Hierarchical Bayesian Regression for experimental psychology: a case study of cognitive control

Area of Science:

• Cognitive Psychology
• Statistical Modeling

Background:

• The replication crisis in experimental psychology necessitates re-evaluating analytical practices.
• There is a growing demand for methods that improve the accuracy and precision of research findings.

Purpose of the Study:

• To explore the utility of Hierarchical Bayesian Regression (HBR) models in addressing the replication crisis.
• To demonstrate the advantages of HBR for statistical inference using the Dual Mechanisms of Cognitive Control (DMCC) task battery.

Main Methods:

• A case study using two existing DMCC datasets.
• Application of two sets of HBR models, with the first set's estimates informing the second set's priors.
• Systematic generation of cumulative posterior distributions for all effects.

Main Results:

• HBR models provided novel insights into proactive/reactive control indicators across four DMCC tasks.
• Demonstrated consistency of effects across datasets.
• Quantified the strength of null effects and modeled response time and accuracy at the trial level.

Conclusions:

• Hierarchical Bayesian Regression (HBR) offers a rigorous and advantageous approach to statistical inference in cognitive psychology.
• HBR enhances the understanding of cognitive control mechanisms by providing precise estimates and modeling complex data properties.
• This methodology can help mitigate issues highlighted by the replication crisis.