Transparent practices in empirical science (a ‘how-to’ workshop)

Event type: 
21 June 2017
12 -2 pm

Rountree Room, D26

A/ Prof Tim Parker
Whitman College, USA

In some disciplines there is discussion of a ‘reproducibility crisis’ triggered by recognition that many published findings do not seem to hold up to further evaluation. Many published results, it seems, are wrong. At the root of this crisis is NOT fabrication of results, but instead a lack of transparency in how science is conducted and reported. Although direct evidence of low reproducibility has been generated for only a few disciplines, there is widespread evidence that the poor transparency which drives low reproducibility is worryingly common in many disciplines, including in ecology and evolutionary biology.


This workshop is designed to help scientists (from any empirical discipline) navigate the reproducibility crisis in two ways. First, participants will learn some* methods for being more transparent in their own work. This should reduce unintended bias and increase the reproducibility of your work. Second, participants will learn to better recognize insufficient transparency in published work, and to understand the potential implications of this insufficient transparency.


The workshop will be hands-on. You will work through the process of maximizing transparency in your own work (come ready to think about your own projects; ideally one project in the planning stage and one project at the writing stage, though you can use projects from any stage and if you don’t have your own project, you will be paired with other participants), and you will practice identifying insufficient transparency in sample work that will be provided at the workshop.


This workshop is scheduled for one hour, but I will be available during the following hour as well if anyone wants continued guidance as they work to improve the transparency of their current projects. 


*There are many great ideas for improving transparency. We will be focusing on a subset of these.