Evidence for p-hacking across scientific disciplines.

<p>A) Evidence for p-hacking from <i>p</i>-values obtained from Results sections. B) Evidence for p-hacking from <i>p</i>-values obtained from Abstracts. The strength of p-hacking is presented as the proportion of p-values in the upper bin (0.045 < p < 0.05) with one-tailed 95% confidence intervals (calculated following Clopper and Pearson [<a href="http://www.plosbiology.org/article/info:doi/10.1371/journal.pbio.1002106#pbio.1002106.ref047" target="_blank">47</a>] using the <i>binom.test</i> function in R). Only disciplines where text-mining of the Results sections returned more than 25 <i>p</i>-values between 0.04 and 0.05 are presented. Marker colour is shaded according to the sample size: with white indicating low samples sizes and red indicating larger sample sizes.