The prompt
Score 10 of 10, Excellent
Use it when You ran an analysis, or were sent one, and need to know what the coefficients, p-values and R-squared actually say before you report them.
- {{question}}
- {{tool}}
- {{data_description}}
- {{audience}}
- {{output}}
Full prompt and why it works
Prompt
The question the analysis was meant to answer: {{question}} Tool that produced it: {{tool}} The data (what each row is, how many rows, the period): {{data_description}} Who I need to explain it to: {{audience}} The output, pasted exactly: """ {{output}} """ Task: Explain the output above in plain words, so that I can tell {{audience}} what it shows and how sure we can be, without overstating it. I am not a statistician. Format: 1. The answer: two or three sentences that answer my question from the output, with the key numbers. 2. Line by line: for each number that matters (for example coefficients, standard errors, p-values, confidence intervals, R-squared and sample size), what it is in one plain sentence and what it says here, with the value copied exactly. Say which parts of the output you skipped. 3. What it does not show: whether the design supports cause and effect or only association, what the model leaves out, and whether the sample is large enough for the claims. 4. Before reporting: assumptions the output cannot confirm on its own, such as outliers, a straight-line relationship or missing data, each with a simple check I can do. 5. A summary for {{audience}}: three bullet points in plain language, with the right level of caution. Requirements: - Under 500 words. - Copy every number exactly as it appears; round only in the summary, and say that you rounded. - A small p-value does not prove a cause, and a large one does not prove there is no effect. Describe what the result suggests and how precise it is, using the confidence interval where there is one. - Do not invent numbers the output does not contain, such as a confidence interval it does not report. Example of one line-by-line entry's form only, with invented numbers: "Price coefficient -2.4: each $1 increase in price goes with about 2.4 fewer units sold per week, holding the other variables in the model constant." If a label in the output is unclear, or I have not said what a variable is or how it was coded, say what you are assuming and what I should confirm.
Variables
- {{question}}
- For example "Does training hours predict first-year sales, once region is taken into account?"
- {{tool}}
- For example "Excel Analysis ToolPak", "SPSS", "R" or "a colleague's report".
- {{data_description}}
- For example "one row per sales rep, 146 reps, 2024".
- {{audience}}
- For example "the regional managers" or "my manager".
- {{output}}
- The whole output table or summary, copied exactly, with its labels.
Why it works
- What to avoid. It separates association from cause and explains what a p-value does and does not prove, the two places statistical results are most often overstated.
- What good looks like. Every number is copied exactly and rounding happens only in the summary, so the explanation can be checked against the output line by line.
- If unsure. Unclear labels and variable coding come back as stated assumptions to confirm, not silent guesses.