Chapter 6: Designing Surveys and Data Collection Instruments
TRUE/FALSE
1) The function of the questionnaire is measurement; measurement error is a serious problem
in questionnaire construction.
Answer: True
Questionnaires are designed to measure a wide variety of marketing-related variables. The
way a question is worded can have a profound impact on the resultant error.
2) Identification data collected on a questionnaire are typically used only internally and are
not reported along with research results for reasons of confidentiality.
Answer: True
Identification data includes information such as the respondent’s name, address, and phone
number and is not reported because of confidentiality.
3) Classification data are typically collected among the first five questions of a questionnaire.
Answer: False
Because respondents may be reluctant to disclose personal information, it is best to place
these questions at the end.
4) Questionnaire design has become a fairly exact science that can be mastered by strictly
following the set of rules for their creation.
Answer: False
Questionnaire design is part scientific undertaking and part art form. Slavishly following
fixed steps may lead researchers to devise overly lengthy, uninspired survey instruments.
5) The questionnaire design stage presumes that the research project is well under way and
that research objective and information needs have already been identified.
Answer: True
Questionnaire design is based on information needs; these must therefore be specified prior to
developing any questionnaires.
6) The more homogeneous the respondent group, the more difficult it is to design a single
questionnaire that is universally appropriate.
Answer: False
The more heterogeneous the group (in terms of geodemographics, attitude, and behaviors),
the more difficult questionnaire design becomes.
7) A question should be phrased so as to permit the respondent to admit a lack of knowledge.
Answer: True
When a question implies that the answer should be known, respondents will often answer it
even though they have limited knowledge of the topic, causing measurement error,
8) The aided recall approach, as the name implies, is intended to overcome respondents’
unwillingness to respond to particular questions.
Answer: False
Aided recall attempts to overcome problems of respondent memory and lack of efficiency in
administering questionnaires by providing the respondent with cues and partial information
regarding the event of interest.
9) In mitigating bias resulting from respondents’ unwillingness to respond to questions
accurately, the labeled response categories approach phrases the sensitive question so that it
refers to “other people,” “people you know,” or even “people like you.”
Answer: False
Referring to “other people,” “people you know,” or “people like you” is the indirect statement
approach to mitigating bias resulting from respondents’ unwillingness to respond accurately.
The labeled response category approach presents respondents with a card that lists the
sensitive response alternatives identified by letters or numbers. The respondent uses the letter
or number to indicate a response to the sensitive question.
10) The counterbiasing statement approach involves asking a nonsensitive question prior to
and immediately after a sensitive question.
Answer: False
The counterbiasing statement is a statement suggesting that the subject behavior is rather
common prior to asking a question about the respondents’ behavior.
11) Open-ended questions are most appropriate for exploratory research purposes.
Answer: True
Open-ended questions allow respondents to freely express views divergent from the
researcher’s expectations, making them useful for exploratory research purposes.
12) Open-ended questions are better suited for self-administered questionnaires than for
questionnaires administered by an interviewer.
Answer: False
They are less suited for self-administered questionnaires, because respondents tend to write
more briefly than they speak.
13) The main issues arising in designing multiple-response questions are the number of
alternatives, the alternatives themselves, and interview bias.
Answer: False
Interviewer bias is not present in multiple-choice questions. The third issue is position bias of
the response alternative.
14) A dichotomous question is an extreme form of the multiple-choice question.
Answer: True
A dichotomous question is a multiple-response question that limits the choices to two (yes or
no, did or did not, agree or disagree) or at most three (if a neutral alternative is added).
15) In multiple-response questions, the order in which the response alternatives are listed
tends to bias the data.
Answer: True
The order of response alternatives can impact the results. If possible, responses alternatives
should be randomized for each respondent.
16) In a telephone survey, recency bias occurs if respondents tend to ignore alternative
responses heard after the first few are given.
Answer: False
This is primacy bias. Recency bias results from choosing among the last alternatives given.
