Disentangling p-hacking from publication bias in finance research

Public trust in scientific knowledge depends crucially on the validity and robustness of published empirical findings. The trust can be compromised by distortions in the pool of published results that arise from p-hacking and publication bias. These problems are particularly substantial in finance, where rich observational data give ample opportunities for multiple estimation of parameters and their selective reporting. In this research project, we develop methodological tools to distinguish between p-hacking and publication bias in meta-analysis, apply the novel methodology to quantify the impact of these biases in four major fields of finance research, and compile a large data set to compare the magnitude of resulting distortions in finance research and related disciplines. Our findings will help scientists understand the underlying causes of the replication crisis and measure the underlying effect of finance phenomena in meta-regression analysis, investors assess their systematic risk exposure, and policymakers find the optimal design of financial markets and banking regulation.

Project team of FSV UK (5)

prof. PhDr. Tomáš Havránek, Ph.D.

Principal investigator

+420 222 112 309  |  +420 224 412 318

tomas.havranek@fsv.cuni.cz

doc. Bc. Jiří Novák, M.Sc., Ph.D.

Team member

+420 222 112 314

jiri.novak@fsv.cuni.cz  |  novakji@fsv.cuni.cz

Mohammad Ali Elminejad Anjileh, M.Sc., Ph.D.

Team member

m.ali.elminejad@gmail.com  |  ali.elminejad@fsv.cuni.cz

Team member

+2 more team members Show all team members