Docstoc

DESIGN OF SURVEY QUESTIONNAIRES

Document Sample
DESIGN OF SURVEY QUESTIONNAIRES Powered By Docstoc
					    Design of Survey
     Questionnaires
    Andrew L. Dannenberg, MD, MPH
 National Center for Environmental Health
Centers for Disease Control and Prevention

              CSTE Fellows
1. Initial Steps in Survey Design

• Decide if questionnaire is best method to
  obtain information (vs. other sources)
• Identify scope of information needed –
  shorter is better
• Outline basic analyses expected to be
  done
• Identify sampling frame and techniques
    2. Select Type of Survey

•   Self-administered - paper
•   Self-administered - computer
•   E-mail or website survey
•   Telephone interview
•   Personal interview
          Comparison of Questionnaire Types
Character-    Self-      Self-         Email or   Telephone   Personal
istics        adminis-   adminis-      website    interview   interview
              tered -    tered -       survey
              paper      computer
Response      40% -      40% - 75%     Variable   70% - 90%   75% -
rate          75%                                             95%
Sample bias   Medium/    Medium/high Medium/      Medium      Low
              high                   high
Cost          Low        Low           Low        Low/        High
                                                  medium
Rapport       Fair       Fair          Poor       Good        Excellent
Control of    Poor       Poor/fair     Poor       Fair        Good
response
situation
Quality of    Poor       Good          Good       Good        Excellent
recorded                 (computer
responses                helps edit)
 Comparison of Questionnaire Types (cont.)
Character-    Self-      Self-        Email or    Telephone    Personal
istics        adminis-   adminis-     website     interview    interview
              tered -    tered -      survey
              paper      computer
Length of     Short/     Medium       Medium/     Medium/      Long
question-     medium                  long        long
naire
Complexity Must be       May be       May be      May be       May be
of question- simple      complex      complex     complex      complex
naire                    using skip
                         patterns
Use of        Good       Good         Good        Not possible Excellent
visual aids
Control of    Poor       Excellent    Excellent   Excellent    Excellent
question                 with skip
order                    patterns
Sensitive     Good       Good         Good        Fair/good    Fair
topics
3. Structure of Questionnaire:
      Opening Sentences
•   Identify self and organization
•   Explain reason without biasing result
•   Include estimated time to complete survey
•   Give permission to refuse specific questions
•   Provide assurance of confidentiality
•   Ask for permission to continue with survey
•   Establish eligibility for participation
3. Structure of Questionnaire:
         Main Content
•   Begin with easy non-intrusive questions
•   Ask most important questions early in survey
•   Use logical order and transition between topics
•   Put sensitive questions at end (such as income)
•   Be culturally sensitive
•   Use pre-coded responses to help data entry
3. Structure of Questionnaire:
      Closing Sentences
• Thank for participation
• Offer how and when results will be
  available
     4. Characteristics of Good
             Questions
•   Use standard questions from other surveys
•   Use clear, simple, unambiguous wording
•   Define terms
•   Assume 8th grade reading level for adults
•   Consider cultural appropriateness of question
•   Open ended questions to gather new ideas
     5. Characteristics of Good
             Answers
•   Clear, mutually exclusive, exhaustive
•   Offer all likely categories of answers
•   Be consistent within groups of questions
•   Offer “Don’t know” and “No opinion” options
•   Use skip patterns when appropriate
•   Rating scales are useful (low to high, 1 to 5)
6. Questionnaire Development
A. Reliability
• Test-retest repeatability:
   – Same subject, different times
• Inter-observer variability:
   – Same subject, two interviewers
• Internal consistency:
   – Related questions yield similar answers

B. Validity
• Need external standard
• Some questions may be validated elsewhere
    7. Survey Pilot Testing
• Develop interviewer instructions
• Do pilot test in similar group
• Identify problems and ambiguities in content and
  format
• Identify questions that are not useful
• Determine best cut points for continuous
  variables
• Determine time needed to complete survey
• Obtain approval for survey from Institutional
  Review Board

				
DOCUMENT INFO
Shared By:
Categories:
Tags:
Stats:
views:26
posted:12/1/2012
language:English
pages:12