Learning Objectives

Define in plain language – and illustrate with social epidemiology examples – the assumptions for valid causal inference in observational research
Contrast propensity score, inverse probability of treatment weight, and multivariable regression for control of confounding
Explain controlled direct effect, natural direct effect, and natural indirect effect in the context of mediation or effect decomposition analysis, particularly in the context of health equity
Given data, and a DAG, estimate and interpret the controlled direct effect, natural direct effect, natural indirect effect and total effect of an exposures-mediator-outcome association

For Monday

Required Readings

Recommended Readings

Oakes & Kaufman, Chapter 17. Ahern & Hubbard: A roadmap for estimating and understanding population intervention parameters
Gilsanz et al. - 2021 - Marginal Structural Models for Life-Course Theorie.pdf

Lecture materials

Handouts/activities

For Friday

Required Readings

Recommended Readings

Lecture materials

Handouts/activities

Assignments Due This Week

Coming Up