17) The issue of whether to include a neutral response alternative in a dichotomous question
is best decided on the basis of the proportion of the respondent group who are truly neutral.
Answer: True
Because only two responses are possible, the central issue in dichotomous question design is
whether respondents should be allowed a neutral or no opinion type of response. If a
substantial number of respondents are truly neutral, the inclusion of the neutral alternative
should increase the accuracy of the results.
18) The questionnaire should be comprehensible to that segment of the population with the
least vocabulary skills.
Answer: True
The words used in the questionnaire must be comprehensible to nearly all persons likely to be
asked to complete it.
19) Leading questions are often needed to accurately assess respondents’ feelings.
Answer: False
A leading question is one that gives respondents a cue as to what the answer should be, thus
biasing the results.
20) When deciding question wording, words like “usually,” “regularly,” “kind,” “normally,”
and “frequently” are considered ambiguous and should be avoided.
Answer: True
Because people interpret these words differently, they are ambiguous and should be avoided.
21) The overriding concern when deciding the sequence of questions is that it retain
respondent interest and attention without introducing bias.
Answer: True
To obtain accurate answers, researchers must be careful not to bias answers and at the same
time maintain interest and attention of the respondent.
22) To reduce sequence bias, ask specific questions first then ask general questions second.
Answer: True
To reduce sequence bias, the general questions should come first.
23) The open-ended response format can be used in the pretest to determine appropriate
response categories for what will become a multiple-response question in the final
questionnaire.
Answer: True
This is an effective method of assuring that all of the possible responses are identified.
24) An observational form should require the observer to summarize rather than list in detail
the behavior observed.
Answer: False
Observational forms should permit the observer to record the behavior in detail, rather than
requiring a summary regarding a number of behavior patterns.
25) Due to substantial extra expenses when doing international marketing research,
questionnaires that are determined to be accurately translated from one language to another
generally do not need to be pre-tested as well.
Answer: False
Complexities of international research can add a great deal of pre- and post-processing time
and typically cost at least double what a survey costs when run domestically. However, it is
not adequate to omit pre-testing.
MULTIPLE CHOICE
1) Measurement error in the design of questionnaires
a. can easily be calculated within a given number of standard deviations
b. may be of unknown direction and magnitude
c. does not require the degree of concern that is typically associated with sampling error
d. results from questions which measure something other than what they are intended to
measure
e. both b and d
Answer: E
Unlike sampling error, measurement error can’t easily be calculated and may be of unknown
direction and magnitude. It is often the largest source of preventable error in marketing
research, far exceeding errors in sampling plans. It results from questions which measure
something other than what they are intended to measure.
2) The _______________ section of the questionnaire asks for such information as the
respondent’s name, address, and phone number.
a. identification data
b. request for cooperation
c. instructions
d. classification data
e. information sought
Answer: A
The identification data section asks for the respondent’s name, address, and phone number. It
is often placed first on the questionnaire.
3) The _______________ component of the questionnaire is an opening statement designed
to enlist the respondent’s help regarding the interview.
a. identification data
b. request for cooperation
c. instructions
d. information sought
e. classification data
Answer: B
The request for cooperation is the second part of the questionnaire and is used to enlist the
respondent’s help in completing the interview.
4) Often researchers will elect to include the purpose of the study in the _______________
section, if disclosing it will not bias respondent outcomes
a. identification data
b. request for cooperation
c. instructions
d. information sought
e. classification data
Answer: B
This information, if provided, is placed in the request for cooperation before the instructions
are given.
5) The instructions are comments to _______________ about how to interact with and
eventually complete the questionnaire
a. the respondent
b. the interviewer
c. both the interviewer and respondent
d. the interviewer or the respondent
Answer: D
The instructions are written for either the interviewer or respondent, depending on the type of
questionnaire and the role of the interviewer in the process.
