| name | design-an-academic-survey |
| category | school |
| description | Design an academic survey with construct definitions, sampling, valid questions, accessible delivery, privacy, pilot evidence, and an analysis plan. Use when self-reported data is appropriate to a research question. |
design-an-academic-survey
Ask only questions needed for the approved research purpose.
When to use
- Use to measure reported knowledge, attitudes, experiences, intentions, or behaviors.
- Do not collect sensitive or identifiable data merely because the survey tool makes it easy.
Procedure
- State the research question, target population, constructs, analysis, and why self-report is suitable.
- Define sampling frame, recruitment route, inclusion rules, response target, and nonresponse risks.
- Reuse validated measures when appropriate and verify permission, population fit, and scoring.
- Write one neutral idea per question with a suitable response scale and explicit reference period.
- Avoid leading, loaded, double-barreled, assumptive, vague, or unnecessarily sensitive questions.
- Design order, branching, randomization, attention checks, and optional responses without coercion.
- Provide consent, purpose, burden, risk, withdrawal, privacy, contact, and incentive information.
- Test accessibility, language, mobile use, screen readers, save behavior, and data minimization.
- Conduct cognitive interviews and a pilot, then revise before approved launch.
- Predefine cleaning, missingness, weighting, exclusions, uncertainty, and open-text handling.
Worked example
A campus survey asks whether students support an effective new policy, combining awareness and approval. Cognitive testing exposes the bias and double question. The revised instrument measures awareness, experience, and approval separately and permits respondents to skip a sensitive item.
Done
- A survey package records constructs, sample, instrument, consent, branching, accessibility, privacy, and analysis rules
- Cognitive-test and pilot reports verify interpretation, completion behavior, technical delivery, and approved revisions
