Mediation analysis is a set of tools for exploring hypotheses about causal pathways, with a special focus on differentiating “direct” from “mediated” associations between an exposure and an outcome.  Many approaches to mediation analysis are based on regression analysis.

In this workshop, we will cover some of the basic ideas behind regression-based mediation analysis, and show how this type of analysis can be performed in Python using the Statsmodels package. All software tools covered in this workshop are free and open source.