questionnaire-design

Start by defining the decision the data will inform, then draft questions using closed formats like Likert scales, ranking, and forced choice. Order questions from factual to opinion-based, pilot test with 3-5 people, and keep the total under 15-20 questions.

Use interviews first to discover requirements through open-ended exploration, then use questionnaires to validate and prioritize findings with larger groups. Questionnaires confirm what you already know; they do not discover new requirements.

Likert scales measure satisfaction and agreement, ranking questions reveal feature priorities, and forced-choice questions expose real tradeoffs. Limit open-ended questions to 1-3 per questionnaire since they are hard to analyze at scale.

Keep questionnaires to 15-20 questions with completion time under 10 minutes. Completion rates drop roughly 20% for every 5 questions beyond 15, and every question should map to a specific decision it informs.

Bias comes from leading questions, double-barreled questions, unbalanced scales, and non-response patterns. Use balanced positive and negative scale options, one concept per question, neutral wording, and track response rates by stakeholder group.

Avoid questionnaires as the first elicitation activity or when you need deep exploratory discovery. They produce shallow, predefined responses without follow-up capability, so use interviews or observation when requirements are still unknown.