McDaniel & Gates – Marketing Research, 9th Edition Instructor’s Manual
CHAPTER 10
The Concept of Measurement
LEARNING OBJECTIVES
1. To understand the concept of measurement.
3. To understand the four levels of scales and their typical usage.
4. To explore the concepts of reliability and validity
KEY TERMS
Measurement Rule
Constructs Constitutive definition
Operational definition Scale
Nominal scales Ordinal scales
Interval scales Ratio scales
Reliability Test-retest reliability
CHAPTER SCAN
This chapter examines the concept of measurement, a process at the heart of marketing research.
Measurement is the process of assigning numbers or labels to objects, persons, states, or events
in accordance with specific rules to represent quantities or qualities of attributes. The process
consists of six steps. 1) Identify the concept of interest; 2) Develop a construct; 3) Define the
concept constitutively; 4) Define the concept operationally; 5) Develop a measurement scale; 6)
McDaniel & Gates – Marketing Research, 9th Edition Instructor’s Manual
CHAPTER OUTLINE
1. The Measurement Process
I. Measurement
2. Step One: Identify the Concept of Interest
3. Step Two: Develop a Construct
4. Step Three: Define the Concept Constitutively
5. Step Four: Define the Concept Operationally
I. Operational Definition
6. Step Five: Develop a Measurement Scale
I. Scale
A. Scale Defined
II. Nominal Level of Measurement
A. Nominal Scales Defined
III. Ordinal Level of Measurement
A. Ordinal Scales Defined
2. Ranking
3. Controversial Use of Ordinal Scales
IV. Interval Level of Measurement
7. Step Six: Evaluate the Reliability and Validity of the Measurement
I. Accurate Data Imply Accurate Measurement
II. Sources of Measurement Errors
A. Systematic Error Defined
B. Random Error Defined
C. Sources of Measurement Differences
III. Reliability
A. Reliability Defined
B. Three Ways to Assess Reliability
IV. Validity
A. Validity Defined
B. Validity: Number of Different Perspectives
8. Summary
CHAPTER SUMMARY
1. MEASUREMENT PROCESS
I. Measurement
A. Measurement Defined–the process of assigning numbers or labels to persons, objects, or
events in accordance with specific rules for representing quantities or qualities of attributes
II. Rule
A. Rule Defined–a guide, a method, or a command that tells a researcher what to do
1. Problems encountered with rules
a. Lack of clarity or specificity
b. Difficulty of devising rules to assess the true value of consumer attributes
See Exhibit 10.1 Measurement Process (p. 277)
2. STEP ONE: IDENTIFY THE CONCEPT OF INTEREST
3. STEP TWO: DEVELOP A CONSTRUCT
I. Constructs
A. Constructs Defined–specific types of concepts that exist at higher levels of abstraction.
1. Constructs–invented for theoretical use and thus are likely to cut across various preexisting
1. Examples of marketing constructs: brand loyalty, involvement, channel power
4. STEP THREE: DEFINE THE CONCEPT CONSTITUTIVELY
McDaniel & Gates – Marketing Research, 9th Edition Instructor’s Manual
I. Constitutive Definition
A. Constitutive (or Theoretical or Conceptual) Definition Defined–
A statement of the central idea or concept under study, establishing its boundaries and fully
distinguishing it from all other similar but different concepts.
B. Constructs of a Scientific Theory–defined constitutively–all constructs, to be capable of
being used in theories, must possess constitutive meaning
5. STEP FOUR: DEFINE THE CONCEPT OPERATIONALLY
I. Operational Definition
1. Because it is overly restrictive in marketing to insist that all variables be operational defined in
directly measurable terms, many variables are stated in abstract terms
6. STEP FIVE: DEVELOP A MEASUREMENT SCALE
I. Scale
A. Scale Defined–a set of symbols or numbers so constructed that the symbols or numbers can
be assigned by a rule to the individuals (or their behaviors or attitudes) to whom the scale is
McDaniel & Gates – Marketing Research, 9th Edition Instructor’s Manual
A. Nominal Scales Defined–scales that partitions data into categories that are mutually
exclusive and collectively exhaustive. An example is gender, where the respondent must choose
either female or male.
III. Ordinal Level of Measurement
A. Ordinal Scales Defined–maintain the labeling characteristics of nominal scales plus an
ability to order data
1. Postulate–an assumption that is an essential prerequisite to carrying out an operation or line of
thinking
2. Transitivity postulate–is the notion that “if a is greater than b, and b is greater than c, then a
is greater than c.”
3. Ranking–the objective of an ordinal scale–ordinal numbers are used strictly to indicate rank
order. Example of an ordinal scale follows:
Please rank the following brands of fax machines from 1 to 5 with 1 being the most preferred and
4. Controversial use of ordinal scales–rate various characteristics–if the researcher can justify
(1) (2) (3) (4) (5)
Very Certain Certain Neutral Uncertain Very Uncertain
See Global Research: Construct Equivalence Problems Often Occur in Global Marketing
Research (p 282)
Examples of construct equivalence problems include:
1) Functional Equivalence. An example is the consideration of beer in England, Germany, and
Scandinavia as an alcoholic beverage but as a soft drink in Italy, Spain, or Greece.
