- Create a construct/item map, pilot record and final questionnaire version.
Define constructs and use
Specify the construct or behavior, timeframe and intended analysis. Existing instruments need appropriate permissions and relevant measurement evidence. A longer questionnaire is not necessarily a better one.
Avoid ambiguous wording
Ask one interpretable question at a time. Combining instructor quality and facilities in one item prevents separate interpretation. Avoid leading terms, unsupported assumptions and unfamiliar jargon. Match answer options to the concept and respondent recall.
Review order and administration
Question order, mode and option presentation can influence responses. Randomization is not suitable for every naturally ordered scale. Test mobile display, completion time, branching and withdrawal.
Use cognitive testing and a pilot
Ask potential respondents how they interpret items and choose answers. Complement this with a small workflow pilot for missingness, timing and interface errors. Assess semantic and cultural equivalence in multilingual instruments; back-translation alone is not sufficient.
Preserve the actual questionnaire
Keep wording, options, order, version date, instructions, branching and pilot changes. Define response-rate denominators and do not describe a convenience sample as representative of all students. Deliver a construct-to-item-to-analysis mapping.
Practical research checklist
- Give each item an explicit purpose.
- Clarify timeframes and options.
- Document pilots and translation review.
- Preserve the administered version.
Worked case and implementation decisions
The following is a fictional teaching case. Do not use its numbers or wording as actual study findings.
“Are you satisfied with the instructor and facilities?” combines two concepts; a respondent may be satisfied with one and not the other. Split it into suitable items with explicit timeframes and options. For a behaviour, “How many times in the last four weeks…” can be clearer than “usually,” if recall is feasible. Cognitive interviews examine interpretation, while a pilot checks flow, timing and form errors. Neither alone establishes construct validity. Preserve the version actually administered.
| Stage | Teaching example | Verification question |
|---|---|---|
| Construct | Teaching satisfaction separate from facilities | One concept per item? |
| Timeframe | Four weeks in the behaviour example | Compatible with recall? |
| Options | Range and non-applicability where justified | Do real answers fit? |
| Pilot | Meaning, branching and mobile display | Are changes logged? |
| Version | Actual wording and order | Does analysis use this version? |
Exercise output: Create a construct/item map, pilot record and final questionnaire version.
Sources and further reading
Official sources for verification and further reading