RL2024-1
Causal Moderated Mediation Analysis – A Causal Investigation of Heterogeneity in Mediation Mechanisms: Methods and Software
Recorded: Tuesday, June 4, 2024

INSTRUCTOR

Xu Qin, University of Pittsburgh

Research questions regarding how, for whom, and where a treatment achieves its effect on an outcome have become increasingly valued. Such questions can be answered by causal moderated mediation analysis, which assesses the heterogeneity of the mediation mechanism underlying the treatment effect across individual and contextual characteristics.

The purpose of this course is to introduce the definition, identification, estimation, and sensitivity analysis for causal moderated mediation effects under the potential outcomes framework. Participants will also learn how to use a user-friendly R package to conduct the analysis and visualize results. The method introduction and the package implementation will be illustrated with a re-analysis of the National Evaluation of Welfare-to-Work Strategies (NEWWS) Riverside data.

The course will include the following activities and objectives:
 

1. A lecture introducing

- background of mediation analysis and moderated mediation analysis
- definition, identification, and estimation of causal mediation effects and causal moderated mediation effects under the potential outcomes framework
- sensitivity analysis for assessing how sensitive results are to potential unmeasured confounding

2. A hands-on training in using an R package to

- estimate the causal mediation effects and causal moderated mediation effects
- conduct sensitivity analysis
- visualize the original analysis results and sensitivity analysis results

The course is targeted at any researchers from PhD students, faculty, or research practitioners. Participants should have basic knowledge of statistical inference and multiple regression at least. Participants are expected to have the R software and the R package installed on their computers before the course.