statistics-check

name	statistics-check
description	Validate statistical methods, check for p-hacking signs, and verify reporting adequacy in biomedical papers.

Statistics Check

Audit statistical methods and reporting in biomedical studies.

Checklist

Study Design

• Sample size justified with power calculation
• Randomization method described
• Blinding status reported
• Primary endpoint pre-specified

Statistical Methods

• Appropriate test for data type (parametric vs non-parametric)
• Multiple testing correction applied (FDR/Bonferroni)
• Effect sizes reported with confidence intervals
• Missing data handling described

Red Flags

• P-value rounding (p < 0.05 without exact value)
• Selective endpoint reporting
• Subgroup analyses without pre-specification
• Baseline table p-values (in RCTs)
• Figure legends without sample sizes
• Error bars undefined (SD vs SEM vs CI)

Reproducibility

• Raw data available
• Analysis code available
• Statistical software and version reported

Output

Save audit to outputs/stats-check-<slug>.md with findings and severity ratings.