Beyond Mediation and Modification: A Deeper Look at Cause and Effect

The rest of this book is organized as follows. Chapter 2 covers the literature review of the Neyman-Rubin’s counterfactual framework of causal inference, general assumptions for causal studies, Pearl’s graphical causal modeling, causal mediation analysis, and causal modification analysis. The section of causal mediation study describes motivations for empirically assessing mediation, traditional mediation analysis, counterfactual mediation analysis, assumptions for identifying the direct and indirect effects, Pearl’s causal mediation formula, and regression-based approaches for mediation study. The section on causal modification study describes motivations for assessing causal effect heterogeneity, definitions and notations for modification, relations between modifier and mediator, assumptions for identifying conditional average treatment effect, and stratification analysis and other machine learning approaches for causal modification study.