The seminar covers both, fundamental and advanced statistical techniques for causal inference from observational data. It starts by replicating the counterfactual concept of causality and its links to regression coefficients. It then discusses the possibilities (and limitations) of estimating causal effects by means of covariate adjustment. Finally, the seminar introduces techniques to deal with unobserved (or unobservable) heterogeneity.
More specifically the seminar covers the following statistical techniques: multiple regressions, exact matching, coarsened exact matching, fixed effects panel regression, instrumental variable regression, and regression discontinuity models. Thereby, the seminar aims at introducing the methods' conceptional ideas and assumptions.
- Kursleiter*in: Prof. Dr. Ulrich Kohler