Examples of p-curves obtained for different scenarios.

Plots show histograms of the distribution of p-values for published (i.e. significant) results for 1,000 findings, in different scenarios. Top panels show results for the Two Peaks model, and bottom ones for the Single Normal model. Left panels show results for a high-bias, low-power, low-prevalence scenario, while right ones have no bias, high power and high prevalence. Specific parameters are (A) Two Peaks, bias = 0.6, power = 20%, prevalence ~ 0.1,PPV ~ 11%(B)Two Peaks, bias = 0, power = 80%, prevalence ~ 0.9, PPV ~ 12%, (C)Single Normal, bias = 0.6, power = 0.2, prevalence ~ 0.1, PPV ~ 99% and (D)Single Normal, bias = 0, power = 0.8, prevalence ~ 0.9, PPV ~ 99%. (TIF)