Bayesian hypothesis testing hierarchical models: A PhD proposal for the innovation of psychological methods
Ruud Wetzels (PhD student)
Psychological Methodology, Department of Psychology, University of Amsterdam
Project: Project financed by University of Amsterdam
Project running from: 1 September 2008 – 1 September 2012
Promotores: Prof. dr H.L.J. Van der Maas (University of Amsterdam), Dr E.M. Wagenmakers (University of Amsterdam)
Summary:
One goal of my PhD project is to do Bayesian inference using all kinds of models that are popular in Psychology. Some examples of such models are ALCOVE (Kruschke, 1992) for category learning or the Expectancy-Valence model (Busemeyer and Stout, 2002) for decision making.
Another goal of the project is to implement and study Bayesian hypothesis testing for hierarchical, possibly order-restricted models. In hierarchical modeling, individual-level parameters are drawn from a group distribution. This way of modeling takes both differences and similarities between participants into account.
In general, the aim is to try to make Bayesian methods more easily available to empirically oriented psychologists that would like to take advantage of the Bayesian methodology but lack the time or the technical skills to implement their own software.
Date of defence: 26 September 2012
Title of thesis: Bayesian model selection with applications in social science.
ISBN: 978-94-6191-404-0
Publisher: Ipskamp Drukkers B.V., Enschede.