2) Conceptual Equivalence. In-groups in the U.S. include one’s own country and out-groups
3) Definitional Equivalence. In France, fragrance is measured on a hot-cold continuum. In the
U.S. and UK, this is not an attribute assigned to fragrances.
2. Arbitrary zero point of interval scales restricts the statements that a research can make about
the scale points
3. Interval scales–amenable to computation of an arithmetic mean, standard deviation, and
correlation coefficient. The more powerful parametric statistical tests such as t tests and F tests
can be applied
V. Ratio Level of Measurement
A. Ratio Scales Defined–have all the characteristics of those scales previously discussed as well
as a meaningful absolute zero or origin–reflects the actual amount of a variable
1. Examples–physical characteristics of a respondent such as age, weight, and height. Other ratio
scales are based on area, distance, money values, return rate, population counts, and lapsed
periods of time
See From the Front Line: Making Sure You Have the Right Level of Scaling (p. 284)
1. What are the differences between nominal and ordinal scales?
2. What level of scaling would be needed when applying cross tabulations or multivariate
procedures to the data?
3. Can multivariate procedures be used when you have ranking data?
7. STEP SIX: EVALUATE THE RELIABILITY AND VALIDITY OF THE
MEASUREMENT
I. Accurate Data Imply Accurate Measurement
A. M = A where M refers to measurement and A stands for complete accuracy–this is rarely
achieved. Instead
M = A + E
where E = errors
II. Sources of Measurement Errors
A. Systematic Error Defined–an error that results in a constant bias in the measurements,
caused by faults in the measurement instrument or process.
B. Random Error Defined–influences the measurements but not systematically. Thus, random
error is transient in nature.
C. Sources of Measurement Differences
2. Differences due to a stable characteristics of individual respondents, such as personality,
values, and intelligence.
4. Differences caused by situational factors such as distractions or others present in the
interview situation.
5. Differences resulting from variations in administering the survey. Interviewers can ask
6. Differences due to the sampling of items included in the questionnaire. When researchers
7. Differences due to a lack of clarity in the measurement instrument. A question may be
ambiguous, complex, or incorrectly interpreted.
8. Differences due to mechanical or instrument factors. Blurred questionnaires, lack of space
to fully record answers, missing pages in a questionnaire, or a balky pen can result in differences
in responses.
See Practicing Marketing Research: The Truth about What Respondents Say (p 287)
Martin Weinberger’s 25 years of experience in the research industry has led him to develop
several perspectives regarding measurement in consumer research.
1. Sometimes consumers don’t tell the truth; sometimes they are willing to reveal it.
3. Sometimes consumers don’t know why they buy the brands they buy.
1. What can researchers do to determine if the consumers’ responses are true?
2. Do you think that the vast majority of respondents attempt to give true answers?
III. Reliability
A. Reliability Defined–a measurement scale is reliable if provides consistent results over time
and is free from random error
B. Three Ways to Assess Reliability
1. Test-Retest Reliability–by repeating the measurement using the same instrument under as
nearly the same conditions as possible
1) It may be difficult to locate and gain the cooperation of respondents for a second testing
3) Environmental or personal factors may change
McDaniel & Gates – Marketing Research, 9th Edition Instructor’s Manual
2. Equivalent Form Reliability–Two equivalent but different instruments are created and given
to the same sample. Equivalent form reliability is determined by measuring the correlation of the
scores on the two instruments.
a. Problems with equivalent forms
2) Even if equivalence can be achieved it may not be worth the time, trouble, and expense
3. Internal Consistency Reliability–ability to produce the similar results using different
samples to measure a phenomenon during the same time period. The theory of internal
consistency rests on the concept of equivalence
a. Equivalence–how much error may be introduced by using different samples of items to
measure a phenomenon; it focuses on variations at one point in time among samples of items.
Can test for item equivalence by assessing the homogeneity of a set of items
b. Split-half technique–scale items should be randomly assigned to one half or the other, the
total scores of the two halves are then correlated
1) Problem with Split-half technique–the estimate of the coefficient of reliability is totally
dependent on how the items were split. Different splits result in different correlations
2) Cronbach alpha technique–can is used to overcome the split-half technique. It involves
4) KR-20 technique–applicable for all dichotomous or nominally scaled items
1. Measuring instrument–reliable if not then it will not yield consistent results when measuring
2. Scale or other measuring device–worthless to a researcher if it lacks validity because it is not
measuring what it is supposed to measure.
B. Validity: Number of Different Perspectives
See Exhibit 10.5 Assessing the Validity of a Measurement Instrument (p 290)
1. Face Validity–the weakest form of validity. It is concerned with the degree to which a
measurement “looks like” it measures what it is supposed to.
2. Content Validity–representativeness or sampling adequacy of the content of the measurement
instrument. Does the scale provide adequate coverage of the topic under study?
2) Second, an exhaustive literature search and focus groups could be conducted to identify all
possible items for inclusion on the scale.
