Thursday, October 25th - 9:00 AM - 12:00 PM
OPEN SCIENCE WORKSHOP
Instructor: Geoff Cumming, La Trobe University, Melbourne
This workshop is in 2 parts. The first (45 minutes), suitable for a broad audience, describes the principles of open science. The second part (2 hours) examines in depth the new statistics associated with open science.
Open Science: Research in the Post-p<.05 World (45 minutes)
Many researchers in the health sciences, as in many other disciplines, believe that the aim of quantitative research is to achieve statistical significance: They see p < .05 as the key to publication, jobs, funding, and fame. But now there is clear evidence that published research in medicine, psychology, and other disciplines is often wrong—this is the replication crisis. Reliance on p < .05 is a major cause of the crisis, and thus has caused massive damage. The rapid rise of Open Science practices in response to the crisis is probably the most important and exciting development in how science is done for many years. It’s already leading to more trustworthy research. Open Science requires preregistration where possible, open and full reporting, and replication. It also requires improved statistical methods, beyond p < .05. I’ll explain what consumers of research need to look for, to have confidence in reported research. I’ll also outline how researchers can adopt Open Science and improved statistical practices.
The New Statistics: What We Need for Open Science (2 hours)
Open Science requires replication, which requires meta-analysis, which in turn requires estimation—i.e. the new statistics, based on confidence intervals. A major obstacle to full adoption of Open Science and estimation is researchers’ deep, deep attachment to p values and statistical significance. I’ll discuss how this attachment might be addressed, and how Open Science and the new statistics is actually easier for students to learn, understand, and use. I will discuss and demonstrate an estimation approach to the analysis of data for a range of measures and designs, with an emphasis on using graphical representations likely to assist understanding and statistical communication. I’ll mention our new introductory statistics textbook that takes an estimation and Open Science approach from the very start. I will be happy to try to tailor the choice of topics and examples in accord with the preferences of workshop participants.
Instructors: Fernando Fernandez-Aranda (Editor-in-Chief, European Eating Disorders Review) Phillipa Hay and Stephen Touyz (Editor(s)-in-Chief, Journal of Eating Disorders) and Ruth Striegel Weissman (Editor-in-Chief, International Journal of Eating Disorders)
This is a unique opportunity to connect with four editors of three pre-eminent eating disorder journals in one room! There will be ample opportunity for interaction, discussion of practical examples, and questions.