-seeking for negative shopping encoun- valid for analysis. Although there is little consensus on the recom-
ters. Consumers who encounter negative experiences tend to share mended sample size for SEM (Sivo et al. 2006), Garver and Mentzer
their experience to obtain advice on how to resolve their problems (1999), and Hoelter (1983) proposed a ‘critical sample size’ of 200.
(Hennig-Thurau et al. 2004). The final control variable is product In other words, as a rule of thumb, any number above 200 is con-
involvement for positive shopping encounters. Consumers whose sidered to provide sufficient statistical power for data analysis.
purchase and use of products are perceived as highly important Another rule of thumb, as proposed by Bagozzi and Yi (1988),
624 J.-R. Fu et al. / Electronic Commerce Research and Applications 14 (2015) 616–630

Table 3
Measurement model.

Positive eWOM scenario Negative eWOM scenario

Construct Indicator Mean stdev Loading t (for k) CR Construct Indicator Mean stdev Loading t CR
Distributive Justice dj01 4.87 1.00 0.88 0.92 Distributive Justice dj01 2.53 1.05 0.91 0.93
(DIST) (DIST)
dj02 4.68 1.02 0.84 14.92 dj02 2.61 1.13 0.90 16.33
dj03 4.89 1.06 0.93 15.93 dj03 2.42 1.04 0.90 17.08
Procedural Justice pj01 5.59 0.99 0.88 0.90 Procedural Justice pj01 2.90 1.28 0.85 0.89
(PROC) (PROC)
pj02 5.68 0.99 0.87 14.81 pj02 2.68 1.30 0.74 8.99
pj03 5.70 1.02 0.85 17.31 pj03 2.76 1.34 0.85 14.42
pj04 5.59 1.03 0.85 16.72 pj04 2.65 1.24 0.82 14.14
Interactional Justice ij01 5.54 1.05 0.89 0.93 Interactional Justice ij01 2.72 1.24 0.78 0.90
(INTER) (INTER)
ij02 5.73 1.00 0.90 19.19 ij02 2.67 1.34 0.92 12.41
ij03 5.48 1.06 0.93 16.96 ij03 2.55 1.20 0.90 12.23
Satisfaction (SAT) sat01 4.60 1.20 0.87 0.91 Satisfaction (SAT) sat01 2.61 1.33 0.98 0.92
sat02 4.62 1.18 0.93 31.56 sat02 2.79 1.29 0.96 31.42
sat03 5.98 1.19 0.90 31.00 sat03 2.82 1.29 0.95 31.34
sat04 5.52 1.29 0.92 30.98 sat04 2.76 1.24 0.91 29.27
Attitude (PATT) patt01 3.43 1.29 0.75 0.83 Attitude (NATT) natt01 4.03 1.39 0.77 0.86
patt02 3.95 1.07 0.76 9.36 natt02 4.31 1.22 0.78 9.13
patt03 3.32 1.24 0.71 6.72 natt03 3.63 1.24 0.75 8.34
patt04 4.59 1.15 0.79 10.04 natt04 4.48 1.21 0.83 11.39
Subjective Norm (SN) sn01 4.20 1.06 0.87 0.93 Subjective Norm (SN) sn01 4.17 1.26 0.96 0.94
sn02 3.98 1.15 0.93 16.67 sn02 4.12 1.25 0.98 30.84
sn03 3.92 1.21 0.90 15.62 sn03 4.11 1.27 0.95 25.70
Positive eWOM pw01 4.63 1.23 0.93 0.96 Negative eWOM nw01 4.66 1.29 0.92 0.96
intention (PWOM) intention (NWOM)
pw02 4.75 1.14 0.94 24.30 nw02 4.64 1.22 0.94 27.27
pw03 4.78 1.11 0.91 22.74 nw03 4.92 1.24 0.89 22.89
pw04 4.39 1.25 0.90 21.46 nw04 4.58 1.25 0.90 23.99
Perceived Behavioral pbc01 5.54 1.21 0.90 0.94 Perceived Behavioral pbc01 5.55 1.18 0.90 0.93
Control (PBC) Control (PBC)
pbc02 5.68 1.10 0.94 18.13 pbc02 5.69 1.06 0.94 16.35
pbc03 5.65 1.13 0.90 16.58 pbc03 5.68 1.08 0.88 15.44
Self Enhancement(SE) se01 5.75 1.06 0.91 0.90 Anxiety Reduction (AR) ar01 4.95 1.24 0.91 0.92
se02 5.36 1.13 0.85 13.82 ar02 5.36 1.18 0.94 18.48
se03 5.41 1.04 0.89 16.87 ar03 5.02 1.05 0.89 19.79
Product Involvement pi01 5.82 0.93 0.86 0.76 Vengeance (VEN) ven01 4.08 1.44 0.92 0.96
(PI)
pi02 5.19 1.03 0.93 10.06 ven02 4.13 1.43 0.93 23.29
pi03 3.93 1.38 0.28 3.09 ven03 3.86 1.51 0.95 26.28
Helping other hoc01 5.34 1.01 0.93 0.93 ven04 3.60 1.49 0.90 22.47
Consumers (HOC)
hoc02 5.32 1.09 0.94 18.14 ven05 3.66 1.42 0.91 19.84
hoc03 5.28 1.10 0.90 18.50 Warning other wa01 5.48 1.19 0.95 0.95
Consumers (WA)
Helping the company htc01 5.49 0.92 0.91 0.95 wa02 5.55 1.17 0.94 21.50
(HTC)
htc02 5.34 0.88 0.93 23.27 wa03 5.23 1.06 0.86 24.52
htc03 5.52 0.97 0.88 18.94 Advice seeking (ADV) adv01 5.17 1.19 0.95 0.94
adv02 5.31 1.16 0.93 20.89
adv03 5.20 1.09 0.91 22.38
Model fit index for CFA Model fit index for CFA
v2 (df) 1264.724 v2 (df) 1293.19
(701) (782)
GFI .946 GFI .940
AGFI .931 AGFI .929
CFI .962 CFI .960
IFI .967 IFI .961
NFI .950 NFI .948
NNFI .957 NNFI .953
RMSEA .039 RMSEA .040

