Daniel W. Heck, Philipps University of Marburg
Cognitive Psychometrics using Hierarchical Multinomial Processing Tree Models
Date & Time: Wednesday, July 21 at 9:00am EST
Many psychological theories assume that different cognitive processes can result in the same observable responses. Multinomial processing tree (MPT) models allow researchers to disentangle mixtures of latent processes based on observed response frequencies. MPT models have recently been extended to account for participant and item heterogeneity by assuming hierarchical group-level distributions. Thereby, it has become possible to link latent cognitive processes to external covariates such as personality traits and other person characteristics. Independently, item response trees (IRTrees) have become popular for modeling response styles. Whereas cognitive and social psychology has usually focused on the experimental validation of MPT parameters at the group level, psychometric approaches consider both the item and person level, thus allowing researchers to test the convergent and discriminant validity of measurements. Bridging these different modeling approaches, Bayesian hierarchical MPT models provide an opportunity to connect traditionally isolated disciplines in psychology.
About the Speaker
Daniel Heck is an Associate Professor of Psychological Methods in the Department of Psychology at Philipps University of Marburg, Germany. His research focuses on Bayesian model selection, cognitive modeling, judgment and decision making, and social desirability. His research has been recognized with the 2021 William K. Estes Early Career Award of the Society for Mathematical Psychology, the nation-wide 2018 Heinz-Heckhausen Award of the German Psychological Society, and the 2018 “Rising Star” recognition of the Association for Psychological Science.