Overview
Sample size calculations are almost universally done wrong in two directions: they use the expected effect size (which is optimistic) rather than the minimum meaningful effect size (which is the correct input), and they treat the result as a single number rather than a function of assumptions that should be stress-tested.
The consequence of an underpowered study is not just a failed experiment — it is a biased literature. Underpowered studies that achieve significance are more likely to be false positives (the winner's curse). Underpowered studies that don't achieve significance are filed away as null results, even though they were never capable of detecting the effect they were looking for.
The Sample Size & Power Analysis Prompt generates a complete power analysis framework: effect size specification from the minimum meaningful effect, sample size calculation for the specific test type, sensitivity analysis across effect size and power assumptions, and a sequential testing specification for studies where early stopping is needed.
What you get: - Minimum meaningful effect size specification - Sample size calculation for 6 test types - Sensitivity analysis across assumptions - Sequential testing specification for early stopping - Post-hoc power analysis for completed studies
Built for: researchers, analysts, and data scientists designing experiments, surveys, and studies who need defensible sample size justifications.