Discrete choice models are statistical tools for examining decisions: where to live, how to get to work, where to go to school, who to be friends with, which political party to vote for, and what dish soap to buy. They are widely applicable in network, segregation, transportation, and marketing research and across the disciplines of business, sociology, political science, and economics.
This course provides masters and doctoral students with a practical overview of discrete choice methods. It will cover the foundational, multinomial logistic regression model, and then consider extensions, including mixed logit and latent class models. Exercises based on both empirical and simulated data will allow students to gain an intuitive understanding of the strengths and limitations of the methods. Applications to agent-based and micro-simulation models will be considered. It is strongly preferred, but not required, that students have experience with regression modeling.