Composite reliability (CR) computed as: qc = (Rk)2/[(Rk)2 + Rvar(e)].

suggests that there should be at least five cases for each latent vari- 4.2. Measurement of the constructs
able in the model for successful model fit. Therefore, our sample
sizes for both groups are acceptable. Table 2 summarizes the The measurement items for the constructs were mostly adopted
demographic characteristics of the respondents. or adapted from previous related studies. The preliminary version
 J.-R. Fu et al. / Electronic Commerce Research and Applications 14 (2015) 616–630 625

Table 4.1
Correlation of latent variables (positive eWOM condition).

The shaded numbers on the leading diagonal are the square root of the variance shared between the constructs and their measures. Off diagonal elements are the correlations
among constructs. For discriminant validity, diagonal elements should be larger than off-diagonal elements.
AVE is average variance extracted (i.e., Proportion of variance in construct that is not due to measurement error) = (Rk2)/[(Rk2)+Rvar(e)].

Table 4.2
Correlation of latent variables (negative eWOM condition).

of this questionnaire was created after conducting literature construct. The constructs were allowed to co-vary in the CFA
reviews and in-depth discussions with three senior professionals. model. The measurement models were evaluated using maximum
Then the survey instrument was reviewed and discussed by prac- likelihood estimation. Internal consistencies and item loadings
titioners and peers within the e-commerce industry. The items greater than 0.70 are considered acceptable (Fornell and Larcker
measuring eWOM intention were adapted from Ajzen (2006). 1981). The results of the SEM analysis (as shown in Table 3)
Items measuring distributive, procedural, and interactional justice revealed an acceptable fit (v2(701) = 1264.72, GFI = 0.946,
were developed by Grégoire and Fisher (2008). The items measur- CFI = 0.962, RMSEA = 0.039 for the positive eWOM scenario, and
ing satisfaction were taken from de Matos et al. (2009). Items v2(782) = 1293.19, GFI = 0.940, CFI = 0.960, RMSEA = 0.040 for the
assessing attitude toward eWOM communication were adapted negative eWOM scenario). The composite reliability (CR) scores
from Kim et al. (2003). Subjective norms and perceived behavioral suggested that the constructs exhibit good internal consistency.
control were measured with the scale developed and validated by Moreover, each item demonstrates high loading on its respective
Liao et al. (2007). Finally, the items measuring helping other con- construct.
sumers, helping the company, self enhancement, warning other To assess discriminant validity, Chin suggested that all con-
consumers, vengeance, anxiety reduction, product involvement, structs should share more variance with their assigned items than
and advice seeking were adapted from Hennig-Thurau et al. with other constructs. In Tables 4.1 and 4.2, the diagonal elements
(2004). The measurement items for the constructs are given in represent the square root of average variance extracted (AVE). A
Appendix A. rule for evaluating discriminant validity is that the AVE square root
should exceed the inter-construct correlations, i.e., the off-diagonal
4.3. Scale validation elements (Chin, 1998). All constructs meet this requirement. The
rule suggests that items should load more highly on their respec-
Before testing the hypotheses, the adequacy of the measure- tive construct than on others. The results of the exploratory factor
ment model was examined by confirmative factor analysis (CFA) analysis (EFA) also show that each item loaded more strongly on its
using the EQS 6.1 program (Bentler and Chou 1987). CFA was con- corresponding construct than on others. Thus these results provide
ducted to assess internal consistency (reliability), item loadings evidence of the acceptable discriminant validity of our measures.
(convergent validity), and discriminant validity. Each item was Finally, using Harmonís’ single factor test, we examined whether
modeled as a reflective indicator of its hypothesized latent our data suffered from common method variance (CMV)
626 J.-R. Fu et al. / Electronic Commerce Research and Applications 14 (2015) 616–630