6) The _______________ section of the questionnaire forms the major, and invariably the
longest portion of the questionnaire.
a. identification data
b. request for cooperation
c. instructions
d. information sought
e. classification data
Answer: D
This part of the questionnaire solicits the information that is requested.
7) The _______________ section of the questionnaire contains the characteristics of the
respondent, primarily geodemographic data.
a. identification data
b. request for cooperation
c. instructions
d. classification data
e. information sought
Answer: D
This data should locate respondents and allow one to place them in factual categories.
8) The classification data section of the questionnaire contains the characteristics of the
respondent, which is primarily
a. geodemographic data
b. personal historical data
c. psychographic data
d. behavioral data
Answer: A
Geodemographic data includes demographic information, such as gender and age, and
geographic information, such as zip code.
9) From a detailed listing of _______________, the researcher can begin to design the
questionnaire
a. the population and respondent group
b. measurement scales and communication media
c. information needs and a clear definition of the respondent group
d. research objectives and sampling plan
Answer: C
To design a questionnaire, researchers need to have a list of the information needs and a clear
outline of the respondent group that will be answering the questions.
10) Each question included in the questionnaire should relate to a specific
a. research objective
b. information need
c. measurement scale
d. marketing problem
Answer: B
Each question included on the questionnaire should relate to a specific information need.
11) Assuming that the desired data are relevant to the decision problem, the researcher must
be sensitive to the respondent’s ability to provide the data, including all of the following
except
a. whether the respondent is uninformed
b. whether the respondent is forgetful
c. whether the respondent is unwilling to answer
d. whether the respondent misremembers
Answer: C
The ability to answer accurately assumes the respondent is willing to answer the question but
does not have the ability to do so.
12) What factors affect a respondent’s ability to remember a particular event?
a. recency of the event
b. importance of the event
c. repetition of the event
d. all of the above
e. both a and c
Answer: D
The probability of forgetting is influenced by the importance of the event and how often it is
repeated. The rate of forgetting is quite rapid over the first few days and continues with the
passage of time.
13) To help individuals recall an event, without biasing information, researchers can use all of
the following methods except
a. unaided recall
b. aided recall
c. labeled response categories
d. recognition
Answer: C
The three main methods are unaided recall, aided recall, and recognition.
14) Several studies suggest that questions that rely on unaided recall can
a. overestimate the actual occurrence of the event
b. underestimate the actual occurrence of the event
c. cause misremembering, i.e. remembering the wrong information
d. be effective if pre-tested with recognition tests
Answer: B
Studies suggest that individuals tend to underestimate the actual occurrence of the event.
15) An example of using _______________ would be asking respondents the question “What
advertisements do you remember seeing last week?”
a. unaided recall
b. aided recall
c. recognition
d. pre-testing
Answer: A
Unaided recall does not provide any clues or hints.
16) The recognition method
a. is a form of unaided recall
b. may create suggestion bias
c. involves presenting to the respondent the actual event
d. all of the above
e. both b and c
Answer: E
With the recognition method, the researcher simply asks whether a respondent has in fact
come across or experienced something. Recognition is the most biasing of the recall methods.
17) To reduce measurement error from respondents misremembering, researchers can use all
of the following methods except
a. use recognition tests
b. use scales that provide appropriate time referents
c. ask the questions in multiple ways to different respondents
d. rely more on aided recall
Answer: A
Using recognition tests may actually increase the occurrences of misremembering.
18) All of the following are approaches that have been developed to mitigate bias resulting
from respondents’ unwillingness to respond to questions accurately except
a. labeled response categories
b. recognition questions
c. indirect statements
d. counterbiasing statements
Answer: B
The approaches developed include counterbiasing statements, indirect statements, labeled
response categories, and randomized response technique.