4) Fourth, the scale could be pretested and an open-ended question asked that might identify
other items to be included.
3. Criterion-Related Validity–refers to the ability of a measuring instrument to predict a
variable that is designated a criterion.
a. Predictive validity–refers to the extent to which a future level of a criterion variable can be
predicted by a current measurement on a scale
b. Concurrent validity is concerned with the relationship between the predictor variable and the
criterion variable, both of which are assessed at the same point in time.
4. Construct validity–theoretical foundations underlying the obtained measurements; degree of
correspondence between researcher’s theory/expected pattern of results and the observed pattern
of results
a. Convergent validity–to the degree of correlation among different measures that purport to
1. Do you think that respondents are more likely to give true answers in a group setting such as a
focus group?
2. What about individual depth interviews?
IV. Reliability and Validity – A Concluding Comment
A, Situation 1–holes all over the target–indicates a complete lack of consistency thus there is no
reliability in the instrument being used
B. Situation 2–very tight pattern (consistency) but the pattern is far removed from the bull’s
See Exhibit 10.6 Illustrations of Possible Reliability and Validity Situations in
Measurement (p. 294)
See Practicing Marketing Research: Tips for Managers on Putting Measurement to Work
(p. 294)
Craig Bailey, president of Customer Centricity, Inc found that the survey instrument his
company was using to measure customer satisfaction was not getting the maximum value
4. Every measurement must have clear and rational goals.
5. If a “carrot and stick” is not clear, measurement will fall into disuse.
Discussion Questions:
1. Do you think that every manager’s bonus in an organization should be based upon
measurement of customer satisfaction?
2. Some people say that all you need to measure customer satisfaction is one question: “Would
you recommend our company to a friend?” Do you think that this is sufficient? (See Real-Life
8. SUMMARY
QUESTIONS FOR REVIEW AND CRITICAL THINKING
1. What is measurement?
2. Differentiate between the four types of measurement scales, and discuss the types of
information contained in each.
Nominal scales partition data into categories that are mutually exclusive and collectively
exhaustive. Every observation will fall into one of several categories and into only one. An
example of a category in a nominal scale is: What is your sex? _____ Male ______Female
Ordinal scales have the labels of nominal scales plus the ability to order data. Numbers are used
to rank the categories. The numbers are not indicative of any value or amount, but are simply to
put in rank order.
3. How does reliability differ from validity? Give examples of each.
4. Give an example of a scale that would be reliable but not valid. Also give an example of a
scale that would be reliable but not valid. Also give an example of a scale that would be
valid but not reliable.
If we were trying to measure customer satisfaction, but the questionnaire consistently measured
5. What are three methods of assessing reliability?
Test-retest reliability – Repeat the same survey two weeks later with the same sample of
people.
6. What are three methods of assessing validity?
Face validity – The researcher has to make a judgment as to whether the measure “looks like” it
measures what it is supposed to.
Content validity – Does the scale provide adequate coverage of the topic under study? For
McDaniel & Gates – Marketing Research, 9th Edition Instructor’s Manual
Criterion-related validity – How well can the measurement instrument predict a variable?
7. Divide the class into teams of four or five. Each team should then go to the Internet and
find the results of a survey with data. Each team should then determine the face validity
and content validity of the research. Also, each team should suggest a method of assessing
reliability of the survey.
Student results will vary.
WORKING THE NET
1. Go to a Web search engine and look up “validity and reliability.” Describe to the class
the new insights you gain into these important concepts.
Responses will vary.
REAL-LIFE RESEARCH
Case 10.1 Can Shopping Styles Be the Same around the World? (p 298)
Key Points:
Global consumer packaged goods (CPG) shoppers should not be defined by their country
but instead by their shopping attitudes and behavior—especially their attitudes toward price and
their propensity to preplan their CPG purchases, according to a study conducted by Ipsos
McDaniel & Gates – Marketing Research, 9th Edition Instructor’s Manual
• Responsible Planners (17 percent): Shoppers in this group are also on a fixed budget
(like the Price-Driven shoppers) and always compare prices. They are the shoppers most
Questions
1. What is the construct in the research? Give a constitutive definition.
2. What level(s) of measurement could the researchers have used? Give examples.
3. What could the researchers have used to measure reliability?
4. How might the researchers approach the measurement of validity?
5. What can the researchers do to demonstrate construct validity?
6. Do you see this scale as having value to marketing researchers? Why?
Case 10.2 Measuring Customer Satisfaction at GE (p 328)
Key Points:
• GE uses a version of “Net Promoter” to measure customer satisfaction. Net promoter
McDaniel & Gates – Marketing Research, 9th Edition Instructor’s Manual
Questions:
1. What is the construct that GE is trying to measure?
2. How might GE assess construct validity?
3. Does “service recovery” play a big role in customer satisfaction?
4. What is the operational definition of the construct?
5. Does the scale adequately address the underlying construct?
The adequateness of the scale is currently debated, particularly in its use as a single item form
McDaniel & Gates – Marketing Research, 9th Edition Instructor’s Manual
6. What level of measurement is being used in the research? Could a higher level have been
used?