Page 136 -

122 CHAPTER 5 Surveys

5.7 OVERALL SURVEY STRUCTURE

Well-written questions are important, but so is the overall structure of the survey instru-
ment. The questions do not exist in a vacuum, rather, they are part of an overall survey
structure. For instance, a survey, in any format, must begin with instructions. These
instructions must make clear how the respondent is to interact with the survey (Babbie,
1990). For instance, are respondents required to fill out all of the questions? Will respon-
dents be required to enter their name or contact information? It is sometimes useful to
put in a description, as a reminder, of who should be filling out the survey (e.g., you must
be aged 65 years or older). If a survey is separated into multiple sections, then those
divisions, and who should fill those different portions, must be made clear. Each section
should be given an appropriate heading. Just as it is important to provide navigation on
a website, a survey should provide navigation to the reader, whether the survey is paper,
e-mail, or web-based. The user (respondent) needs to know where on the survey they
should go, in what order. Sometimes, it is also helpful to provide contact information if
the respondent has any questions (such as a telephone number or e-mail address).
Different formats of surveys (paper, e-mail, web-based) may require that infor-
mation or instructions be presented to the respondent. For instance, in a paper survey,
are there ovals or checkboxes? Should a checkmark be placed in them, should an X
be placed in the box, or should the box be filled in? Should items be circled? Are
respondents required to fill out all of the questions? These directions must be made
clear. For an e-mail survey, should answers be typed in directly following the ques-
tion on the same line, or on a line or two below it?
Layout of the survey instrument can also be important. For paper surveys, it
is important to make sure that there is enough white space so that the respondent
does not feel overwhelmed by the amount of information on a page (Babbie, 1990).
Obviously, a balance needs to be struck. While respondents may worry if they see a
30-page survey, on the other hand, stuffing all of the survey questions onto two pages
may prove to be problematic. Only white paper should be used, and a large enough
font, in standard text, should be used (Dillman, 2000). Booklet printing (with two
staples in the middle of the booklet) is preferred to one staple in the upper left-hand
corner, but that is still preferred to any type of unusual folding or paper shapes that
users may have trouble understanding (Dillman, 2000). In addition, do not use ab-
breviations to cut down on the amount of space needed, as they may cause confusion
among respondents (Babbie, 1990). For a web-based survey, links are often provided
directly in the survey, so that the respondent can click on the link and get a pop-up
window with more detailed information. While pop-up windows are generally not
good interface design, they work very well for giving short bits of information to
users while they are in the process of responding to a survey.
Survey questions generally may be asked in any order which makes sense in the
context of the research. However, it is important to keep in mind that questions relat-
ing to a similar topic or idea should be grouped together (Dillman, 2000). This tends
to lower the cognitive load on respondents and allows them to think more deeply
about the topic, rather than “switching gear” after every question. Because some