Causal Inference is the process of measuring how specific actions change an outcome. For example, you might want to determine whether a recent marketing campaign increased sales, whether leveraged buyouts increase the probability of bankruptcy, or whether government lockdowns slowed the spread of covid-19. But for questions about cause and effect, running a basic regression can often get us the wrong answer. You have probably heard the warning that “correlation is not causation”. In this course we will explore what we mean by “causation”, how correlations can be misleading, and how to measure causal relationships when we can’t perform a perfect randomized experiment. The course will emphasize applied skills, and will revolve around developing the practical knowledge required to conduct causal inference in R.
Students should have some experience with R, and a basic understanding of Ordinary Least Squares (OLS) regression, including how to interpret coefficients, standard errors, and t-tests.
- Develop a conceptual framework for analyzing causal questions that allows you to critically examine when statistical estimates can be interpreted as causal;
- Learn practical skills in R, including how to simulate data to test statistical estimators, and how to implement standard casual estimators; and,
- Learn how to estimate causal effects in real-world settings.