Beliefs Theory of Planned Behavior

0.297
Helping other (5.257) Attitude
consumers 0.405
R2=0.445 (7.354)
0.117
(1.570) 0.451 Positive
Helping the (8.096)
0.213 eWOM
Company
Subjective (4.287) Intention
0.199 Norm R2=0.550
(3.386)
0.081
Self (1.990)
Enhancement Perceived
Behavioral 0.188
Control 0.115
(4.299)
(2.473)

Justice Theory 0.063

(1.598)
Distributive
Justice 0.050 Past
(0.339) Experience

0.134
Satisfaction
Procedural R2=0.166
Justice (1.980)
Product
0.286
Involvemen
Interactional (3.760)
Justice

1.
Structural path estimates are the standardized parameter estimates, and the associated t-values are in
parentheses. Solid-lines stand for significant paths and dotted-lines for non-significant ones.
Measurement model is not shown above for purposes of clarity.
2.
The strength of the relationships among the constructs is represented by standardized path coefficients.
3.
χ2(721)=1323.102, GFI=.941, AGFI=.921, CFI= .959, IFI=0.947, NFI= .949, NNFI=.943,
RMSEA=0.042.
Fig. 3. Result of SEM analysis (positive eWOM).

Beliefs Theory of Planned Behavior

0.314
Warning other (7.429) Attitude
Consumers R2=0.552 0.369
0.149 (6.647)
(3.294) 0.496 Negative
(10.974)
Vengeance 0.248 eWOM
0.048 Subjective (4.652) Intention
Norm R2=0.646
(0.884)
0.134
(2.527)
Anxiety Perceived
Reduction Behavioral
Control -0.205
0.002
(4.153)
(0.715)
Justice Theory
0.089
Distributive (1.650)
Justice 0.440
(7.848) Past
Satisfaction Experience
Procedural 0.095 R2=0.218
Justice (1.069)
Advice
Seeking
Interactional 0.125
Justice (1.401)

χ (804)= 1523.759, GFI=.939, AGFI=.918, CFI= .950, IFI=.949, NFI= .947, NNFI=0.945, RMSEA=0.048.
. 2

Fig. 4. Result of SEM analysis (negative eWOM).

 J.-R. Fu et al. / Electronic Commerce Research and Applications 14 (2015) 616–630 627

Table 5
Direct and in-direct effect of structural model.

Positive eWOM Negative eWOM

Antecedent Direct effect Indirect effect Total effect Antecedent Direct effect Indirect effect Total effect
HTC 0.033n.s. 0.033n.s. WA 0.116 0.116
SE 0.050 0.050 REV 0.055 0.055
n.s.
HOC 0.094 0.094 AR 0.018 0.018n.s.
DIST 0.004n.s. 0.004n.s. DIST 0.090 0.090
n.s.
PROC 0.027 0.027 PROC 0.019 0.019n.s.
n.s.
INTER 0.058 0.058 INTER 0.026 0.026n.s.
SAT 0.188 0.188 SAT 0.205 0.205
PBC 0.081 0.081 PBC 0.134 0.134
SN 0.213 0.086 0.299 SN 0.248 0.123 0.431
PATT 0.405 0.405 NATT 0.369 0.369
EXP 0.115 0.115 EXP 0.002n.s. 0.002n.s.
PI 0.063n.s. 0.063n.s. ADV 0.053n.s. 0.002n.s.

