Concepts (I)

Mediation Analysis

The section provides an overview of the concept and methods of mediation analysis in the context of epidemiology and statistical modeling. The section discusses total effects, indirect and direct paths, and highlights the limitations of the Baron and Kenny approach (Baron and Kenny 1986). It introduces the counterfactual definition, emphasizing the importance of adjusting for confounders and providing insights into the mechanics of mediation analysis through imputation and weighting methods. The section also covers sensitivity analysis, proportion mediated, and methodological extensions, including multi-category mediators and software references for conducting mediation analysis.

Reading list

Key reference: (Rochon, Bois, and Lange 2014; Lange et al. 2017)

Video Lessons

Mediation Analysis - Baron and Kenny (1986)
Setting up Mediation analysis with counterfactual definitions and why confounding adjustment helps
Mediation analysis mechanism under counterfactual definition (i) Outcome imputation & (ii) Weighting
Mediation analysis using survey data, assumptions, extensions, software and references

Video Lesson Slides

References

Baron, Reuben M, and David A Kenny. 1986. “The Moderator–Mediator Variable Distinction in Social Psychological Research: Conceptual, Strategic, and Statistical Considerations.” Journal of Personality and Social Psychology 51 (6): 1173.
Lange, Theis, Kristoffer W Hansen, Rune Sørensen, and Søren Galatius. 2017. “Applied Mediation Analyses: A Review and Tutorial.” Epidemiology and Health 39.
Rochon, Justine, Andreas du Bois, and Theis Lange. 2014. “Mediation Analysis of the Relationship Between Institutional Research Activity and Patient Survival.” BMC Medical Research Methodology 14 (1): 9.