STATISTICAL PRIMER
State Center For Health Statistics
Department of Human Resources • Division of Health Services • P.O. Box 2091 • Raleigh, NC 27602-2091 • 919/733^728
No. 7 o j October 1986
N.C
QUESTIONNAIRE DESIGN
by
J. Michael Bowling
ABSTRACT
An overview of the survey process is presented with special emphasis on questionnaire design. Topics covered are types ofquestions,
structure of questions, question wording, and questionnaire construction. This information should help persons in state and local
agencies to collect better data through more effective questionnaire design and implementation.
The collection, processing, and dissemination of
information have become increasingly important to the
health care provider, researcher, and administrator faced
with competing alternatives for the expenditure of scarce
resources. Sample survey design, once relegated by many
students to the dusty shelf of pedagogic esoterica, has
become a useful tool to a more diverse group of data users
employing more sophisticated computer technologies.
The intent of this Statistical Primer is to provide this newly
emerging audience of data users with an overview of the
"survey process" placing special emphasis upon ques-tionnaire
design.
Conducting A Survey
A survey is a systematic means of collecting informa-tion
directly from people (respondents) often through the
use of a questionnaire. A census is a special type of survey
in which efforts are made to gather information from all
members of the population. In contrast, a sample survey
normally involves the selection of a small percentage of
eligible members of a population representative of the
larger group. If appropriate steps have been taken to
eliminate bias in the selection of sample respondents, the
sample size and variability of the sampling units can be
used to determine the precision of estimated means,
proportions, and other population indices being measured
in the survey.
There are three basic types of surveys: face-to-face,
telephone, and mail. The choice of method is dependent
upon, among many other things, the research goal of the
survey, the population chosen for study, and the survey
budget.
Warwick and Lininger ( 1 ) have delineated seven inter-dependent
stages in conducting a sample survey, most of
which also apply in the case of a census of all eligible
respondents. The following is an overview of these stages.
Planning is the preliminary stage of the survey. During
the planning stage researchers, should conduct an exten-sive
literature review of previous publications to be used
as a basis for questionnaire design. The formulation of a
"research question" as a focus for the project is often
helpful in ensuring continuity between the stages of the
survey which culminate in data analysis and reporf
writing.
Sampling is the process of selecting a subset of eligible
respondents from a population. The key to the cost-effectiveness
of choosing a small number of respondents
and generalizing to a much larger whole is the representa-tiveness
of the sample. Consultation with a survey statisti-cian,
while beneficial to all stages of survey design, is most
crucial in the sampling stage. The foundation of a good
sample survey rests upon the sample ( 1 ).
Questionnaire design is the stage of the survey pro-cess
in which the goal of the project, often embodied in
the research question, is translated into a series of ques-tions.
The principles of designing a survey questionnaire
are the same whether the scope of a survey project is
limited to an interoffice canvas or whether it is a multis-tage
national survey. A discussion of these principles
is presented in most of the remainder of this report.