(Podsakoff et al. 2003). A principal component analysis was con- helping other consumers (b = 0.297, p-value < 0.01) and self
ducted on all indicators. The results suggested that CMV was not enhancement (b = 0.199, p-value < 0.01) significantly influence
significant in our data because no single factor emerged to account attitude, supporting H9 and H11. Helping the company failed to
for the majority of covariance among variables. achieve statistical significance (b = 0.117, p-value > 0.1), indicating
that H10 is not supported. For negative eWOM, warning other con-
5. Hypothesis testing sumers was significant (b = 0.314, p-value < 0.01) as was ven-
geance (b = 0.149, p-value < 0.01), which supported H12 and H13.
The fourteen hypotheses were tested collectively using struc- Finally, anxiety is not significantly related to attitude (b = 0.048,
tural equation modeling (SEM). SEM can support simultaneous p-value > 0.1), thus H14 is not supported. The predicting variables
latent variables with multiple indicators, mediating effects, and of the endogenous constructs and their direct, indirect, and total
causality hypotheses. The results of the SEM analysis revealed an effects on the dependent variable are presented in Table 5.
acceptable fit (v2(721) = 1323.102, GFI = .941, RMSEA = .042 for
the positive eWOM scenario, and v2(804) = 1523.759, GFI = .939,
RMSEA = 0.048 for the negative eWOM scenario). Figs. 3 and 4 6. Discussion and conclusion
summarize the results.
The variance explained (R2) for the structural equations for 6.1. General discussion
intention to engage in positive and negative eWOM communica-
tion is 55.0% and 64.6%, respectively. First, for consumer satisfac- This empirical study manifests the magnitude and relative
tion, the coefficients between satisfaction and positive eWOM importance of various antecedents for intention toward con-
(b = 0.188, p-value < 0.01), and negative eWOM (b = 0.205, sumers’ eWOM communication. It significantly contributes to the
p-value < 0.01) are significant, indicating that H1 is supported. current eWOM literature in several ways. First, the findings reveal
Inspection of the antecedents of satisfaction suggests that the three that satisfaction shows limited association with eWOM intentions.
justice perceptions seem to have inconsistent influence across the This result confirms East (2000) observation which postulated that
two conditions. For the positive eWOM condition, procedural models that rest largely on satisfaction/dissatisfaction are unlikely
justice (b = 0.134, p-value < 0.01, supported H3) and interactional to account for much effect in WOM intentions. This study further
justice (b = 0.286, p-value < 0.01, supported H4) significantly extends past research to the online environment and concludes
influence satisfaction, but distributive justice fails to reach that, although satisfaction/dissatisfaction may be a necessary con-
statistical significance (b = 0.05, p-value > 0.1, H2 is not support). dition for eWOM to occur, it is far from sufficient to explain con-
In contrast, for the negative eWOM condition, distributive justice sumers’ eWOM intention.
is significantly related to satisfaction (b = 0.440, p-value < 0.01, Second, with its ability to account for a substantial variance in
H2 is support), but procedural justice (b = 0.095, p-value > 0.1, H3 the dependent variable, the current study offers evidence of the
is not support) and interactional justice (b = 0.125, p-value > 0.1, applicability of the TPB in understanding consumers’ eWOM
H4 is not support) are not. Therefore, we conclude that H2, H3, intention, and provides new theoretical insights. It suggests that
and H4 are partially supported. consumers would consider the consequences and implications of
Second, for the hypotheses related to TPB, subjective norms eWOM communication before engaging in it. Furthermore, the
appears to be a strong predictor of positive eWOM (b = 0.213, relative importance is not consistent across the two conditions,
p-value < 0.01) and negative eWOM (b = 0.248, p-value < 0.01) which suggests that consumers’ decisions to engage in positive
intention, thus H6 is supported. Subjective norm also strongly vs. negative eWOM communication seemed to be based on differ-
influences attitude (b = 0.451, p-value < 0.01 for positive eWOM, ent criteria. The results in Table 5 reveal that attitude appears to be
b = 0.496, p-value < 0.01 for negative eWOM), providing support the primary determinant of positive eWOM intention, followed by
for H7. Attitude (b = 0.405, p-value < 0.01 for positive eWOM, subjective norms then satisfaction. In contrast, for negative
b = 0.369, p-value < 0.01 for negative eWOM) and perceived behav- eWOM, subjective norms is the most important determinant, while
ioral control (b = 0.081, p-value < 0.01 for positive eWOM, attitude represents the second largest determinant. This suggests
b = 0.134, p-value < 0.01 for negative eWOM) are significantly that the effect of social approval and pressure will be more salient
related to intention, which supports H5 and H8. when the consumer claims that a seller did not perform as
Third, examining the antecedents of attitude, the variance promised, while the spreading of positive eWOM depends mainly
explained for attitudes toward positive and negative eWOM com- on the assessment of whether such behavior is appropriate. In
munication are 45.3% and 55.2%, respectively. For positive eWOM, addition, perceived behavioral control significantly improves
628 J.-R. Fu et al. / Electronic Commerce Research and Applications 14 (2015) 616–630