19) An approach that lists alternative responses to a sensitive question by letters or numbers
and permits the respondent to answer with a letter or number employs
a. randomized responses
b. labeled response categories
c. counterbiasing statements
d. indirect responses
e. split design
Answer: B
The labeled response categories method presents the respondent with a card that lists the
sensitive response alternatives and has the responses identified by letters or numbers. The
respondent uses the letter or number to indicate a response to the sensitive question.
20) In mitigating bias resulting from respondents’ unwillingness to respond to questions
accurately, the _______________ approach begins the question with a statement suggesting
that the subject behavior is rather common and then ask the respondent the question.
a. counterbiasing statement
b. indirect statement
c. labeled response categories
d. randomized response technique
Answer: A
Sometimes the unwillingness to provide an accurate response can be overcome by using a
counterbiasing statement, suggesting the behavior is rather common.
21) In mitigating bias resulting from respondents’ unwillingness to respond to questions
accurately, the _______________ approach phrases the question with words such as “other
people” or “people you know.”
a. counterbiasing statement
b. indirect statement
c. labeled response categories
d. randomized response technique
Answer: B
With sensitive or embarrassing questions that individuals may be reluctant to answer
honestly, using an indirect approach will often be successful.
22) Open-ended questions
a. are costly
b. reduce interviewer bias
c. permit diverse responses which may not have been anticipated
d. all of the above
e. both a and c
Answer: E
Open-ended question allows respondents to freely express views divergent from the
researcher’s expectations. Disadvantages include high potential for interviewer bias, because
interviewers usually must summarize the respondents’ answers and delete the aspects deemed
unimportant, and the time and cost associated with coding the responses.
23) All of the following would be considered advantages of open-ended questions except
a. can serve as an excellent first question on a topic
b. are relatively easy to code and analyze
c. can establish rapport and gain the respondent’s cooperation
d. exert minimal influence on subsequent responses, because respondents are not biased by
predetermined set of response alternatives
Answer: B
Open-ended questions are time consuming and costly to code and analyze.
24) A major disadvantage of using _______________ questions is interviewer bias while
recording the respondent’s answers.
a. open-ended
b. multiple-choice
c. dichotomous
d. indirect
Answer: A
Of the three types of question formats, open-ended questions have the highest potential for
interviewer bias since the interviewer must attempt to record the answer exactly as it was
given, which is rarely done.
25) To code verbal data in the “excruciatingly correct” manner requires considerable time and
_______________ people.
a. three
b. five
c. six
d. eight
Answer: C
Although coding can be done with fewer people, using 6 ensures a much higher level of
accuracy.
26) When designing multiple-response questions, response alternatives should be
_______________, which means the set of responses should include all of the possible
alternatives.
a. collectively exhaustive
b. mutually exclusive
c. position unbiased
d. label inclusive
Answer: A
Collectively exhaustive implies all possible alternatives are included. Sometimes this is done
by adding an “other” category.
27) When designing multiple-response questions, response alternatives should be
_______________, which means the respondent should be able to identify exactly one
alternative that clearly represents their response.
a. collectively exhaustive
b. mutually exclusive
c. position unbiased
d. label inclusive
Answer: B
Mutually exclusive means that a respondent is able to identify one response that clearly
represents their answer.
28) Relative to open-ended questions, multiple-choice questions
a. increase interviewer bias
b. are more difficult to administer
c. reduce the time and cost associated with data processing
d. all of the above
e. both a and c
Answer: C
Multiple-choice questions reduce interviewer bias (in attempting to interpret verbal
responses) and the cost and time associated with data processing. They are also typically easy
and quick to administer.
29) In deciding on how many response alternatives to put in a question, the rule of thumb is
that the explicit alternatives should contain _______________ percent of the responses.
a. 75
b. 85
c. 95
d. 99
Answer: C
Less than 5 percent of the responses should be “other,” and 95 percent should fall within the
listed alternatives.
