Mixed Models for clustered data (January 2019)

Course Title    

Mixed Models for Clustered Data:

Aim

Applying and understanding mixed models for continuous, dichotomous and count outcome variables in relation to explanatory variables for cross-sectional and longitudinal data.

Organization

Within two weeks, five sessions of four hours will be provided. Per session there will be 2.5 hours of lectures and 1.5 hours performing exercises. In total 20 contact hours are provided.

Intended for

Researchers, PhD students, and master students interested in statistical models for analysing data with cluster structures, for example in longitudinal or multilevel data.

Content

The family of mixed models is a very useful statistical toolbox for the analysis of clustered data (e.g. members of the same family, patients within one hospital, repeated measures of individuals, etc.).
This course will provide knowledge on the concepts of linear mixed models and generalized linear mixed models for cross-sectional and longitudinal data for numerical or dichotomous outcome variables.

The course will explain and practice with general estimating equations (GEE), maximum likelihood estimation (MLE), and restricted maximum likelihood estimation (REML).

Using many practical examples, the model specification, analysis and interpretation of the results will be explained in this course. In the accompanying workshops, the course participants have the possibility of guided training using SPSS.

Th e main topics of the course will be

§ Analysis of variance models

§ Linear mixed models

§ Generalized linear mixed models

§ Model specification approaches

Applicationform Mixed Models