SURVEY RESEARCHthought about violent crimes, drunk driving may
have appeared to be a less important issue to them.
By contrast, after we asked about drunk driving as
a crime, respondents may have expressed less con-
cern about crime in general.
We need to remember that context effects are
strong if the question is ambiguous because respon-
dents will draw on the context to interpret and under-
stand the question. Previous questions on the same
topic and ones immediately preceding a question can
have a large context effect. For example, Sudman et
al. (1996:90–91) contrasted three ways of asking
how much a respondent followed politics. When
they asked the question alone, about 21 percent of
respondents said they followed politics “now and
then” or “hardly at all.” When they asked the
question after asking about something that the
respondent’s elected representative recently did,
the percentage who said they did not follow nearly
doubled (39 percent). The knowledge question made
many respondents feel that they did not really know
much. When a question about the amount of “pub-
lic relations work” the elected representative pro-
vided to the area came between the two questions, 29
percent of respondents said they did not follow pol-
itics. This question gave respondents an excuse for
not knowing the first question; they could blame
their representative for their ignorance. The context
of a question can make a difference, and researchers
need to be aware of it at all times: “Question com-
prehension is not merely a function of the wording
of a question. Respondents use information provided
by the context of the question to determine its
intended meaning” (Sudman et al., 1996:69).Layout and Format.There are two format or lay-
out issues: the overall physical layout of the ques-
tionnaire and the format of questions and responses.
Questionnaire Layout. Layout of a questionnaire
is important both to an interviewer and for the
respondent.^43 Questionnaires should be clear, neat,
and easy to follow. Put identifying information (e.g.,
name of organization) on questionnaires and give
each question a number. Never cramp questions
together or create a confusing appearance. A few
cents saved in postage or printing will ultimately
cost more in terms of lower validity resulting from
a lower response rate or of confusion of interview-
ers and respondents. A professional appearance
with high-quality graphics, space between ques-
tions, and good layout encourages accuracy and
completeness and helps the questionnaire flow. If
using an interview format, create a face sheetas part
of the questionnaire for administrative use. The face
sheet should include the time and date of the inter-
view, the interviewer’s name, the respondent’s iden-
tification number, and the interviewer’s comments
and observations on the interview.
Give interviewers and respondents instructions.
It is best to print instructions in a different style from
the questions (e.g., in a different color or font) to dis-
tinguish them. This helps an interviewer to distin-
guish between questions for respondents and
instructions intended for the interviewer alone.
Layout is crucial for mail and Web question-
naires because there is no friendly interviewer to
interact with the respondent. Instead, the question-
naire’s appearance persuades the respondent.
Include a polite, professional cover letter on let-
terhead stationery with mail surveys, identifying the
researcher and offering a telephone number for ques-
tions. Details matter. Respondents will be turned off
if they receive a bulky brown envelope with bulk
postage addressed to Occupant or if the question-
naire does not fit into the return envelope.
Web surveys are still new, and researchers are
just learning which design features are most effec-
tive, but visual design details matter (see Web sur-
vey discussion later in this chapter).Question Format.You must decide on a format for
questions and responses. Should respondents circle
responses, check boxes, fill in dots, or write in a
blank? The principle is to make responding clear and
unambiguous. Boxes or brackets to be checked and
numbers to be circled are usually clearest. Also, list-
ing responses down a page rather than across makes
them easier to see (see Expansion Box 8, Question
Format Examples). Use arrows and instructions for
contingency questions. Visual aids are helpful. For
example, hand out thermometer-like drawings to
respondents when asking whether their feeling