The class has three broad goals, which I enumerate below. Each goal is associated with a number of specific learning outcomes, which you should be able to perform as the class concludes. Students who complete the class should be able to
- formalize a research question about causal relations using causal graphs and counterfactuals. This includes
- translating theoretical arguments about causal relations into a corresponding graphical model,
- specifying the causal relation of interest and defining this relation in terms of counterfactual contrasts,
- being able to conceptually distinguish this causal relation from statistical association.
- use graphical models to devise strategies for identifying the causal relation of interest. For this, students
- demonstrate that they are capable to derive empirical implications from a graphical causal model,
- understand the theoretical assumptions necessary to test these implications,
- critically evaluate whether these assumptions hold in applied social research.
- obtain an estimate approximating the causal relation of interest and, if feasible, test underlying assumptions. To do so, students
- adapt existing
Statacode for their purposes,
- correctly interpret the resulting estimates,
- understand and perform tests of the validity of the analyses.
- adapt existing
As a bonus, I hope the class leads you to appreciate the ubiquity of causal inference in research, work, and daily life along with the recent advances in the concepts and tools to systematically undertake it. The class, by far, doesn’t cover everything there is. But it should provide you with enough understanding to learn more on your own.