Week 7 - Survey Research (Core Notes)

Question 1

What are the 7 steps of survey research?

Answer 1

1. Plan and design survey research study:
   - what issues do you want to survey?, cross-sectional/longitudinal design?, target population?
2. Select sample (s)
3. Build and refine survey instrument
4. Collect survey data
5. Enter and ‘clean’ data (locate and eliminate errors where possible)
6. Analyse survey data
7. Interpret and report results.

Question 2

In survey research, what are the range of data collection methods available?

What reasons would be considered in the election of such methods?

Answer 2

Interviews, questionnaires, mail questionnaires, group -administered questionnaires, electronic surveys.

Depends on extraneous variables trying to account for + practical + ethical considerations of your research.

E.g in person interviews = clear up ambiguities, have high completion rate - but expensive, and sometimes may be personal (uncomfy)
E.g web based surveys are inexpensive + large sample size - though privacy and anonymity problems + sample may not be representative of population (volunteer sampling).

Question 3

Principles 1 and 2 (constructing survey instruments)

Answer 3

Principle 1: Write items to match the research objectives
- i.e cover all the areas you need to cover through extensive review of all literature.
Principle 2: Write appropriate items for respondents
- Not using pretentious language
- Consider reading leaves/demographic/culture of participants
- Must be understandable and meaningful to participants
- Natural/familiar language.

Question 4

Principles 3, 4, 5 (constructing survey instruments)

Answer 4

3. Write short simple questions
   - concise, avoid jargon, unambiguous, easy, simple to answer.
4. Avoid loaded or leading questions
   - Provoke emotionally charged reaction
   - Implicitly suggest how participant should respond e.g context on feminism then are you a feminist? (Usually can be answered with yes/no response).
5. Avoid double barrelled questions
   - two or more issues on single question - makes them more complex to answer.

Question 5

Principles 6 and 7 (constructing survey instruments)

Answer 5

6. Avoid double negatives
   - sentence construction that contain two negatives, can be misread - erroneous answer.
7. Decide on close ended and or open ended questions
   - Open ended questions good when researcher unsure what respondent is thinking (qualitative/exploratory research)
   - Close ended questions, good for quantitative research - easier to standardise data

Researchers may wish to include both in one type of survey.

Question 6

Principle 8 and 9 (the principles of constructing survey instruments)

Answer 6

8. Use mutually exclusive and exhaustive response categories
   - categories should not overlap - this can skew results.
   - Response categories are exhaustive when categories include all possible responses.
9. Choose a type of response category
   - For close ended questions - such response categories include;

• Binary forced choice - participants choose one of a pair of attitudinal objects
• Rankings - participants are asked to put their responses in ascending or descending order
• Checklists - participants asked to check all response categories that apply.

Note** Binary forced-choice options can reduce response set, difficult for item analysis. Not recommended

Question 7

Principle 10 (constructing survey instruments)

Answer 7

10. Use multiple items to measure complex constructs

• gender, weight or age can be easy to measure
• Complex/abstract constructs (self esteem/intelligence) can be harder > thus, multiple questionnaire items needed to measure these constructs.

Question 8

Principle 11 (constructing survey instruments)

Answer 8

11. Make sure the questionnaire is easy to use
    - Ordering of questions: Positive first, then negative last - captures attention - thus put demographic questions last.
    - Questionnaire length: Should be short (mail questionnaires), telephone interviews less than 15 min, face to face could be longer.
    - Contingency questions: directing participants to different follow up questions can be confusing (like different categories)
    - Response bias: Social desirability bias (make yourself look good) - minimise this by ensuring anonymity and if using binary forced questions, frame each choice with equal desirability.
    - Response set: May not want to pick extremes and thus choose middle choices/always agree and stuff - solution: use an even number of response categories on rating scale. Multiple choice question types helps reduce response set but also reduces reliability.

Question 9

Principle 12 - constructing survey instruments

Answer 9

12. Pilot test questionnaire until it is perfected
    - to identify and fix problems, to practise the protocols, and clear up ambiguity before your study (just like in experimental research.
    - Think aloud technique: Participants their thoughts while engaged in activity i.e questionnaire