Only open to Master’s and Research Master’s students from Psychology.
In this course the focus will be on a whole range of regression models. Starting with the linear least squares regression model with numeric and categorical explanatory variables and a continuous normally distributed response variable we expand to regression models for other type of response variables, such as Poisson regression models for counts and logistic regression models for dichotomous response variables. After introducing these models, we discuss the unifying theory of generalized linear models, that encompasses linear, Poisson, and logistic regression among other regression models. In the second part of the course, we discuss regression models for data with dependencies, that is correlated data. Having correlated data, the critical independence assumption in generalized linear models is violated. To deal with this violation, a new class of regression models will be introduced, called multilevel models. Different forms of correlated data are discussed, together with model opportunities. For all regression models discussed throughout the course we will discuss assumptions and diagnostics to verify the assumptions.
On completion of the course, students:
Have a good understanding of the general framework of regression models;
Are able to formulate questions about the general framework of regression models;
Are able to fit the various regression models using statistical software;
Can identify an appropriate regression model, given a data set and a description of the research design;
Can assess whether a model fits the data well or not;
Can interpret the estimated parameters from the regression models.
For the timetable of this course please refer to MyTimetable
NOTE As of the academic year 2021-2022, you must register for all courses in uSis.
You do this twice a year: once for the courses you want to take in semester 1 and once for the courses you want to take in semester 2.
Registration for courses in the first semester is possible from early August. Registration for courses in the first semester is possible from December. The exact date on which the registration starts will be published on the website of the Student Service Center (SSC)
By registering for a course you are also automatically registered for the Brightspace module. Anyone who is not registered for a course therefore does not have access to the Brightspace module and cannot participate in the first sit of the exam of that course.
Also read the complete registration procedure
Mode of instruction
We will meet once a week for 4 hours to discuss about one or two chapters from the book and the exercises. The meetings will be guided by the questions that students hand in.
The final grade is the average of two exam grades, a midway exam and a final exam. The student is expected to actively participate, which is measured by handing in questions about the working material of the week.
The Institute of Psychology follows the policy of the Faculty of Social and Behavioural Sciences to systematically check student papers for plagiarism with the help of software. Disciplinary measures will be taken when fraud is detected. Students are expected to be familiar with and understand the implications of this fraud policy.
Roback, P, and Legler, J. (2021). Beyond Multiple Linear Regression: Applied generalized linear models and multilevel models in R. CRC press.
Note that the book is also freely available online, via https://bookdown.org/roback/bookdown-BeyondMLR/ and the R-code and data sets for all examples in the book can be found on https://github.com/proback/BeyondMLR