Modelling Multiverses and Correcting for P-hacking
The goal of this PhD project is to formalize and validate the new multiverse model for p-hacking and to create novel tools that can be used to correct for biases in meta-analysis.
The new model potentially offers fundamental understanding of multiverse analyses and how selection for significant outcomes (p-hacking) might create biases in estimates of causal effects and inflated false positive rates in (psychological) experiments. The new model need to be worked out in detail and assessed formally and in relation to simulated data.
Conducting a systematic review of simulation studies of p-hacking in the context of psychological experiments and to synthesizing these results in the new model.
Developing and refining new meta-analytic tools to correct for p-hacking based on the new multiverse model.
Prof. dr. J. Wicherts
prof. dr. M. van Assen
dr. R. van Aert
1 September 2023 – 1 September 2027