Modern Methods for Causal Inference

Catalog description

The goal of this course is to introduce a core set of modern statistical concepts and techniques for causal inference from randomized and observational studies, and to demonstrate how to use them to answer complex research questions in health research. The students will acquire knowledge on causal inference methods, including potential outcomes, directed acyclic graphs, and nonparametric structural equation models. This course focuses on aspects related to the identification of casual effects from randomized and observational studies. The course will also cover some estimation techniques such as inverse probability weighting, g-computation, matching, and doubly robust estimators based on machine learning. Time permitting, the course will cover one or more of the following topics: survival analysis, longitudinal data, mediation analyses, or effect modification. This course will use the free software R to perform all statistical analysis.