P-value Interpretation Assistant
π How to use this tool
- Step 1: Select your outcome type (mean, risk, odds, etc.) and fill in sample size, pre-specification, and MCID (minimum clinically important difference).
- Step 2: Enter the reported effect size, 95% confidence interval, and p-value.
- Step 3: Optionally, add baseline risk, enable Bonferroni or Bayes factor adjustments.
- Step 4: Click Interpret / Infer. The assistant provides: strength of evidence, clinical relevance vs MCID, CI precision/compatibility, multiplicity context, and optional NNT/NNH & Bayes factor.
- Step 5: Read the colored flags for quick interpretation hints and copy the summary sentence for your report or manuscript.
Apply Bonferroni for 3 tests: Adjusts p-values to control for multiple comparisons when testing several outcomes. For example, if three outcomes were analyzed, each p-value is multiplied by 3 to maintain an overall 5% error rate.
Show conservative Bayes factor from p: Converts the p-value into an approximate Bayes factor (Sellke et al., 2001) to express how much more likely the data are under the null hypothesis versus the alternative. Smaller Bayes factors indicate stronger evidence against the null.
1) Study context
Use absolute units (e.g., 1 day; 8 percentage points).
Quick Examples
2) Effect & uncertainty
If you provide CI, weβll also comment on precision and compatibility.
3) Optional (binary outcomes)
Interpretation
Copy-ready summary
β
π§© Solved Example (click to auto-fill and infer)
Scenario: Pain score difference (0β10 scale) favors treatment by β1.2 points (95% CI β1.9 to β0.5), p=0.001, MCID = 1.0, pre-specified primary outcome; 3 total outcomes tested.
