3M Selection and Strategic Models
Prof Curtis S. Signorino, University of Rochester
5 - 16 August (two week course / 35 hrs)
Detailed Course Outline [PDF]
Course Content
We will examine methods for incorporating self-selection, choice, and strategic interaction into statistical analyses using experimental and non-experimental data. The material will draw heavily from the literatures on selection, random utility, basic game theory, quantal response equilibrium (QRE), and structural estimation of strategic models. Students will be introduced to the theoretical and applied work in these areas, and will have an opportunity to apply these techniques using the R statistical programming language.
Course Objectives
Individual choice is central to social and political behaviour. Modelling choice can be as important for statistical analysis as it is for theorizing. The course objectives are (1) to demonstrate the bias that results from ignoring selection and strategic choice in our statistical models, (2) to teach students how to use and interpret existing techniques (e.g., Heckman and strategic models), and (3) to provide students with the basic tools to develop and implement new selection or choice-based models of their own.
Course Prerequisites
Students are expected to be familiar with calculus (integration and differentiation), basic probability and inferential statistics, ordinary least squares, simultaneous equations, maximum likelihood estimation (MLE), and the R statistical programming language. Actual programming experience (in R or any other language) is not required. Students should also have a basic understanding of introductory game theory (e.g., Nash equilibrium and subgame perfection).
Reading
No texts are required. All readings will be from articles or lecture notes provided by instructor.