400011 SE Regression models for categorical data (2021S)

This course covers regression models for categorical data. We start by revisiting the linear model and evaluate the linear probability model (LPM) as a first approach to analyzing categorical data. We discuss different estimation techniques and the fundamentals of maximum likelihood estimation (MLE) will be introduced. Next, we explore a broad range of generalized linear models (GLMs) including models for binary, multinomial and ordered outcome variables. We learn how to interpret and visualize their model results by deriving quantities of interest. Finally, we will cover further extensions of these models, such as hierarchical/multilevel models and models for panel data.

Each session will consist of a short lecture followed by practical exercises using a statistical software (R or Stata).

By the end of this course, participants will be able to analyze different types of categorical data using regression techniques widely used in the Social Sciences. They will have a solid understanding of the statistical foundations of these models. They will also be able to interpret those models correctly and apply them to their own work.

Prior knowledge of linear regression and the familiarity with any statistical software will be helpful but are not required in order to complete the course successfully.