explanatory power in predicting eWOM intention, in spite of the the antecedents of positive and negative eWOM intention in the
limited amount of variance explained. online environment. The existence of inconsistent relative impact
Third, while a means of enhancing satisfaction is through the regarding the relationships between eWOM intention and its
perception of justice, it is worth noting that the three justice per- antecedents for positive and negative conditions underscore the
ceptions have differential effects on consumers’ satisfaction across importance of investigation aimed at understanding these distinc-
the two conditions. For negative shopping experience, the feeling tions. These distinctions shed new light on understanding why
of satisfaction is largely driven by their perception of distributive consumers engage in eWOM communication.
justice, that is, the degree to which the exchange was perceived We believe our study makes some meaningful extensions and
as justice in the light of tangible matters. This is consistent with contributions to the literature in various ways. First, it contributes
Wu (2013) and Martinez-Tur et al. (2006) in the consumer to the theory by developing a research model anchored in the TPB,
complaint literature. In contrast, for positive shopping experience, and extends it by integrating the justice theory and the social
satisfaction is significantly influenced by the perceptions of inter- psychology literature. By treating eWOM intention as a focal
actional and procedural justice. Thus, when forming perceptions construct, this research goes beyond the traditional WOM and
of satisfaction for a positive encounter, consumers seem to weigh responds to the call of Wien and Olsen (Wien and Olsen 2012) to
buyer–seller interaction and policies more heavily. examine the determinants of eWOM. Second, we found that satis-
Finally, extending the original TPB with consumers’ beliefs of faction exerted a limited influence on consumers’ eWOM intention.
altruism, reciprocity and egoism was deemed valuable. Across Consumers seemed to place more weight on the consequences and
the two conditions, altruism (helping or warning other consumers) implications of the eWOM communication behavior before posting
is the most important determinant of attitude towards eWOM their opinion or comments online. This finding contributes to
communication. In addition, for positive shopping experience, clarifying the relatively small influence of satisfaction. Finally,
consumers with higher self-enhancement value (egoism) tend to we found that consumers’ positive vs. negative eWOM intentions
have a more favorable attitude towards eWOM communication. are associated with different antecedents. This has been implied
For negative shopping experience, consumers report more favor- in East et al. (2007) research, but to the best of our knowledge
able attitudes toward eWOM communication when they believe has never been explicitly examined in the word-of-mouth
that a bad seller should get what it deserves (reciprocity). How- literature. The findings in this study advance the existing theories
ever, helping the company (reciprocity) and anxiety reduction and contribute to future research aimed at enhancing our under-
(egoism) seem to be irrelevant to consumers’ attitudes. standing of consumers’ eWOM intentions. In sum, we believe such
a deeper understanding of antecedents of eWOM intention would
6.2. Limitations and future research assist managers in designing appropriate management policies to
increase favorable outcomes.
The findings of this study should be interpreted in light of a From a managerial standpoint, it is imperative for marketers to
number of limitations. create an environment that is conducive for stimulating positive
First, out study asked respondents to recall a recent shopping eWOM and decreasing negative eWOM as well. For practitioners,
experience before answering the questionnaires. Their evaluations the implications of this study are noticeable. First, our findings
were based on a transaction-specific experience (Oliver 2010) suggest that cultivating positive normative beliefs of significant
which was limited to a given occasion or period. However, con- members’ motivation to comply, as well as positive attitudes is
sumers’ eWOM intention may also be triggered by cumulative important for fostering favorable outcomes. Attitude toward
evaluations (Wien and Olsen 2012) which have developed over eWOM communication is largely driven by altruism. In addition,
time according to the experiences individuals have had with a consumers who intend to post positive eWOM appear to be more
seller. It is suggested that future research test empirically how driven by underlying attitudinal factors, whereas consumers who
these differences might affect consumers’ eWOM intention. consider posting negative eWOM are more driven by social
Second, although common method bias was not significant, the pressure. While companies cannot directly control consumer-to-
use of a post hoc self-report design does not rule out the presence consumer messages, they are still in a position to influence the
of common method variance. Future research can adopt procedural conversations among consumers (Riivits-Arkonsuo and Leppiman
methods to reduce this potential threat. Third, another weakness 2013). Marketers should try to identify ‘social influencers’ in
of the study is the cross-sectional nature of the research design. social media by encouraging users to spread positive eWOM and
Thus, caution must be exercised in the interpretation of these discouraging them from sharing negative information within their
observed associations. Future research can use a longitudinal personal networks (Chu and Kim 2011).
design to better understand the causal relationships among vari- Third, for companies striving to encourage positive eWOM, our
ables. Fourth, the generalizability of the study is limited by the studies suggest that judgment of the process – the procedural and
use of a convenience sample of university students since the use interactional aspects – is more likely to enhance satisfaction and in
of the student sample is likely to impede concluding how con- turn trigger positive eWOM intention. As such, it appears that jus-
sumers other than students would respond. However, given that tice policies and procedures, as well as courteous employee-
university students tend to be heavy users of the Internet and com- consumer interactions, influence consumer perceptions of satisfac-
prise an important segment to marketers, we believe the use of a tion. Although companies should strive to provide distributive jus-
student sample is justified. Finally, a substantial proportion of tice, they may see a greater impact by concentrating on procedural
the variance in eWOM intention remained unexplained. It is possi- and interactional justice. From a procedural and interactional jus-
ble that some variable not included in the models would account tice perspective, clear job descriptions and specific training, as well
for substantial variance. Hence, the research of eWOM intention as compensation policies, would be helpful in providing better
can be enriched if additional theories are applied to the integration service.
of our research framework. Finally, for companies trying to reduce negative eWOM,
distributive justice is the key antecedent of satisfaction. As such,
6.3. Implications companies suffering from negative eWOM should invest more
resources in fixing their failures (such as offering refunds,
The study aims to build on the well-developed theory, TPB, the discounts, and compensation) than in employee communication
justice theory and the social psychology literature, to investigate skills and customer service training. While service policies and
 J.-R. Fu et al. / Electronic Commerce Research and Applications 14 (2015) 616–630 629

