Introduction to Causal Mediation Analysis

This document is a guide that introduces causal mediation analysis, a method used to estimate the causal effect of a treatment on an outcome by evaluating the mechanisms through which this effect occurs. It explains the decomposition of the total effect into direct and indirect effects, where indirect effects involve changes in a mediator variable that subsequently influence the outcome. The guide outlines the research process involved in causal mediation analysis, including specifying research questions, identifying treatment and mediator variables, and selecting appropriate estimation methods. It also discusses various approaches to mediation analysis, including classical regression models and potential-outcomes frameworks. The document provides examples and workflows for researchers to follow when conducting causal mediation analysis, emphasizing the importance of understanding both direct and indirect effects in causal inference. Additionally, it introduces the 'mediate' command for fitting causal mediation models, detailing how to estimate effects using this command in statistical software.