Bayesian Hierarchical Models in Psychological Science
Jeffrey N. Rouder, Jordan M. Province · 2019
This chapter provides a gentle introduction to Bayesian hierarchical modeling. It aims to make the material accessible to a wide audience including those who have a limited aptitude in mathematics and statistics. Bayesian analysis starts with a different definition of probability than what most psychologists learn and teach in their statistics courses. Bayesian probability offers an elegant and intellectually pleasing alternative to traditional notions of probability. To illustrate the advantages of Bayesian hierarchical modeling, we provide an application for the process of lexical access during reading. In the course of reading, printed strings on the page must be matched to words in the lexicon to build semantic meaning. With modern advances in Bayesian analysis, Bayesian analysis of hierarchical nonlinear models of psychological processes is possible and relatively straightforward. Researchers sometimes draw a sharp distinction between models that specify plausible psychological mechanisms and those that just stipulate noise.