employee-consumer interactions are often seen as an integral part Fornell, C., Larcker, D.F., 1981. Evaluating structural equation models with
unobservable variables and measurement error. Journal of Marketing
of the service offering (Schneider et al. 1998), companies can
Research 18, 39–50.
decrease negative eWOM by providing a fair level of tangible out- Fornell, C., Wernerfelt, B., 1987. Defensive marketing strategy by customer
comes to consumers. complaint management: a theoretical analysis. Journal of Marketing Research,
337–346.
Garver, M.S., Mentzer, J.T., 1999. Logistics research methods: employing structural
Acknowledgement equation modeling to test for construct validity. Journal of business logistics 20,
33.
The authors would like to acknowledge a research Grant (NSC Grégoire, Y., Fisher, R.J., 2008. Customer betrayal and retaliation: when your best
customers become your worst enemies. Journal of the Academy of Marketing
102-2410-H-151 -018 -MY2) from the Ministry of Science and
Science 36, 247–261.
Technology in Taiwan, R.O.C. Grégoire, Y., Laufer, D., Tripp, T.M., 2010. A comprehensive model of customer direct
and indirect revenge: understanding the effects of perceived greed and
References customer power. Journal of the Academy of Marketing Science 38, 738–758.
Hansen, H., Samuelsen, B.M., Andreassen, T.W., 2011. Trying to complain: the
impact of self-referencing on complaining intentions. International Journal of
Ajzen, I., 1991. The theory of planned behavior. Organizational Behavior and Human
Consumer Studies 35, 375–382.
Decision Processes 50, 179–211.
Hennig-Thurau, T., Gwinner, K.P., Walsh, G., Gremler, D.D., 2004. Electronic word of
Ajzen, I., 2006. Perceived behavioral control, self-efficacy, locus of control, and the
mouth via consumer opinion platforms: What motivates consumers to
theory of planned behavior. Journal of applied social psychology 32, 665–683.
articulate themselves on the Internet? Journal of Interactive Marketing 18,
Alexandrov, A., Lilly, B., Babakus, E., 2013. The effects of social-and self-motives on
38–52.
the intentions to share positive and negative word of mouth. Journal of the
Hoelter, J.W., 1983. The analysis of covariance structures goodness-of-fit indices.
Academy of Marketing Science 41, 531–546.
Sociological Methods & Research 11, 325–344.
Alicke, M.D., Braun, J.C., Glor, J.E., Klotz, M.L., Magee, J., Sederhoim, H., Siegel, R.,
Howard, P.E.N., Rainie, L., Jones, S., 2001. Days and nights on the internet the impact
1992. Complaining behavior in social interaction. Personality and Social
of a diffusing technology. American Behavioral Scientist 45, 383–404.
Psychology Bulletin 18, 286–295.
Hsiao, C.H., 2011. Predicting online consumer complaints in Northern Taiwan.
Ariely, D., 2007. The Customers’ Revenge. Harvard Business Review 85, 31–43.
African Journal of Business Management 5, 5281–5291.
Bagozzi, R.P., Yi, Y., 1988. On the evaluation of structural equation models. Journal
Kim, C., Kim, S., Im, S., Shin, C., 2003. The effect of attitude and perception on
of the Academy of Marketing Science 16, 74–94.
consumer complaint intentions. Journal of Consumer Marketing 20, 352–
Basil, M.D., Brown, W.J., Bocarnea, M.C., 2002. Differences in univariate values
371.
versus multivariate relationships. Human Communication Research 28, 501–
Koh, N.S., Hu, N., Clemons, E.K., 2010. Do online reviews reflect a product’s true
514.
perceived quality? An investigation of online movie reviews across cultures.
Batson, C.D., 2010. Corsini Encyclopedia of Psychology Altruism. John Wiley & Sons
Electronic Commerce Research and Applications 9, 374–385.
Inc.
Lee, S.J., Lee, Z., 2006. An experimental study of online complaint management in
Bendapudi, N., Berry, L.L., 1997. Customers’ motivations for maintaining
the online feedback forum. Journal of Organizational Computing and Electronic
relationships with service providers. Journal of Retailing 73, 15–37.
Commerce 16, 65–85.
Bentler, P.M., Chou, C.-P., 1987. Practical issues in structural modeling. Sociological
Lee, J., Park, D.H., Han, I., 2008. The effect of negative online consumer reviews on
Methods & Research 16, 78–117.
product attitude: An information processing view. Electronic Commerce
Bodey, K., Grace, D., 2006. Segmenting service ‘‘complainers” and ‘‘non-
Research and Applications 7, 341–352.
complainers” on the basis of consumer characteristics. Journal of Services
Liao, C., Chen, J.-L., Yen, D.C., 2007. Theory of planning behavior (TPB) and customer
Marketing 20, 178–187.
satisfaction in the continued use of e-service: an integrated model. Computers
Brandt, R., 1972. Rationality, egoism, and morality. The Journal of Philosophy, 681–
in Human Behavior 23, 2804–2822.
697.
Lim, H., Dubinsky, A.J., 2005. The theory of planned behavior in e-commerce:
Brown, T.J., Barry, T.E., Dacin, P.A., Gunst, R.F., 2005. Spreading the word:
making a case for interdependencies between salient beliefs. Psychology &
Investigating antecedents of consumers’ positive word-of-mouth intentions
Marketing 22, 833–855.
