CHAPTER 8
UNDERSTANDING MEASUREMENT, DEVELOPING
QUESTIONS, AND DESIGNING THE QUESTIONNAIRE
LEARNING OBJECTIVES
In this chapter you will learn:
8-1 To understand the basics of measurement regarding people, places, and things
8-2 To recognize the three types of measures used by marketing researchers
8-3 To become familiar with scale formats commonly used by marketing researchers
8-5 To learn the dos and don’ts of question wording
8-7 To understand the advantages of computer-assisted questionnaire design software
8-8 To comprehend coding of questionnaires
CHAPTER OUTLINE
Basic Concepts in Measurement
• Measurement
• Properties
• Objective properties
• Subjective properties
• Scale development
Types of Measures
• Nominal Measures
• Ordinal Measures
• Scale Measures
o Ratio scale
o Interval scale
Interval Scales Commonly Used in Marketing Research
• The Likert Scale
o Lifestyle inventory
• The Semantic Differential Scale
o Halo effect
• The Stapel Scale
• Issues with Interval Scales Used in Marketing Research
o Symmetric interval scale
o Nonsymmetric interval scale
• The Scale Should Fit the construct
Reliability and Validity of Measurements
Designing a Questionnaire
• The Questionnaire Design Process
Developing Questions
• Four Dos of Question Wording
o The question should be focused on a single issue or topic
o The question should be brief
o The question should be grammatically simple
o The question should be crystal clear
• Four Don’ts of Question Wording
o Do not “lead” the respondent to a particular answer
o Do not use “loaded” working or phrasing
o Do not use a “double–barreled” question
o Do not use words that overstate the case
Questionnaire Organization
• The Introduction
o Who is doing the survey
o What is the survey about
o How did you pick me
o Motivate me to participate
o Am I qualified to take part
• Question Flow
Computer-Assisted Questionnaire Design
• Question Creation
• Skip and Display Logic
• Data Collection and Creation of Data Files
• Coding the Questionnaire
• Coding the Questionnaire
• Pretesting the Questionnaire
• Ready-Made Respondents
• Data Analysis, Graphs, and Downloading Data
KEY TERMS
Measurement Properties
Objective properties Subjective properties
Scale development Nominal measures
Ordinal measures Scale measures
Interval scale Ratio scale
Workhorse scales Likert scale
Lifestyle inventory Semantic differential scale
Halo effect Symmetric interval scale
Nonsymmetric interval scale Reliable measure
Valid measure Questionnaire
Questionnaire design Question bias
Question development Question evaluation
Leading question Loaded question
Double-barreled question Overstated question
Face validity Questionnaire organization
Introduction Undisguised survey
Disguised survey Incentives
Anonymity Confidentiality
Screening questions Question flow
Warm-up questions Transitions
Skip question Classification questions
Computer-assisted questionnaire design Skip logic
Display logic Coding
“All that apply” question Pretest
Stapel scale
TEACHING SUGGESTIONS
1. Measurement of a construct requires an operational definition. Instructors may want
to review this concept and elaborate on how an operational definition indicates how a
construct’s subjective and/or objective properties are measured on a scale. For
instance, “purchase intention” may be operationally defined as a respondent’s
indication on a 7-point scale where 1 corresponds to “not at all likely” and 7 pertains
to “very likely” to purchase the brand at his/her next purchase.
2. Scale development in an academic context is much more rigorous than in an applied
context. If an instructor desires to bring this difference to students’ attention,
3. To help students understand the idea of measurable qualities, have them generate as
many different rating aspects of one or more of the following:
• A convenience store
• A department store
• A new automobile
Use class discussion to point out how customers may harbor positive ratings on some
qualities, but negative ratings on others, and how this information has important
marketing implications.
4. Numerical versus letter grades can be used as an example of how higher-level scaling
assumptions provide more information than lower-level ones. For instance, take
numerical grades. Suppose one student has an 89 average, and another student has an
81 average. Assuming that 80-89 defines the “B” letter grade range, both are B
students, but the 89 student is close to an A, while the 81 student is close to a C. This
closeness is not contained in the ordinal letter grade system, but it is in the numerical
ratio system.
Class discussion can focus on whether it is better to have a ratio scale (1-100 scale) or
an ordinal scale (A-F scale) as a way to measure the goodness of students. How
would students feel about a nominal system of pass-or-fail. Hopefully students will
desire a more sensitive scale rather than a less sensitive one. The point is that with a
5. It is useful to remind students of the inability of observation and most other
qualitative methods to look into the minds of those people being studied. The reason
for scaling stems from a need to standardize respondents’ answers into a consistent
format that can be used to summarize and compare their answers.
