Please feel free to use and adapt lecture slides (with their LaTeX sources), R code and data in your own teaching. Each course site gives its license and how to cite it.
Six lectures on causal inference with panel data, from two-way fixed effects to heterogeneity-robust DID, synthetic control and latent factor models, with R code in panelView and fect. description
Panel data are the most common basis for causal claims in the social sciences. For decades the default analysis was a two-way fixed effects regression. Over the past ten years that default has been challenged, and dozens of alternative estimators have appeared. This short course works through the change, from the classical models to the methods now in use.
The first half covers the parametric panel models and what their assumptions require, then difference-in-differences (DID) and factorial DID as distinctive research designs. It closes with the result that two-way fixed effects recovers a weighted average of treatment effects whose weights can turn negative when units are treated at different times, as well as other pathologies.
The second half covers the responses: heterogeneity-robust DID estimators, the synthetic control method and latent factor models, as well as doubly and triply robust methods.
Each lecture pairs the identification argument with applied examples and diagnostics: what to plot, what to test, and what to report when the assumptions are in doubt. Implementation is in R, using panelView, fect, and related packages.
Participants should have taken a graduate or advanced-undergraduate level regression course. Prior experience with R is helpful but not required.
Earlier version (August 2021). This workshop series gives an overview of newly emerged causal inference methods using panel data (with dichotomous treatments). We start our discussion with a review of the difference-in-differences (DiD) method and conventional two-way fixed effects (2WFE) models. We then discuss the drawbacks of 2WFE models from a design-based perspective and clarify the two main identification regimes: one under the strict exogeneity (SE) assumption (or its variants) and one under the sequential ignorability (SI) assumption. In Lecture 2, we review the synthetic control method and discuss its extensions. In Lecture 3, we introduce the factor-augmented approach, including panel factor models, matrix completion methods, and Bayesian latent factor models. In Lecture 4, we take a different route and discuss matching and reweighting methods to achieve causal inference goals with panel data under the SE or SI assumptions. We also discuss hybrid methods that enjoy doubly robust properties. These lectures were first given at Washington University in St. Louis in August 2021. A caution: they reflect my thinking at the time, and my views on some of these methods have since shifted; where the two versions differ, the current lectures reflect my present view.
A graduate introduction to regression and the probability and statistics behind it; the first course in Stanford PoliSci’s four-course methods sequence. description
This is the first course in a four-course sequence on quantitative political methodology at Stanford Political Science. Political methodology is a growing subfield of political science which deals with the development and application of statistical methods to problems in political science and public policy. The subsequent courses in the sequence are 450B, 450C, and 450D. By the end of the sequence, students will be capable of understanding and confidently applying a variety of statistical methods and research designs that are essential for political science and public policy research.
This first course provides a graduate-level introduction to regression models, along with the basic principles of probability and statistics which are essential for understanding how regression works. Regression models are routinely used in political science, policy research, and other disciplines in social science. The principles learned in this course also provide a foundation for the general understanding of quantitative political methodology. If you ever want to collect quantitative data, analyze data, critically read an article that presents a data analysis, or think about the relationship between theory and the real world, then this course will be helpful for you.
You can only learn statistics by doing statistics. In recognition of this fact, the homework for this course will be extensive. In addition to the lectures and weekly homework assignments, there will be required and optional readings to enhance your understanding of the materials. You will find it helpful to read these not only once, but multiple times (before, during, and after the corresponding homework).
This course was originally designed by Jens Hainmueller, and maintained and updated by me.
Co-taught with Simone Paci, the course pairs how AI systems work and how to use them well with the politics of AI, from labor markets and elections to data centers and chips; students design a project that uses or studies AI on a social or political problem. description
Artificial intelligence is now both a political object and a tool for political and social life. This course, co-taught with Simone Paci, covers both. No advanced programming or computer science background is needed.
The course runs two parallel tracks. On Mondays, we cover the political economy of AI. Those sessions ask who demands it and to what effect (labor markets, firms, public opinion), how politicians and public institutions put it to use, and what conflicts follow over data centers, chips and export controls, and data governance.
On Wednesdays, we teach how AI systems work and how to use them well, covering statistical and deep learning, natural language processing, large language models, post-training and reasoning models, computer vision, and AI agents, including their reliability and safety.
Weekly sections carry the hands-on work, such as calling AI models from light Python code, small data exercises, and tool building. The two tracks meet in a final project that uses or studies AI in relation to a concrete social or political challenge, presented in themed panels at the end of the quarter.