Tips for Structured Self Report

    Although
self-reports represent a powerful mechanism for obtaining data, researchers who
use this approach should always be aware of the risk of response biases that
is, the tendency of respondents to distort their responses. Perhaps the most
pervasive problem is people’s tendency to present a favorable image of
themselves. 

    Social desirability response bias refers to the tendency of some
individuals to misrepresent their responses consistently by giving answers that
are consistent with prevailing social values. This problem is often difficult
to combat. Subtle, indirect, and delicately worded questioning can sometimes
help to alleviate this response bias. The creation of a permissive atmosphere
and provisions for respondent anonymity also encourage frankness.

    Some
response biases are most commonly observed in composite scales. These biases
are sometimes referred to as response sets. Scale scores are seldom entirely
accurate and pure measures of the critical variable. A number of irrelevant
factors are also being measured at the same time. 

    Because response set factors
can influence or bias responses to a considerable degree, investigators who
construct scales must attempt to eliminate or minimize them. Extreme responses
are an example of a response set that introduces biases when some individuals
consistently select extreme alternatives (eg, “strongly agree”). 

    These extreme
responses distort the findings because they do not necessarily signify the most
intense feelings about the phenomenon under study. There is little a researcher
can do to counteract this bias, but there are procedures for detecting it. Some
people have been found to agree with statements regardless of content. Such
people are called yea-sayers, and the bias is known as the acquiescence
response set. 

    A less common problem is the opposite tendency for other
individuals, called nay-sayers, to disagree with statements independently of
question content. The effects of response biases should not be exaggerated, but
it is important that researchers who are using self-reports give these issues
some thought. 

    If an instrument or scale is being developed for general use by
others, evidence should be gathered to demonstrate that the scale is
sufficiently free from response biases to measure the critical variable.

Developing
Structured Self-Report Instruments

    A
well-developed interview schedule or questionnaire cannot be prepared in
minutes or even in hours. To design useful, accurate instruments, researchers
must carefully analyze the research requirements and attend to minute details.
The steps for developing structured self-report instruments follow closely
those outlined. However, a few additional considerations should
be mentioned. 

    Once data needs have been identified, related constructs should
be clustered into separate modules or areas of questioning. For example, an
interview schedule may consist of a module on demographic information, another
on health symptoms, a third on stressful life events, and a fourth on
health-promoting activities. 

    Some thought needs to be given to sequencing
modules, and questions within modules, to arrive at an order that is
psychologically meaningful and encourages candor and cooperation. The schedule
should begin with questions that are interesting, motivating, and not too
sensitive. The instrument also needs to be arranged to minimize bias. 

    The
possibility that earlier questions might influence responses to subsequent
questions should be kept in mind. Whenever both general and specific questions
about a topic are included, general questions should be placed first to avoid
“coaching.” Every instrument should be prefaced by introductory comments about
the nature and purpose of the study. In interviews, the introductory comments
would be read to respondents by the interviewer, and often incorporated into an
informed consent form. 

    In SAQs, the introduction usually takes the form of a
cover letter that accompanies the questionnaire. The introduction should be
carefully constructed because it represents the first point of contact with
potential respondents. An example of a cover letter for a mailed questionnaire
is presented. 

    When a first draft of the instrument is in
reasonably good order, it should be discussed critically with people who are
knowledgeable about questionnaire construction and with experts on the
instrument’s substantive content. The instrument should also be reviewed by
someone capable of detecting technical problems, such as spelling mistakes,
grammatical errors, and so forth. 

    When these various people have provided
feedback, a revised version of the instrument can be pretested. The pretest
should be administered to individuals who are similar to current participants.
Ordinarily, 10 to 20 pretests are sufficient.

Tips
for Developing Structured Self-Report Instruments

    Although
we all are accustomed to asking questions, the proper phrasing of questions for
a study is an arduous task. In this section, we provide some tips on wording
questions and response options for self-report instruments. Although most
advice is specific to structured self-reports, some suggestions are equally
appropriate for qualitative interviews.

Tips
for Wording

    Questions
In wording questions for self-reports, researchers should keep four important
considerations in mind. 

1.Clarity. Questions should be worded clearly and
unambiguously. This is usually easier said than done. Respondents do not
necessarily understand what information is needed and do not always have the
same mind-set as the researchers. 

2. Ability of respondents to give
information. Researchers need to consider whether respondents can be expected
to understand the question or are qualified to provide meaningful information. 

3. Bias. Questions should be worded in a manner that will minimize the risk of
response biases. 

4. Sensitive information. Researchers should strive to be
courteous, considerate, and sensitive to the needs and rights of respondents,
especially when asking questions of a private nature. Here are some specific
suggestions with regard to these four considerations:

    Tips
for Formatting an Instrument The appearance and layout of an instrument may
seem a matter of minor administrative importance. However, a poorly designed
format can have substantive consequences if respondents (or interviewers)
become confused, miss questions, or answer questions they should have omitted. 

    The format is especially important in questionnaires because respondents
typically do not have a chance to seek assistance. The following suggestions
may be helpful in laying out an instrument.