Positive political science involves connecting our observations of the social world with causal mechanisms. We are going to focus on a particular problem: to what extent can we used observed data to measure, discover, and test underlying causal claims? The course will include a combination of statistical theory, hands--on data analysis, and causal reasoning. The goal of the course is to produce students who can understand, apply, and ultimately further quantitative political methodology.
Why do people vote the way they do? Can universal health insurance lead to a longer lifespan? What countries are more or less likely to erupt in civil conflict? Assessing these questions requires the ability to think analytically about data and statistics. This course will provide an introduction to causal inference, probability theory, and estimation. The focus of this course will be on hands-on data analysis and the practical application of basic statistical methods to real-world, relevant problems.
Political methodology is a rapidly changing and fascinating subfield of political science. From text analysis to scaling models to network models to machine learning methods, political science sits at the forefront of applied statistical research on several fronts. This course will first cover the foundations necessary to work at the frontier of political methodology, after which we will move through selected topics.
In this summer study group, we work through Asumptotic Statistics by Aad van der Vaart in some detail. The goal is to work through the book one proof at a time and fill in the gaps in the explication of his proofs (which are numerous). Topics include notions of convergence of random variables, M-Estimation and local asymptotic normality in the parametric setting, then revisiting it all again in the nonparametric setting. The typical day consists of turning to the next page of the book, working through the next theorem in close detail, then going to the next...Read more about Summer Study Group: Asymptotic Statistics