W4 - Survey research

Question 1

What are the 2 purposes of surveys?

Answer 1

1. Information gathering
   - Exploratory
   - Descriptive
2. Theory testing & building
   - Explanatory
   - Predictive

Question 2

What are the processes to develop questionnaires?

Answer 2

1. When to design:
   - No established tool exists to measure
   - Reliability/validity of established measure is in doubt
   - Avoid jangles: different labels for same thing
2. Piloting - test on a small group of people
   - Identify problematic items and revise
   - Ask respondents for feedback
3. General design principles: short, readable and appropriate response options

Question 3

What needs to be considered when designing surveys?

Answer 3

1. Instructions: clear and easy-to-follow
2. Order: can divide into sections (by topic or question types), easy/general -> specific
3. Demographic: characteristics of sample & include relevant questions
4. Types of questions: open (rich data, need analysis strategy) & close (unambiguous, clear response)
5. Clearly written

Question 4

What are the general rule to avoid to write good survey questions

Answer 4

• Double-barrelled questions - unclear
• Ambiguity -> multiple interpretations
• Negations (OR double-negative) -> positive tone instead
• Emotive language -> leading Ps
• Jargon
• Response bias (social desirability effects)
• Response acquiescence (mixed response)

Question 5

What are the different formats of rating scale?

Answer 5

1. Dichotomous (Yes/No)/Multichotomous (multiple options to choose)
2. Likert scale (multi-point responses, ensure spacing of response options)
3. Non-verbal scale (suitable for children, replace texts with images)
4. Graphic rating scale (Ps can draw their own answer as a mark on the continuous line)
5. Ranking scale (measure relative importance)
6. Semantic differential scale (where opinion lies on a spectrum)

Question 6

What are the ways to increase response rate?

Answer 6

• Keep questionnaire short
• Keep questionnaire simple & clear
• No extra cost to Ps
• Sending reminder
• Offer an incentive (compensation)

Question 7

How to assess quality of questionnaire?

Answer 7

Psychometric: science of measuring psychological constructs (e.g. personality, cognitive ability test) -> questionable validity sometimes

Criteria:
1. Temporal consistency: same results under same conditions?
2. Internal consistency: does it actually measure construct?
3. Construct validity - assessed by:
- Convergent: correlates with tests of related constructs
- Discriminant: doesn’t correlate with tests of different constructs