An in-class exercise can be used to illustrate the need for and benefits of using
standardized scales. Select some subject with some controversial or emotional aspect
to it. Possible college examples include: dropping football from the athletic program
to save money, requiring freshmen and sophomores to live on campus, creating a
walking campus with no student or faculty vehicles allowed, or not awarding a
diploma unless the graduating student signs an agreement to contribute 1 percent of
his or her income to the alumni fund. Have students write down how they “feel”
about the issue. Gather up the written comments and assign the task of compiling the
reactions to a team of three or four students. Once done, have other students observe
how the team goes about the compilation task.
Class discussion should bring out the fact that the qualitative information required
judgments along the way by the compilation team. Next, have students rate their
feelings on the controversial topic using a 7-point negative-to-positive scale. Have a
student team compile these responses with the rest of the class as observers. It should
be apparent that the use of a rating scale greatly simplifies the judgments and
compilation activities.
6. Life styles and their measurement is a good research topic. Assign the topic to a
student or a student team to perform library research on it and report findings to the
class. You may want to expand the assignment to include VALs which is an Active
Learning Exercise in the chapter. Sending students to the SRI VALS home page is
always a fun experience for students. With a multimedia teaching room, instructors
can visit SRI’s web site and demonstrate VALS in class.
7. It is sometimes possible to demonstrate the halo effect to students by having them
inspect the questionnaires from a survey that used a semantic differential or similar
scale. Sometimes a respondent will place check marks all on one side, the other, or
down the middle of the scale. Alternatively, students can see that a respondent has
used the scale correctly when the check marks are consistent with the reverse
positions of the scale items.
8. Contrasting a telephone questionnaire to one that is self-administered will emphasize
how telephone administration constrains the choice of scales and further requires the
interviewer to describe the scale to the respondent. With self-administered
questionnaires, the format and instructions are facilitated by the visual context.
Divide the class into two teams and have one team design a telephone questionnaire,
while the other designs a mail questionnaire. The topic might be to determine the
satisfaction level of car buyers with the automobile’s style, handling, fuel efficiency,
and passenger compartment roominess. The need to tailor the questions to the form
of data collection can be emphasized by comparing the two questionnaires.
9. The 4 “do’s” and 4 “do not’s” described in the chapter are not easy for students to
remember. A “divide and conquer” strategy is to assign each student (or student
team) one “do” or one “do not” with instructions to formulate a question that
illustrates the error quite well. Students can share the questions they formulate in
class. You might have a contest by class vote as to which one is the best (easiest to
relate to or remember) for each of the 4 do’s and each of the 4 do not’s items.
Prepare a list of the winners for distribution the next time you teach the class.
10. There are other computer-assisted questionnaire design programs on the market.
Many of these companies have web sites, and some allow downloading of evaluation
copies of their products. One approach is to have students or student teams review
and demonstrate different computer-assisted questionnaire design software. Of
course, we are partial to Qualtics, but this approach will reveal alternatives with
particular capabilities or features that apply to special types of questionnaire
situations or question form. If student need help identifying these programs, here is a
list that was supplied on the “SurveyMonkey” web site at one time.
Zoomerang™
SuperSurvey
Greenfield Online
Infopoll
Perseus
PollCat
Inquisite
Cool Surveys
Survey System
Apian Software
Hosted Survey
SurveyView
StatPac
SurveyGold
Survey Select
InstantSurvey
EZSurvey
Mercator
SurveyCrafter
PollPro
SurveyHeaven
Surveywire
ObjectPlanet
SurveySaid
mantaINSIGHT
Active Websurvey
SumQuest
CustomerSat
StatSurvey
SurveySite
QuickSense
SurveyTrends
LiveSurveys
InSite
Websurveyor
ACTIVE LEARNING EXERCISES
Construct a College Student Life Style Inventory
Students must develop a life style statement for each dimension using the
recommendations in the Likert scale section – “With this scale, it is best to use “flat” or
plain statements and let the respondent indicate the intensity of his or her feelings by
using the agree-disagree response continuum position.” Representative statements are
provided in the table.
College Life Style
Dimension
Your statement
Studying
I am a studious person.
Going out
I enjoy going out.
Working
I work to gain experience in my major.
