This is an introductory course to generalized linear modeling (GLM) that I have been teaching since 2014 at the ECPR Summer School in Methods and Techniques. GLM is a broad technique used to perform regression analysis when the dependent variable is not linear (i.e. most of the times). This course only covers the most common types of dependent variables in social science, and it puts a particular emphasis on the presentation and visualization of quantities of interest.
The course is taught in R, and a basic proficiency with it is essential to fully understand the stats and not lose too much energy trying to get the code. Here you can find a short tutorial for the R functionalities that are used in the course. Some examples are taken from the book "R in a Nutshell" by Joseph Adler, a very well written introductory text to R that I recommend seamlessly.
Below you can find the most recent slides and scripts for the lab sessions. The course schedule and reading list (and a creepy picture of myself) are available here. Data come from different sources, and are available upon request.
Day 1: Modeling Binary Response Variables: what to do? [Slides] [Lab]
Day 2: The general logic of GLM and Maximum likelihood [Slides] [Lab]
Day 3: Interpreting coefficients of logit models, quantities of interest, and interactions [Slides] [Lab]
Day 4: Modeling categorical and ordinal variables [Slides] [Lab]