Experiment design before the build
researchWrite the hypothesis, the metric, and the analysis before the test exists.
Survey analysis plan before the data
Write the questions and cuts before you look at responses, so the analysis is not fishing.
When to use it
When a survey is still open, or when you have data but no analysis plan.
The prompt
Copy into your assistant
You are writing an analysis plan, not the analysis.
Input: the survey questions and the decision the survey is meant to inform.
Rules:
- State the decision first, then the cuts that would change it.
- Name the comparison groups and why they matter.
- Say what result would be surprising, and what result would be inconclusive.
- Do not propose significance tests that the sample size cannot support.
Output: Decision, primary cuts, secondary cuts, what would change our mind, limitations.
If the sample size is unknown, say the plan is provisional and name what sample would be needed.What to change
Paste the raw question wording — wording effects are a larger risk than the analysis choices at typical survey sizes.
Why it works
Analysis plans written before the data prevent the most common survey failure: picking the cut that produced an interesting number.
More research prompts
- Experiment design before the build Write the hypothesis, the metric, and the analysis before the test exists.
- Source quality triage Rank a source list by authority and recency before writing from it.
- Competitor teardown from public pages Structure what a competitor says about itself, with quotes, and without speculation.