Questionnaire Design

1
Questionnaire Design
2
Six Steps of Questionnaire Design

• Decide what information should be sought
• Decide what type of questionnaire should be used
• Write a first draft
• Re-examine and revise
• Pretest the questionnaire
• Edit the questionnaire and specify procedures for
  its use.

3
Step 1

• Articulate the problem that is under study e.g.
  It has been suggested that GPs dont spend enough
  time on preventive care activities.
• Determine what findings you need in order to
  answer the research question.
• Ascertain whether single questions or multiple
  questions are required for each finding.

4
Step 2

• Determine type of questionnaire will be used.
• e.g. mail-out, interviewer administered, phone
  survey, individual interviews.
• Identify whether relevant instruments already
  exist.

5
Step 3

• Draft questionnaire.
• Consider wording, ordering and formatting of
  questionnaire.

6
Step 4

• Re-examine and re-write questionnaire.
• Have questionnaire reviewed by experts.
• If dealing with a controversial topic, have
  questionnaire reviewed by both sides of the
  issue.
• If no consensus can use split ballot technique
  (different wording for same questions 2
  versions of questionnaire).

7
Step 5

• Pre-test the questionnaire.
• Administer questionnaire to small sample under
  survey conditions.
• Ensure pre-test respondents resemble actual
  survey respondents.
• Interview respondents about reactions to items
  and questionnaire as a whole (identify ambiguous
  or offensive items).
• A second pre-test may be necessary if major
  changes required.

8
Step 6

• Edit questionnaire.
• Specify administration procedures.

9
Preparing Questionnaire Items

• Closed questions
• Provide specific response alternatives.
• Open-ended questions
• Merely specify the area to be addressed in a
  response.

10
(No Transcript)
11
Closed Questions

• Advantages
• Provide respondents with a standard frame of
  reference.
• No coding problems.
• Reduces chance of interviewer bias.
• Reduces data entry time.

• Disadvantages
• More difficult to write.
• May lead to fence sitting when middle alternative
  is offered.
• Limited response categories.
• Possible bias due to limited response options.

12
Open-ended Questions

• Advantages
• Quick easy to write.
• Wide range of responses.
• Able to elicit all possible answers.
• Useful when many pre-codes are required.
• Rich information.
• Quotable quotes.

• Disadvantages
• Intensive coding needed.
• Post coding can introduce bias.
• May restrict range of responses if question does
  not adequately cue respondents.
• Costly time consuming.
• Limits statistical inference
• More difficult to analyse.

13
Good Questionnaire Items

• Simple, direct and familiar language.
• Clear and specific.
• Short as possible (max. 20 words).
• No leading, loaded or double barrel questions.
• Only one question per statement.
• Avoid universals e.g. only, just, merely.
• Include all conditional information prior to key
  idea.
• Edit for readability.

14
Task

• Each group to come up with three closed questions
  and two open-ended questions for their topic.
• Do not use basic demographic questions
• Apply principles of good question construction.

15
Measurement Scalesfor categorical items
16
Nominal Scale

• Categorise responses into a number of distinct
  categories.
• There is no implication of grading responses.
• Example
• What is your sex? Male 1 Female 2

17
Ordinal Scale

• Ordinal scales attempt to order or rank
  responses.
• Add the relationship of greater than or less
  than to the measurement process.
• Can know the rank order but not how far apart
  respondents are.
• e.g. What is the highest level of education you
  have completed?
• Primary school 1
• Some high school 2
• High school graduate 3
• University graduate 4

18
Interval Scale

• Can specify how far apart two stimuli are on a
  give dimension.
• Differences of the same numerical size in scale
  value are equal.
• Does not have a meaningful zero point.
• Standard mathematical operations (add, subtract,
  division, multiplication) can be performed on
  interval data.

19
Interval scale Example

• Difference between 50 and 70 correct answers on
  an aptitude test is equal to difference between
  70 and 90 correct answers. Thus the ratio of
  scale intervals is meaningful.
• However, there is no meaningful zero point.
• For example, a score of zero on a verbal aptitude
  test would not necessarily mean that the person
  had absolutely no verbal ability.

20
Ratio Scale

• A ratio scale has the same properties as an
  interval scale with the addition of an absolute
  zero point.
• Makes the ratio of scale values meaningful.
• e.g.Physical scales measuring time, weight and
  distance can be treated as ratio scales.

21
Measurement Scales for Non-Categorical Items
22
Likert Scales

• Commonly used in attitudinal research
• Example
• For each of the following statements please
  circle one of the numbers below according to your
  level of agreement with the statement.
• The GP should be the first point of contact for
  patients seeking to loose weight?
• Strongly agree Agree Strongly
  disagree
• 1 2 3 4 5

23
(No Transcript)