and behaviors in a retailing context. Journal of the Academy of Marketing
Manstead, A.S.R., 2000. The role of moral norm in the attitude-behavior relation. In:
Science 33, 123–138.
Hogg, D.J.T. (Ed.), Attitudes, behavior, and social context. Lawrence Erlbaum,
Brown, J., Broderick, A.J., Lee, N., 2007. Word of mouth communication within
Mahwah, NJ, pp. 11–30.
online communities: Conceptualizing the online social network. Journal of
Martinez-Tur, V., Peiro, J., Ramos, J., Moliner, C., 2006. Justice perceptions as
Interactive Marketing 21, 2–20.
predictors of customer satisfaction: the impact of distributive, procedural, and
Chang, M.K., 1998. Predicting unethical behavior: a comparison of the theory of
interactional justice. Journal of Applied Social Psychology 36, 100–119.
reasoned action and the theory of planned behavior. Journal of Business Ethics
Mattila, A.S., Patterson, P.G., 2004. The impact of culture on consumers’ perceptions
17, 1825–1834.
of service recovery efforts. Journal of Retailing 80, 196–206.
Chelminski, P., Coulter, R.A., 2011. An examination of consumer advocacy and
Maxham III, J.G., 2001. Service recovery’s influence on consumer satisfaction,
complaining behavior in the context of service failure. Journal of Services
positive word-of-mouth, and purchase intentions. Journal of Business Research
Marketing 25, 361–370.
54, 11–24.
Cheung, C.M.K., Lee, M.K.O., 2012. What drives consumers to spread electronic word
Mdakane, S., Muhia, A., Rajna, T., Botha, E., 2012. Customer relationship satisfaction
of mouth in online consumer-opinion platforms. Decision Support Systems 53,
and revenge behaviors: examining the effects of power. African Journal of
218–225.
Business Management 6, 10445–10457.
Chin, W.W., 1998. The partial least squares approach to structural equation
Merchant, K.A., Van der Stede, W.A., 2007. Management control systems:
modeling. In: Marcoulides, G.A. (Ed.), Modern Methods for Business Research.
performance measurement, evaluation and incentives. Prentice Hall.
Lawrence Erlbaum Associates, Mahwah, NJ, pp. 295–336.
Moldovan, S., Goldenberg, J., Chattopadhyay, A., 2011. The different roles of product
Chu, S.-C., Kim, Y., 2011. Determinants of consumer engagement in electronic word-
originality and usefulness in generating word-of-mouth. International Journal
of-mouth (eWOM) in social networking sites. International Journal of
of Research in Marketing 28, 109–119.
Advertising 30, 47–75.
Oliver, R.L., 2010. Satisfaction: A Behavioral Perspective on the Consumer. ME
Davidow, M., 2003. Have you heard the word? The effect of word of mouth on
Sharpe, New York.
perceived justice, satisfaction and repurchase intentions following complaint
Pfeffer, J., Zorbach, T., Carley, K.M., 2014. Understanding online firestorms: negative
handling. Journal Of Consumer Satisfaction Dissatisfaction And Complaining
word-of-mouth dynamics in social media networks. Journal of Marketing
Behavior 16, 67–80.
Communications 20, 117–128.
de Matos, C.A., Rossi, C.A.V., Veiga, R.T., Vieira, V.A., 2009. Consumer reaction to
Phau, I., Sari, R.P., 2004. Engaging in complaint behaviour: an Indonesian
service failure and recovery: the moderating role of attitude toward
perspective. Marketing Intelligence & Planning 22, 407–426.
complaining. Journal of Services Marketing 23, 462–475.
Podsakoff, P.M., MacKenzie, S.B., Lee, J.-Y., Podsakoff, N.P., 2003. Common method
Dellarocas, C., Wood, C.A., 2008. The sound of silence in online feedback: Estimating
biases in behavioral research: a critical review of the literature and
trading risks in the presence of reporting bias. Management Science 54, 460–
recommended remedies. Journal of applied psychology 88, 879.
476.
Richins, M.L., 1982. An investigation of consumers’ attitudes toward complaining.
Dichter, E., 1966. How word-of-mouth advertising works. Harvard Business Review
Advances in Consumer Research 9, 502–506.
44, 147–160.
Riivits-Arkonsuo, I., Leppiman, A., 2013. Consumer online word-of-mouth-analysis
East, R., 2000. Complaining as planned behavior. Psychology & Marketing 17, 1077–
through an experience pyramid model. International Business - Baltic Business
1095.
Development. Peter Lang Verlag, Tallinn, pp. 297–317.
East, R., Uncles, M.D., 2008. In praise of retrospective surveys. Journal of Marketing
Schneider, B., White, S.S., Paul, M.C., 1998. Linking service climate and customer
Management 24, 929–944.
perceptions of service quality: Tests of a causal model. Journal of Applied
East, R., Hammond, K., Wright, M., 2007. The relative incidence of positive and
Psychology 83, 150.
negative word of mouth: A multi-category study. International Journal of
Shechter, M., Freeman, S., 1994. Nonuse Value: Reflections On the Definition and
Research in Marketing 24, 175–184.
Measurement, Valuing the Environment: Methodological and Measurement
Fishbein, M., Ajzen, I., 1975. Belief, attitude, intention and behavior: An introduction
Issues. Springer, Newyork, pp. 171–194.
to theory and research. Addison-Wesley Pub. Co..
630 J.-R. Fu et al. / Electronic Commerce Research and Applications 14 (2015) 616–630

