Course discription and topics
Mixed models are used to analyse correlated observations. Correlated observationsobservations can occur, for instance, when subjects are clustered within neighbourhoods, patients are clustered within hospitals, students are clustered within schools, etc. Besides this, correlated observations also occur in longitudinal studies where the repeated measurements over time are clustered for each individual. Mixed model analysis provides a very elegant and powerful tool to deal with this clustering, i.e. to deal with correlated observations.

This four-day course will explain the basic concepts of mixed models. It is an applied course, so the emphasis lies on the interpretation of the results from the mixed model analyses and not on the mathematical background. The course centres on the two most important applications of mixed models – multilevel analysis and longitudinal data analysis. Lectures are given in the morning and in the afternoon a computer practical is given using the statistical programs STATA, SPSS and MLwiN.

The following topics are discussed during the course:
• Basic principles of mixed model analyses
• Example of a mixed model analysis with a two-level structure
• Example of a mixed model with a three-level structure
• Logistic mixed model analysis
• The use of mixed model analyses in longitudinal studies
• Other methods of analysing longitudinal data
• Comparison between the different methods of analysing longitudinal data
• Alternative models for analysing longitudinal data
• Sample size calculations
• Software for mixed model analysis

Learning objectives
• The student understands the basic principles of mixed model analysis
• The student understands the role of mixed model analysis in analysing longitudinal data
• The student understands the differences between different methods of analysing longitudinal data
• The student understands how alternative models can be used to answer specific research questions
• The student is capable of performing a mixed model analysis using various software programs.
• The student understands the output of a mixed model analysis and is capable of interpreting this output.