Addressing Researcher Degrees of Freedom in Applications, Methodological Research, and Teaching
MCML Authors
Maximilian Mandl
Abstract
Maximilian Mandl
Abstract
This thesis examines researcher degrees of freedom (RDF), the flexibility in data analysis that can lead to differing, sometimes biased, results. It introduces the minP permutation method as a more powerful alternative to traditional corrections for multiple testing and applies it in a neurosurgical study. Further, it develops approaches to address pleiotropy in Mendelian Randomisation, highlights the risks of selective interpretation in methodological research through the 'storytelling fallacy', and demonstrates these with real-world examples. Finally, it presents an educational seminar designed to raise student awareness of RDF and promote open science practices. (Shortened.)
BibTeXKey: Man25