Singh, J., Wilkes, R.E., 1996. When consumers complain: a path analysis of the key Verhagen, T., Nauta, A., Feldberg, F., 2013. Negative online word-of-mouth:
antecedents of consumer complaint response estimates. Journal of the Academy Behavioral indicator or emotional release? Computers in Human Behavior 29,
of Marketing Science 24, 350–365. 1430–1440.
Sivo, S.A., Fan, X., Witta, E.L., Willse, J.T., 2006. The search for” optimal” cutoff Voorhees, C.M., Brady, M.K., 2005. A service perspective on the drivers of complaint
properties: Fit index criteria in structural equation modeling. The Journal of intentions. Journal of Service Research 8, 192.
Experimental Education 74, 267–288. Wangenheim, F.v., Bayón, T., 2007. The chain from customer satisfaction via word-
Sparks, B.A., Browning, V., 2010. Complaining in Cyberspace: The Motives and of-mouth referrals to new customer acquisition. Journal of the Academy of
Forms of Hotel Guests’ Complaints Online. Journal of Hospitality Marketing & Marketing Science 35, 233–249.
Management 19, 797–818. Ward, J.C., Ostrom, A.L., 2006. Complaining to the masses: the role of protest
Stern, P.C., 2000. New environmental theories: toward a coherent theory of framing in customer-created complaint web sites. Journal of Consumer
environmentally significant behavior. Journal of social issues 56, 407–424. Research 33, 220–230.
Sun, T., Youn, S., Wu, G., Kuntaraporn, M., 2006. Online word-of-mouth (or mouse): Wien, A.H., Olsen, S.O., 2012. Evaluation context’s role in driving positive word-of-
an exploration of its antecedents and consequences. Journal of Computer mouth intentions. Journal of Consumer Behaviour 11, 504–513.
Mediated Communication 11, 1104–1127. Wojnicki, A.C., Godes, D., 2008. Word-of-mouth as self-enhancement. HBS
Sundaram, D.S., Mitra, K., Webster, C., 1998. Word-of-mouth communications: a marketing research paper.
motivational analysis. Advances in Consumer Research 25, 527–531. Wu, I.L., 2013. The antecedents of customer satisfaction and its link to complaint
Tax, S.S., Brown, S.W., Chandrashekaran, M., 1998. Customer evaluations of service intentions in online shopping: an integration of justice, technology, and trust.
complaint experiences: implications for relationship marketing. Journal of International Journal of Information Management 33, 166–176.
Marketing 60, 60–76. Yap, K.B., Soetarto, B., Sweeney, J.C., 2013. The relationship between electronic
Tedeschi, J.T., Lindskold, S., Rosenfeld, P., 1985. Introduction to social psychology. word-of-mouth motivations and message characteristics: The sender’s
West Publishing Co.. perspective. Australasian Marketing Journal.
Thogersen, J., Juhl, H.J., Poulsen, C.S., 2009. Complaining: a function of attitude, Yoo, K.H., Gretzel, U., 2008. What motivates consumers to write online travel
personality, and situation. Psychology and Marketing 26, 760–777. reviews? Information Technology & Tourism 10, 283–295.
Utz, S., Kerkhof, P., van den Bos, J., 2012. Consumers rule: How consumer reviews Zeelenberg, M., Pieters, R., 2004. Beyond valence in customer dissatisfaction: a
influence perceived trustworthiness of online stores. Electronic Commerce review and new findings on behavioral responses to regret and disappointment
Research and Applications 11, 49–58. in failed services. Journal of Business Research 57, 445–455.