Substantive questions in empirical social science research are often causal. Does voter outreach increase turnout? Do political institutions affect economic development? Are job training programs effective? This graduate-level class will introduce students to both the theory and the practice behind making these kinds of causal inferences. We will cover causal identification, potential outcomes, experiments, matching, regression, difference-in-differences, instrumental variables estimation, regression discontinuity designs, sensitivity analysis, dynamic causal inference, and more. The course will draw upon examples from political science, economics, sociology, public health, and public policy. Taught with R, with a mix of lectures, discussion, and in-class computing.

• Syllabus

Lecture notes:

• 3 - Potential Outcomes
• 4 - Randomized Experiments
• 5 - Randomization Inference
• 6 - Observational Studies
• 7 - Matching
• 8 - Post-treatment bias and weighting
• 9 - Regression
• 10 - Panel data and fixed effects
• 11 - Differences in differences
• 12 - Instrumental Variables
• 13 - Regression Discontinuity
• 14 - Causal Mechanisms
• 15 - Dynamic Causal Inference

Homeworks:

• Homework 1
• Homework 2 - Roads Data
• Homework 3 - Abductions Data
• Homework 4 - NSW Experimental Data NSW Observational Data
• Homework 5 - Washington paper Washington Data
• Homework 6 - Card and Krueger paper Card and Krueger data
• Homework 7 - Acemoglu, Johnson, and Robinson data Lee data