Applying hierarchical bayesian modeling to experimental psychopathology data: An introduction and tutorial.

Area of Science:

• Clinical Psychology
• Statistical Modeling
• Psychopathology Research

Background:

• Bayesian methods are increasingly popular across scientific fields but are underrepresented in graduate clinical statistics training.
• Significant barriers, including advanced mathematical concepts and complex computations, limit researchers' adoption of Bayesian approaches.
• There is a need to enhance the accessibility of Bayesian inference for researchers in psychopathology.

Purpose of the Study:

• To provide a clear introduction to the Bayesian inference framework for psychopathology researchers.
• To offer a practical tutorial on implementing Bayesian methods, specifically hierarchical Bayesian modeling (HBM).
• To demonstrate the application of HBM using real experimental psychopathology data.

Main Methods:

• A primer on core Bayesian inference concepts and implementation considerations for Bayesian estimation.
• Application of hierarchical Bayesian modeling (HBM) to a dataset from schizophrenia, bipolar disorder, and healthy control groups.
• Modeling individual responses and group differences using probability functions suitable for hierarchical data structures.

Main Results:

• Demonstration of HBM for analyzing psychophysical gaze perception data in clinical and control groups.
• Provision of accessible code and data for replicating and visualizing results.
• Comparison of Bayesian results (posterior probabilities) with those from traditional statistical methods.

Conclusions:

• Hierarchical Bayesian modeling provides a flexible framework for analyzing complex psychopathology data.
• This tutorial facilitates the integration of Bayesian methods into the statistical toolkit of psychopathology researchers.
• The study promotes greater accessibility and understanding of advanced Bayesian techniques in clinical science.