A poorly designed questionnaire can quietly undermine an entire study — collecting data that looks complete but doesn't actually measure what the researcher intended. Here's how to avoid the most common design mistakes.
A common mistake is writing questionnaire items first and figuring out what they measure later. The more reliable approach works backward: start from your research question and the specific constructs you need to measure, then design items deliberately to capture each construct. If you can't clearly state which construct a given item is meant to measure, it probably shouldn't be in the questionnaire.
A question like "Don't you agree that modern teaching methods are more effective?" nudges respondents toward a particular answer. Compare this to the more neutral "How effective do you find modern teaching methods?" — the second version doesn't presuppose an answer.
A question like "Is your teacher knowledgeable and approachable?" actually asks two separate things at once. A respondent who finds their teacher knowledgeable but not approachable has no way to answer accurately. Split these into two separate items.
Words like "often," "regularly," or "adequate" mean different things to different respondents. Where possible, replace vague terms with specific, quantifiable options ("2-3 times per week" instead of "often").
Academic or technical terminology familiar to the researcher may not be familiar to respondents. Write items at a reading level appropriate to your actual target population, not the level you'd use writing for your thesis committee.
Early questions can influence how respondents interpret and answer later ones. General practice is to place simpler, less sensitive questions first, building rapport before moving to more sensitive or complex items later in the questionnaire.
Switching between different response formats (a 5-point scale here, a 7-point scale there, yes/no elsewhere) without clear reason increases respondent confusion and cognitive burden. Keep formats consistent within sections where possible.
Forcing a response on a question that genuinely doesn't apply to a respondent produces inaccurate data. Where relevant, include "not applicable" or skip logic so respondents aren't forced into a misleading answer.
Administering your questionnaire to a small group before the actual study — even just 10-15 people similar to your target population — reliably surfaces wording problems, confusing items, and timing issues that are much harder to spot when you're the one who wrote the questionnaire. This single step catches more design problems than any amount of additional careful re-reading by the researcher alone.
| Check | Why it matters |
|---|---|
| Every item maps to a specific construct in your research question | Prevents collecting irrelevant or unusable data |
| No leading, double-barreled, or vague items | Improves validity of individual item responses |
| Consistent response formats within sections | Reduces respondent confusion and fatigue |
| Pilot tested with a small sample first | Catches wording and timing problems before full deployment |
Related: Thesis Structure 101 · Scholar's Corner overview