Measuring the complexity of psychological models.

Model selection is a very important aspect of scientific inquiry, that balances goodness-of-fit and complexity. A greater understanding of the complexity of models is therefore an important prerequisite for improving model selection. The overall objective of this project is to gain a better understanding of complexity. A first project is the systematic comparison of existing complexity measures, something that is not been done so far. A second project is the development of informative priors for psychological models. The third project is the development of a new complexity measure, the Prior Predictive Complexity (PPC). A final project is relating the PPC to model selection. Software will be made available to facilitate the dissemination of the PPC into mainstream psychology.

Keywords:Bayesian statistics, Model selection, Cognitive models, Priors, Complexity measure

Disciplines:Biological and physiological psychology, General psychology, Other psychology and cognitive sciences, Psychological methods, Mathematical and quantitative methods, General pedagogical and educational sciences, Social theory and sociological methods, Political theory and methodology

Project type:PhD project