Proceedings of the Annual Meeting of the Cognitive Science Society

Cognitive measurement models decompose observed behav-ior into latent cognitive processes. For situations with morethan one condition, such models allow to test hypotheses onthe level of the latent processes. We propose a fully Bayesianensemble model approach to test hypotheses on the level ofthe latent processes in situations in which multiple measure-ment models or model classes exist. In the first step, one needsto perform a Bayesian model selection step comparing the hy-potheses within each model class. Aggregating the results ofthe first step yields ensemble posterior model probabilities. Weprovide an example for a working memory data set using anensemble of a resource model and a slots model.