Week 3

Amanda Kay Montoya - Assistant Professor, University of California

Combining Statistical and Causal Mediation Approaches

Mediation analysis is an important analytical method for psychologists because it can help address questions about processes, or why an effect occurs. Mediation is an inherently causal process, and causal mediation analysis can be used to evaluate the assumptions required to estimate causal effects as well as methods for evaluating these assumptions. Statistical mediation can be used to estimate indirect effects and make inference about those effects. Statistical mediation is ubiquitously used in psychology, but the ideas from causal mediation analysis are discussed much less. In this presentation, I introduce researchers to causal mediation analysis and how to integrate practices from this approach into existing (statistical) mediation analysis methods. I present principles of mediation analysis in the modern age which combine practices from statistical and causal mediation analysis to present recommended practices for planning, conducting, and reporting mediation analyses.

This presentation is a companion to Montoya, A. K. (in press). Combining Statistical and Causal Mediation Analysis. Handbook of Research Methods in Social and Personality Psychology (H. T. Reis, T West, C. M. Judd, Ed.) Cambridge University Press. Available as preprint: https://psyarxiv.com/f8paj/