P-Value Precision and Reproducibility

摘要: P-values are useful statistical measures of evidence against a null hypothesis. In contrast to other estimates, however, their sample-to-sample variability is usually not considered or estimated, and therefore fully appreciated. Via systematic study log-scale p-value standard errors, bootstrap prediction bounds, reproducibility probabilities for future replicate p-values, we show that p-values exhibit surprisingly large in typical data situations. addition providing context discussions about the failure results replicate, our findings shed light on relative value exact vis-a-vis approximate indicate use *, **, *** denote levels 0.05, 0.01, 0.001 significance subject-matter journals right level precision reporting when judged by widely accepted rules rounding estimates.