30) When a list of numbers (such as prices or how many times per week a respondent visits a
store) is given for the alternatives in a multiple-response question, there is a high probably
that position bias will occur. In this case, the bias would be
a. toward the lowest numbers
b. toward the central position
c. toward the highest numbers
d. toward the first number in the series
e. toward the last number in the series
Answer: B
When numbers are given in an array, the position bias is towards the center of the array.
31) When response alternatives are listed for a multiple-response question, the
_______________ alternative has a greater chance of being chosen.
a. first
b. last
c. center
d. penultimate
Answer: A
The first alternative tends to be selected more often than any other.
32) In a telephone interview, if respondents tend to choose among answers they heard last
from a list of alternative responses, then _______________ is present.
a. modal answer bias
b. recency bias
c. primacy bias
d. interviewer bias
Answer: B
Recency bias involves choosing an answer they heard last.
33) A(n) _______________ question is one that gives the respondent a cue as to what the
answer should be.
a. ambiguous
b. biasing
c. counter-biasing
d. leading
e. double-barreled
Answer: D
Leading questions often reflect the researcher’s or decision maker’s viewpoint and hints at
what the answer should be.
34) The use of a brand or company name in a question
a. can cause the respondent to believe that the company or brand is the sponsor of the survey
b. can promote a more positive response to the brand or company
c. make affect open-ended questions that follow
d. all of the above
Answer: D
By using a brand name or company name in a question, respondents may believe that
company is the sponsor and will tend to be more positive to that brand or company. In
addition, that brand or company can impact responses to open-ended questions that follow.
35) A(n) _______________ question includes words or phrases that are emotionally colored
and that suggest a feeling of approval or disapproval.
a. leading
b. ambiguous
c. biasing
d. double-barreled
Answer: C
By using emotionally charged words or phrases, researchers can bias the responses.
36) The question “Are you in favor of curtailing the amount of sugar allowed in children’s
cereals?” violates which of the following guidelines for question wording?
a. use simple language
b. avoid implicit assumptions
c. avoid estimates
d. avoid biasing questions
e. avoid double-barreled questions
Answer: B
The idea that sugar is detrimental is implicit as an assumption built into built into the
question. A better-worded question makes the nature of the question or choice explicit, like
“Are you in favor of curtailing the amount of sugar allowed in children’s cereal if it will
result in lower rates of tooth decay, the use of artificial sweeteners, and a 10 percent cost
increase?”
37) The question “How many cans of soft drinks do you purchase in a year?” violates which
of the following guidelines for question wording?
a. avoid implicit alternatives
b. avoid implicit assumptions
c. avoid estimates
d. avoid biasing questions
e. avoid double-barreled questions
Answer: C
This is a poorly worded question because respondents would have to estimate how many they
purchased in a year. The time frame should be a smaller unit, such as week or possibly a
month.
38) The question “What is your evaluation of the restaurant’s appearance and service?”
violates which of the following guidelines for question wording?
a. avoid leading questions
b. avoid implicit assumptions
c. avoid estimates
d. avoid biasing questions
e. avoid double-barreled questions
Answer: E
Double-barreled questions ask two questions in one. In this case appearance and service are
two concepts.
39) Sequence any questions that are embarrassing, sensitive, complex, or dull
a. at the beginning of the questionnaire
b. near the beginning of the questionnaire
c. well down in the questionnaire
d. with a more interesting question before and after
Answer: C
These type of questions should be asked only after rapport is built with the respondent, late in
the question sequence.
40) Which of the following are not among the guidelines for determining question sequence?
a. Arrange questions in logical order.
b. Ask general questions last.
c. Place uninteresting and difficult questions early in sequence to make the rest of the
questionnaire seem easier.
d. Use a simple and interesting opening question.
e. both b and c
Answer: E
Questions should be arranged in a logical order; general questions should be asked before
specific questions; uninteresting and difficult questions should be placed late in the sequence;
and the opening question should be simple and interesting.
SHORT ANSWER
1) Identify the five sections of a questionnaire.
