Indiana University Bloomington

Psychological and Brain Science P657
Replicability And Mediation

Contact: Prof. Eliot R. Smith (
Offered: Spring, 2016
Pre-Requisites: Recommended: Graduate-level statistics coursework up to and including multivariate analysis (i.e., multiple regression). Please contact the instructor if you are interested in the course but uncertain about whether you have the necessary background.
Syllabus: No Syllabus Avaliable
Description: Psychology has recently been claimed to be in a “crisis” due to the non-replicability of much published research. For example, the Open Science Collaboration (Science, 28 August 2015) found significant replication results for only 36% of the studies they attempted to replicate across social and cognitive psychology. This course will address underlying causes, effects, and potential remedies for this situation. Is psychology unique in facing issues of nonreplicability? What structural changes in the field might help? Perhaps most important, what can individual researchers do to improve the situation?

The second major topic of the course is the analysis of mediation, ways to answer the question of how and why an experimental manipulation or other cause has a particular effect. Because psychology seeks to identify underlying theoretical processes rather than simply to document empirical relationships, methods of establishing mediation are becoming increasingly important. We will consider state-of-the-art methods as well as current controversies in this area. As part of the course, students will conduct mediation analyses on appropriate datasets of their own, or datasets that will be made available.