400013 SE Generalised Linear Models (2025S)

This course focuses on Generalised Linear Models (GLMs) and their applications. We begin by revisiting the linear model and evaluating the linear probability model (LPM) as an approach to analysing categorical data. We discuss various estimation techniques and introduce the fundamentals of maximum likelihood estimation (MLE). We then explore a wide range of GLMs, including models for binary, multinomial, ordered, and count data. We practise presenting and visualising model results using post-estimation techniques, which are particularly useful for interpreting interaction effects in non-linear models. Finally, we cover extensions of regression models, such as hierarchical/multilevel models and models for panel data. We also address frequently asked questions in the context of regression models, such as when to cluster standard errors and when to use fixed- or random-effects specifications.

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

By the end of this course, participants will be able to analyse different types of data using generalised linear models, which are widely used in the Social Sciences. They will have a solid understanding of the statistical foundations of these models. They will be able to interpret GLMs correctly, visualise model results in an appealing way, and apply GLM techniques to their own research problems.

Prior knowledge of linear regression and familiarity with any kind of statistical software are helpful but not strictly required to participate in and complete the course successfully.