Exercising
I exercise regularly
Shopping
Shopping is a large part of my life.
Dating
I usually go out on dates as much as the average student.
Spending money
I like to make purchases.
Can you identify what is “bad” about a question and correct it?
Bad Version of the
Question
What’s the error?
Good Version of the
Question
How do you feel about car
seats for infants?
Lacks focus
How do you feel about
requiring infants to ride in
car seats?
When your toddler wants to
ride in the car with you as
you run errands or pick up
your older children at
school, practice, or some
friend’s home, do you use
an infant
car seat?
Not grammatically simple
Does your toddler usually
ride in an infant car seat?
If using an infant car seat is
not convenient for you to
use, or when you are in a
hurry and your toddler is
crying, do you still go
ahead and use the infant
car seat?
Double barrel (not
convenient; in a hurry;
toddler is crying)
Make into three questions
or three question parts: Do
you use an infant car seat
when:
It is not convenient?
You are in a hurry?
Your toddler is crying?
How much do you think you
should have to pay for an
infant seat that retrains and
protects your toddler in
case someone runs into
your car or you lose control
of your car and run into a
light post or some other
object?
Not grammatically simple
How much do you think
you should pay for an infant
seat that protects you child
in a car accident?
Shouldn’t concerned
parents of toddlers use
infant car seats?
Leading question
Parents of toddlers should
use infant car seats. (agree-
disagree scale)
infant car seats should be
used for your loved ones?
Do you think that parents
who are responsible citizens
and who are aware of
driving dangers use infant
car seats?
Double barreled
Should parents use infant
seats with their children?
If you had an accident with
your toddler on board, do
you believe an infant car
seat could protect your
child from being maimed?
Overstated
Do you believe that infant
car seats are valuable?
Decide on Questionnaire Order in a Questionnaire
How to measure?
Location on the Questionnaire
and Reason(s) for This Location
Describe the restaurant
concept and asked
intentions to purchase
there on a scale.
Place after the warm-ups as the
description takes concentration on
the part of the respondent and
there is an intentions scale used
Determine respondents’
preferences for each of the
several possible design
features on a preference
scale
Logically follows the description
of the new restaurant so
respondents can relate these
questions to the restaurant concept.
Describe standard entrees
and example house
specials and obtain how
much respondents are
willingness using price
ranges
Logically follows the preferences
for new restaurant so respondents
can relate these questions to the
expected prices.
Determine farthest driving
distance respondents are
willing to drive to the new
restaurant for each
location feature.
Logically follows the description
of the new restaurant and may be
before or after the preferences so
respondents can relate these
questions to the expected prices.
Ask for demographics of
Demographics are placed at the
the respondents
end of the questionnaire
Determine normal use of
various local media such
as newspaper, radio,
television and obtain
specifics such as what
newspaper sections are
read, what radio
programming, and what
local television news
times watched
These are quick and easy to
answer, so they can be warm-ups
or otherwise placed at the
beginning of the questionnaire.
ANSWERS TO END-OF-CHAPTER QUESTIONS
1. What is measurement? In your answer, differentiate an object from its properties,
both objective and subjective.
2. Distinguish the three measures used in marketing research.
3. How does an interval scale differ from a ratio scale?
4. Explain what is meant by a continuum along which subjective properties of an object
can be measured.
5. What are the arguments for and against the inclusion of a neutral response position
in a symmetric scale?
6. Distinguish among a Likert scale, a lifestyle scale, and a semantic differential scale.
With the modified Likert scale, respondents are asked to indicate their degree of
agreement or disagreement on a symmetric agree-disagree scale for each of a series of
7. What is the halo effect, and how does a researcher control for it?
8. Provide questions to measure each of the constructs that follow. Before you construct
the measure, consult a source book to find a concise definition of the construct.
Relate the definition and then provide the question.
Suggested questions follow each construct.
a. Brand loyalty
9. How does reliability differ from validity? In your answer, define each term.
10. What is a questionnaire, and what functions does it serve?
11. What is meant by the statement that questionnaire design is a systematic process?
Questionnaire design is a systematic process in which the researcher contemplates
various question formats, considers a number of factors characterizing the survey at
12. What are the four guidelines or “dos” for question wording?
13. What are the four “don’ts” for question wording? Describe each.
The Question Should Not “Lead” the Respondent to a Particular Answer. A leading
question is worded or structured in such a way as to give the respondent a strong cue
14. What is the purpose of a questionnaire introduction, and what things should it
accomplish?