ReproductibiliTEA !

By Ladislas Nalborczyk
NeuroSpin, CEA-Saclay
Computational cognitive neuroscientist and author of the Barely significant blog

Numerous authors have highlighted the limitations of the Null-Hypothesis Significance Testing (NHST) approach and the near exclusive reliance on p-values and significance testing. Despite being taught and used as the default paradigm for statistical inference, one of the main barriers to its correct use remains its complexity, often hidden by misleading intuitive interpretations, and the false belief that it would be somehow aligned with the falsificationist paradigm. In this talk, I will draw on the work of Paul Meehl to show how these misinterpretations and misconceptions have impeded theoretical progress in Psychological sciences (broadly speaking). I will then consider how we can confront the ongoing theory crisis through a coherent philosophy of science. Ultimately, this will require us to develop stronger theories and to bridge the gap between substantive theories and statistical hypotheses. I will illustrate how this can be implemented, at the statistical level, by a recently proposed “Bayesian workflow”, which promotes an iterative model building, criticism, and comparison approach.

📍 Timone campus, main building, blue wing, 5th floor, Amphi 5
🗓 Thursday, February 15, 2024
🕐 3 PM : discussion between students and speaker
🕐 4 PM : seminar and questions, open to all students, researchers and lab staff
☕ Followed by an afternoon tea

👍 Can be validated as 2 hours of PhD training : PhD students : please register now on AMeTICE!