Causal Inference


While working as a research fellow at the WZB, I had the chance of co-teaching a class on Causal Inference at the Hertie School for the students in the Masters in International Affairs and Masters in Public Policy. This course represented the 2nd part of the sequence in statistical modeling for the MIA and MPP students.

My co-instructor in the class was Max Schaub. Throughout the class we were helped by a wonderful team of teaching assistants: Adelaida Barrera and Sebastian Ramirez-Ruiz. Both Adelaida and Sebastian were invaluable during the course of the class, from running laboratory sessions, to developing code files, and many other tasks in between. All code files developed for the seminar sessions were entirely authored by them, and they deserve full credit for this. We couldn’t have wished for better support, for which we remain grateful to this day!

Assuming prior knowledge in simple and multiple linear regression modeling, it introduced students to a new perspective on studying causes and effects in social science research. Based on a framework of causality, the course agenda covered various strategies to uncover causal relationships using statistical tools. We started with reflecting about causality, the ideal research design, and then learned to use a framework to study causal effects. We then revisited common regression estimators of causal effects and learned about their limits. Next, we focused on matching, instrumental variables, difference-in-differences and fixed effects estimators, regression discontinuity designs, and techniques to explore moderated and mediated relationships. All classes divided time between theory and application.

If interested, consult all slides and code files here.