PRE-CONFERENCE WORKSHOP I: BAYESIAN CAUSAL MEDIATION ANALYSIS

The goal of the workshop is to describe statistical, methodological, and conceptual aspects of Bayesian causal mediation analysis. The one-day workshop consists of four parts. Part 1, covers definitions, history, and applications for the mediation model followed by estimation of mediation effects including assumptions, statistical tests, and confidence intervals. The methods described in this section serve as the foundation for causal inference methods in Part 2 and Bayesian mediation analysis in Part 3. Part 2 describes the potential outcomes framework and applies it to the single mediator model. Participants will learn about the causal estimators for the single mediator model and the assumptions necessary to make causal inferences. Part 3 describes methods for Bayesian mediation analysis and the extension of these techniques to Bayesian causal mediation analysis. Differences between frequentist and Bayesian frameworks are described, and inferences from the posterior distributions of the causal estimators are discussed. In Part 4, participants are taught how to perform Bayesian causal mediation analysis in SAS, Mplus, and R.