Design a survey to answer {{question}}. Help me avoid the standard survey pitfalls.
Context:
- Who I'm surveying: {{audience}}
- How I'll use the results: {{decision the data will inform}}
- Length tolerance: {{how long they'll actually fill out — 2 min? 10 min?}}
Output:
1. **The 5-8 questions** I should actually ask — fewer is better, only what informs the decision
2. **For each question** — the format (multiple choice / scale / open text), and why that format vs alternatives
3. **Leading questions to avoid** — common phrasings that bias the answers
4. **The "catch" question** — one that catches careless or fake responses
5. **Open text strategy** — when to use, what to ask, how to make analysis tractable
6. **What the response distribution will look like** if my hypothesis is right vs wrong — set expectations
Skip "how would you rate us 1-10" and other surveys-by-rote. Each question must earn its place.surveysresearchfeedback