The generalized linear model (GLM) is a flexible generalization of ordinary least squares regression. The GLM allows the linear model to be related to the response variable via a link function together with an error function. Starting with the familiar linear regression and ANOVA, the course will expand the linear model to include link functions such as the logit with binomial and the log with Poisson error distributions, thereby enabling students to model outcome variables that are not continuous. Attention will be paid to likelihood estimation methods and the checking of model assumptions.

Literature: Faraway, JJ. Extending the linear model with R : Generalized Linear, Mixed Effects and Nonparametric Regression Models, Second Edition. Chapman and Hall/CRC , 2016. Note: textbook is recommended, but not required.

At the end of the course, the student will:

• know the role of link functions and error distributions
• be familiar with the most commonly used generalized linear models
• know when to apply which model in practice
• know the most commonly used methods for checking model appropriateness and model fit
• be able to perform GLM analyses using the appropriate software (R and SPSS)
• be able to interpret the output and report the results of GLM analyses in terms of the context of the research question

At least one course in basic statistical methods up to and including simple and multiple linear regression; familiarity with likelihood methods (Wald, score and likelihood ratio tests). Students will (preferably) have completed the courses Introduction to Statistics, Classical Methods in Data Analysis, Modern Methods in Data Analysis or their equivalents.

Familiarity with the statistical package R is required!

All elements have to be awarded with at least a 5.5 in order to pass the final Assessment.