Figure 13.
Two scenarios for p-hacking ‘workflows’ and their effect on false-positive rates. (a) p-hacking workflow in a t-test. (b) p-hacking workflow in a regression analysis. p-hacking strategies are applied sequentially; at each stage, incorrect rounding with a rounding level of p < 0.051 is applied in addition. The solid grey line shows the nominal α-level of 5%.