Guidelines for Asking Questions

Derived from the work of Stanley L. Payne. The Art of Asking Questions (Princeton, New Jersey:
Princeton University Press, 1951).

Understanding the Issue:

1.   Decide what it is that you are really after. What is it about a topic that has "hooked" you?
The sooner you can gain clarity about this, the sooner your questions will move toward the focus
you hope for.

2.   Formulate a good research question, one that has many possible answers and a variety of
facets and perspectives, out of which your sub-questions will come. If your research question can
be simply and quickly answered, why ask questions of a variety of respondents anyway?

3.   Try to assess the stage of development of your issue. It would be a mistake, for example, to
conduct a public opinion poll on a topic about which there is little public opinion; the asking of the
question then becomes the making of opinion that was not there before. It would likewise be a
mistake to base your research on issues that have already been resolved in the minds of most
respondents; asking probing questions about possible alternative answers may be thought provoking
for respondents, but in the end you will probably get  pretty much the same answers.

4.   Select a variety of question formats for your questions, matching format to the question
itself. If there is a fairly well defined range of possible answers, for example, you may select a
multiple choice format. 

5.   As in any research, keep asking yourself, "What am I taking for granted?"


The Open-Ended or Free-Answer Question

6.   Ask yourself, Is it really necessary to ask this as an open-ended question? Remember that,
while such responses may be extremely rich, they are also extremely difficult to interpret later in
succinct and/or quantitative form.

7.   If the different points of view on a question are generally well-known, consider converting
the open-ended question to a multiple-choice format. Because the multiple-choice format forces
respondents to answer in terms of pre-determined categories, you may lose some of the richness of
open-ended responses and may fail to capture the outlying answer or opinion. On the other hand,
at the point of analysis you will be forced to interpret these responses in terms of categories anyway;
if these categories are well-known, a multiple-choice format may save you time and energy.

8.   Make open-ended questions as clear, focused, and directive as possible. If they are too broad,
you will get responses from every direction and according to every dimension...a frustrating set of
data to try to make sense of later.
9.   Even though a question is open-ended, you may be able to provide check boxes or scales for
some elements of the response. This is a useful technique for setting the boundaries within which
an open-ended response should operate.


The Two-Way or Either/Or Question

10.  Avoid implied alternatives. Making both sides of the question explicit lends greater clarity
to the response; otherwise, who knows what the respondent concluded was the unspoken
alternative?

11.  Always provide  "don't know" or "no opinion" alternatives except in rare circumstances;
without this option the respondent, wishing to be helpful and cooperative, may feel forced to select
a response which s/he neither holds nor has even thought before the question was posed.

12.  Consider whether there is a reasonable middle-ground which some respondents might take.
If so, you must then decide whether to offer it as a possible response for everyone.

13.  Ordinarily, the choices should be made mutually exclusive. If they cannot be made so, then
you should add a "both" category and perhaps include the combination idea in the question itself.



The Multiple-Choice Question

14.  Even more than in the two-way question, the choices offered in the multiple-choice question
must be mutually exclusive.

15.  None of the alternatives should be excluded or overlooked if a true expression of choices
is desired. If combinations of alternatives are possible, those combinations should be included
among the choices offered.

16.  The choices should be well balanced; that is, an equal number of alternatives should be
presented on every side of the question. The number of alternatives offered does affect the
distribution of replies.

17.  Decide whether you want respondents to express one choice or more than one choice, and
then indicate this clearly. If you want only one response but get several, you may lose important
data.

18.  Always provide for "don't know," "no opinion," or "none of the above" answers to allow
persons whose views fall outside the expected set of responses to answer with integrity.




Other Concerns about Question Format

19.  As a general rule, avoid double-barreled questions that combine more than one idea or set
of alternative responses.

20.  Arrange questions in logical, almost narrative order. Use the questions to stimulate an
orderly conversation from the respondent. Use serialized questions when the same introduction and
possible alternative responses apply.

21.  Avoid the use of double negatives.


The Treatment of Respondents

22.  Avoid the appearance of talking down or otherwise insulting the intelligence of your
respondents. On the other hand, avoid prestige influences like appeals to one's wisdom, knowledge,
fairness, affluence, physical attributes, morals, and devotion to duty.

23.  Use good grammar, but don't make your questions sound stilted.

24.  Don't use slang, or try to be "folksy."

25.  Do what you can to help respondents who, as a general rule, wish to be helpful and provide
you with the information you want. Be careful, however, that in your desire to be helpful you do not
imply that a certain answer would be "better" than other answers. Convey respectful neutrality.

26.   When a respondent's replies are inconsistent, you may discover something by pointing out
this apparent inconsistency in as non-confrontational a way as possible.

27.   Remember that the fine distinctions you as a researcher may have in mind will often not be
understood by a respondent.


About the Words Themselves

28.  Use as few words as necessary. Most questions can be composed in twenty words or less.

29.  Use simple words whenever possible. Avoid jargon.

30.  If you use a synonym, make sure it is actually synonymous with the idea at hand.

31.  Instead of the indefinite "how much?" you can save work for yourself by indicating the terms
in which you want the answers to come -- percentages, dollars, miles, etc.

32.  A personalized question may produce different responses from an impersonalized one.

                                             JWK, 1/96