This course is closed, registration is no longer possible.
Correlation is not causation, but when can you give causal interpretations after analysing non-experimental data?
This course introduces quantitative causal inference using observational data in the social and economic sciences. You discuss the complexities of causal research, modern methods that can enable estimating effects of phenomena that were not randomized, and limitations in doing so. You will investigate a variety of specific topics in causal inference, including the design of observational studies, potential outcomes, difference-in-differences designs, discontinuity designs, matching, instrumental variables analysis for non-compliance, synthetic control methods, difference-in-differences designs with many time periods, formal sensitivity analysis, and other special topics as time and your interest permit, including causal mediation (the analysis of mechanisms) and dynamic or time-varying treatments.
You will learn how to implement these methods using the R statistical language.
