Unit 2: Measurement and Survey

1. Measurement in Marketing Research

Measurement refers to the assignment of numbers or symbols to characteristics of objects
according to specific rules. In marketing research, we rarely measure physical objects
directly; instead, we measure characteristics such as customer satisfaction, brand loyalty,
perceived quality, or trust.

Measurement is important because it converts abstract marketing concepts into numerical

data that can be analyzed statistically. Without proper measurement, it is impossible to
compare respondents, test hypotheses, or draw meaningful conclusions.

For example, customer satisfaction cannot be observed directly. Researchers measure it

using indicators such as satisfaction with product quality, service experience, or delivery
time.

2. Concept, Construct, and Variable

A concept is a general abstract idea representing a phenomenon. Examples include
satisfaction, loyalty, trust, and brand image. Concepts exist at a theoretical level and
cannot be directly measured.

A construct is a concept that has been specifically defined for research purposes.
Constructs translate abstract concepts into forms suitable for scientific measurement. For
example, 'customer satisfaction' may be defined as the overall evaluation of a consumer’s
purchase and consumption experience.

A variable is the measurable representation of a construct. Variables can take different

values for different respondents. For example, satisfaction score (1–5), number of repeat
purchases, or likelihood to recommend.

3. Operationalization
Operationalization is the process of defining how a concept will be measured in a
research study. It converts theoretical ideas into measurable variables and indicators.

The steps in operationalization include: defining the concept clearly, identifying its
dimensions, selecting indicators that represent those dimensions, and choosing
appropriate measurement scales.

For example, the concept of customer satisfaction may be operationalized through

variables such as satisfaction with product quality, delivery speed, customer service, and
overall experience.
Logical Flow will beto define:

Concept > Construct > Variable > Indicators

4. Levels of Measurement
There are four levels of measurement commonly used in marketing research.

Nominal Scale: This is the simplest form of measurement. It classifies data into
categories without any order or ranking. Examples include gender, brand name, or region.

Ordinal Scale: This scale provides ranking or order among categories but does not
indicate the magnitude of difference between ranks. For example, ranking brands from
most preferred to least preferred.

Interval Scale: Interval scales have ordered categories with equal intervals between
values but lack a true zero point. Temperature scales such as Celsius are examples.

Ratio Scale: Ratio scales possess all properties of interval scales but also have an absolute
zero point. Examples include age, income, number of purchases, and sales volume.

5. Attitude Measurement
Attitudes refer to a learned predisposition to respond favorably or unfavorably toward an
object such as a product, brand, advertisement, or company.

Attitudes are important because they influence consumer behaviour. Marketers measure
attitudes to understand perceptions, predict purchasing decisions, and evaluate brand
positioning.

Attitudes generally consist of three components: cognitive (beliefs about the object),
affective (feelings toward the object), and behavioral (intentions or actions toward the
object). Some of the scales to measure attitude are:

Likert Scale

Semantic Differential Scale

Stapel Scale

6. Likert Scale
A Likert scale measures respondents’ degree of agreement or disagreement with a
statement. Respondents indicate their level of agreement on a scale such as: Strongly
Disagree, Disagree, Neutral, Agree, Strongly Agree.
For example: 'This brand offers good value for money.' Respondents select the option
that best reflects their opinion.

Advantages of Likert scales include ease of construction, simplicity for respondents, and
reliability in measuring attitudes. However, limitations include response biases such as
central tendency bias and agreement bias.

7. Semantic Differential Scale

The semantic differential scale measures attitudes using bipolar adjective pairs.
Respondents rate an object between two opposite adjectives.

For example:

Innovative ___ ___ ___ ___ ___ Traditional or

Reliable ___ ___ ___ ___ ___ Unreliable.

This scale is particularly useful for measuring brand image and consumer perception
because it captures the emotional meaning associated with a product or brand.

8. Stapel Scale
The Stapel scale measures attitudes using a single adjective and a numerical scale usually
ranging from +5 to -5.

For example, respondents may rate a brand on the attribute 'Innovative' using the scale +5
(very accurate description) to -5 (very inaccurate description).

The advantage of the Stapel scale is that it does not require opposite adjective pairs,
which can sometimes be difficult to create.

9. Questionnaire Design
A questionnaire is a structured set of questions used to collect information from
respondents. It is the most common instrument for collecting primary data in marketing
research.

A well-designed questionnaire ensures that questions are clearly understood, responses

are accurate, and the data collected is relevant to research objectives.

10. Steps in Questionnaire Design

The process begins with specifying the information needed based on research objectives.
The researcher must then determine the type of questionnaire, design question content,
decide on response formats, and carefully word each question.

Questions must be arranged in a logical sequence and the overall layout should be clear
and easy to follow.

Finally, the questionnaire must be pretested through a pilot study to identify and correct
problems before full data collection.

Process:

• Specify information needed.

• Determine type of questionnaire.

• Determine question content.

• Design response format.

• Determine question wording.

• Determine question sequence.

• Design layout.

• Pretest questionnaire

11. Sampling Concepts

Sampling is the process of selecting a subset of individuals from a population to represent
the entire population.

Studying the entire population (census) is often expensive and time-consuming. Sampling
allows researchers to collect data from a smaller group while still making generalizations
about the population.

Key terms include population (entire group), sample (subset selected), sampling unit
(individual element), and sampling frame (list of population elements).

12. Probability Sampling Techniques

In probability sampling each element of the population has a known and non-zero chance
of being selected. This allows researchers to make statistical inferences about the
population.

Simple Random Sampling: Every element has an equal probability of selection. Selection
is usually done using random numbers or computer-generated lists.
Systematic Sampling: Elements are selected at regular intervals from a list. For example,
selecting every 10th customer from a database.

Stratified Sampling: The population is divided into homogeneous subgroups called strata,
and samples are drawn from each stratum. For example, selecting respondents from
different age groups.

Cluster Sampling: The population is divided into clusters (such as cities or regions), and
some clusters are randomly selected for study.

13. Non-Probability Sampling Techniques

In non-probability sampling the probability of selecting each element is unknown. This
method is often used in exploratory research or when a sampling frame is unavailable.

Convenience Sampling: Respondents are selected based on ease of access. For example,
surveying students in a classroom.

Judgment Sampling: The researcher selects respondents based on expertise or judgment

about who will provide useful information.

Quota Sampling: The population is divided into groups and specific quotas are assigned
to ensure representation of each group.

Snowball Sampling: Existing respondents help identify additional respondents. This

technique is useful for studying hard-to-reach populations.

14. Sampling Errors

Sampling error is the difference between the sample result and the true population value.
It occurs because only a portion of the population is studied.

Non-sampling errors arise from mistakes during data collection, such as interviewer bias,
respondent misunderstanding, or data processing errors.

In many cases non-sampling errors can have a greater impact on research accuracy than
sampling errors.

15. Fieldwork and Data Cleaning

Fieldwork refers to the process of collecting data from respondents using surveys,
interviews, or observations. Field investigators must follow standardized procedures to
ensure data quality.
After data collection, responses must be edited, coded, entered into a database, and
cleaned.

Data cleaning involves identifying missing values, correcting inconsistencies, and

removing errors so that the dataset is ready for statistical analysis.