Definition: P-curve is a tool for identifying potential publication bias and makes use of the distribution of significant p-values in a series of independent findings. The deviation from the expected right-skewed distribution can be used to assess the existence and degree of publication bias: if the curve is right-skewed, there are more low, highly significant p-values, reflecting an underlying true effect. If the curve is left-skewed, there are many barely significant results just under the 0.05-threshold. This suggests that the studies lack evidential value and may be underpinned by questionable research practices (QRPs; e.g., p-hacking). In the case of no true effect present (true null hypothesis) and unbiased p-value reporting, the p-curve should be a flat, horizontal line, representing the typical distribution of p-values.

Related terms: File-drawer, Hypothesis, P-hacking, p-value, Publication bias (File Drawer Problem), Questionable Research Practices or Questionable Reporting Practices (QRPs), Selective reporting, Z-curve

References: Bruns and Ioannidis (2016), Simonsohn et al. (2014a), Simonsohn et al.(2014b), & Simonsohn et al. (2019)

Drafted and Reviewed by: Bettina M. J. Kern, Sam Guay, Kamil Izydorczak, Charlotte R. Pennington, Robert M. Ross, Olmo van den Akker

Note that we are currently working on an automated mechanism to link references cited above with their full-length version that can be found at https://forrt.org/glossary/references with all references used so far.