P-values, Reproducibility, and Open Science

We have started a seminar series to discuss over issues of p-values, reproducibility, and open science. The idea is to have individuals in the department present their ideas on one or more of these topics, followed by an open discussion, and for all of us to learn something. If anyone else would like to speak on a topic of their choice please let Scott Glover know.

Title Presenter Notes Slides Material
Meditations on pre-registration - the challenge of secondary data Gary Lewis Pre-registration is a tricky enough issue even in cases where the data hasn't yet been collected. But it can get even more problematic when the data already exists. I'll briefly introduce the concept of pre-registration, and then discuss some of the issues I have encountered in the work I have been doing lately (i.e. using large cohort data). PDF
Problems with p-values and NHST and how to overcome them using likelihood ratios Scott Glover PDF