Interpret Survey Results Honestly
Get a careful read of survey data — what it suggests, what it can't tell you, and the most likely ways it could mislead.
When to use this
When you've run a survey and need an interpretation that doesn't over-claim from a small sample.
The prompt
You are a careful analyst who treats survey data with the right amount of skepticism.
Inputs:
- **What I was trying to learn**: [the question that motivated the survey]
- **Who answered**: [audience, size, how they were recruited]
- **The data**: [paste responses, summary tables, or counts per question]
- **What I'm tempted to conclude from it**: [your draft interpretation]
Do this:
1. **What the data actually shows** — restate the responses faithfully. Numbers and percentages. No editorializing yet.
2. **What it suggests, with confidence levels** — for each main finding: claim, evidence, your confidence (Low/Medium/High), and what would raise that confidence.
3. **What it CANNOT tell us** — limits of the sample, biases in recruiting, leading questions, the difference between what people say and what they do.
4. **Pressure-test my draft interpretation** — what am I likely over-claiming? Under-claiming?
5. **What I'd ask next** — 2–3 follow-up questions or experiments to validate the most important finding.
Be willing to tell me my draft is too strong. Don't pad with reassurance.
What you'll get back
A faithful summary of the data, claim-by-claim confidence levels, named limits, a critique of the draft interpretation, and 2–3 follow-up questions.
How this is structured in English
Notice the English patterns this prompt uses — they're worth borrowing for your own requests.
- The right amount of skepticism 'The right amount' is the operative phrase — not zero, not paralyzing. Calibrated doubt is the analyst's skill.