Answer: The sections are:
1) identification data
2) request for cooperation
3) instructions
4) information sought
5) classification data
2) What can cause respondents to be unable to answer a question accurately?
Answer: Respondents may be unable to answer accurately if they are uninformed, forgetful,
or (unintentionally) misremember.
3) What are four methods for mitigating bias from respondents being unwilling to answer
sensitive questions?
Answer: Methods to mitigate this bias involve using a counterbiasing statement, an indirect
statement, labeled response categories, or the randomized response technique.
4) What are the steps for coding of responses to open-ended questions?
Answer: Step 1. Two reviewers independently read through all responses.
Step 2. Each reviewer forms a comprehensive list of key words and phrases.
Step 3. Together with a third reviewer, the two lists should be merged into one list.
Step 4. Have two additional individuals sort the responses based on the categories created in
the third step.
Step 5. The two individuals from Step 4 with another new individual should come to an
agreement on the sorting of responses.
Note that the first 3 steps should be considered adequate, whereas steps 4 and 5 are
considered “excruciatingly correct” or too costly to be practicable.
5) Give an example of a double-barreled questions.
Answer: Student examples will obviously vary. A double-barreled question is one in which
the wording calls for two responses, such as “What is your evaluation of the snowmobile’s
ride and acceleration?” Test to see if the question can be split into two separate questions that
could have different answers.
ESSAY
1) Outline the broad interrelated guidelines in constructing a questionnaire.
Answer: The guidelines are:
1) review preliminary considerations
2) decide on question content
3) decide on response format
4) decide on question wording
5) decide on question sequence
6) decide on physical characteristics
7) carry out pre-testing and revision
8) make the final draft
2) Describe the randomized response technique—why it is used, how it is applied, and an
example of its use.
Answer: With the randomized response technique, the respondent is presented with two
questions, one sensitive, one not. A randomizing device, such as a coin toss, is used to decide
which question the respondent answers. The interviewer does not know the answer to the
non-sensitive question and does not get to see the result of the randomizing device.
For example, the sensitive question might be “Have you had hemorrhoids in the past month?”
and the non-sensitive question might be “Do you have the same birthday as your best
friend?” If the randomizing device has a probability of one half, the proportion of
respondents whose answer would be “Yes” on the sensitive question can be calculated as
equal to twice the incidence of respondents answering “Yes” overall minus the probability of
answering “Yes” on the non-sensitive question (in this case, one twelfth).
3) Discuss the advantages and disadvantages of open-ended questions.
Answer: Advantages include:
1) can serve as an excellent first question on a topic
2) can establish rapport and gain the respondent’s cooperation in answering more specific
questions
3) exert minimal influence on subsequent responses
4) can provide researchers with additional insights
Disadvantages include:
1) high potential for interviewer bias
2) time and cost associated with coding the responses
3) extra weight given to respondents who are more articulate
4) Discuss the advantages and disadvantages of using multiple-response questions.
Answer: Advantages include:
1) reduce interviewer bias
2) reduce cost and time associated with coding open-ended questions
3) good for self-administered questionnaires
4) researchers can limit responses to a particular set that matters
Disadvantages include:
1) design of effective questions requires considerable time and cost upfront
2) typically an exploratory study using open-ended questions is required beforehand
3) tend to bias the data by the order in which alternative responses are listed
5) How questions are worded impacts the type of response that will be received. Discuss the
general guidelines for appropriate question wording.
Answer: Guidelines for appropriate question wording include:
1) simple language
2) use unambiguous words
3) avoid leading questions
4) avoid biasing questions
5) avoid implicit alternatives
6) avoid implicit assumptions
7) avoid estimates
8) avoid double-barreled questions
9) consider the frame of reference
Test Bank for Modern Marketing Research: Concepts, Methods, and Cases
Fred M. Feinberg, Thomas Kinnear, James R. Taylor
9781133188964, 9781133191025, 9780759391710
