An Article by Assistant Professor Hiroyuki Muto Has Been Published in the Japanese Journal of Psychonomic Science
An article by Assistant Professor Hiroyuki Muto, entitled “Introduction to hierarchical Bayesian modeling for experimental psychologists: A tutorial using R and Stan,” has been published in the Japanese Journal of Psychonomic Science (Vol. 39, No. 2).
Statistical modeling is a method to understand and predict phenomena by representing the data generating process with a probability model and applying the probability model to the data. In recent years, with the spread of Bayesian estimation software using the MCMC method, a general-purpose estimation algorithm, it has become possible to fit complex models to data with flexibility and ease, which was difficult to do with conventional methods. Combining such Bayesian statistical modeling with psychological experiments is expected to become a powerful tool for empirical and mathematical verification of the mechanism of human information processing.
This article provides a tutorial-style explanation of Bayesian statistical modeling in practice, using three well-known experimental paradigms in perceptual and cognitive psychology (the mental rotation task, the psychophysical task, and the Eriksen flanker task) as examples. In particular, it focuses on the concept of hierarchical Bayesian modeling, which is highly expandable and applicable to a wide range of phenomena, including the ability to perform analyses that simultaneously consider both inter-individual variability and inter-trial variability. In addition, this article briefly discusses several topics relevant to the practice of statistical modeling (e.g., model selection, research transparency, and secondary analysis).
Hiroyuki Muto (2021)
“Introduction to hierarchical Bayesian modeling for experimental psychologists: A tutorial using R and Stan”
Japanese Journal of Psychonomic Science, 39(2), 196-212. (Open Access, Japanese version only)