MKTG660

In this seminar we will focus on causal models from observational data. Physical randomization with experiments are one of the cleanest ways to build causal models. However physical randomization - either through a lab experiment or a field experiment - is not always feasible. Often times we are restricted to delineate causal effects from observational and secondary data. This seminar will cover foundational elements of modern practices of causal inference using observational data. We will cover potential outcomes model as well as discuss in detail the most common designs: sub-classification and propensity score, regression discontinuity, instrumental variables and difference in differences.