Analysis of correlated data: Mixed linear models

• Construct and apply factor structure diagrams for complex experimental designs.
• Perform statistical analyses based on the theory of mixed linear models using the statistical software R.
• Explain the theory of mixed linear models.
• Distinguish between random and fixed effects.
• Compare and distinguish between different relevant models and statistical methods.
• Perform, explain, and discuss statistical analyses of data from unbalanced block and split-plot experiments.
• Perform, explain, and discuss statistical analyses of data from unbalanced longitudinal studies.
• Perform, explain, and discuss hierarchical statistical analyses including analyses based on variance component models and regression models with varying coefficients.
• Perform, explain, and discuss statistical analyses for repeated measurements including identification of various correlation structures.
• Combine and modify the various techniques.

The course will cover basic theory and application of mixed linear models. This includes fixed and random effects but also more general correlation structures relevant to the analysis of repeated measurements/longitudinal data.

In short: The course gives theoretical and practical tools for performing statistical analysis of data structures which do not satisfy the independence assumptions made in introductory statistics courses.

The statistical software R will be used.

02402/02403/02323/02411/02418, In addition to an introductory statistics course (e.g. 02402) it is recommended to have at least two relevant statistics courses. The two most relevant courses are 02411 and 02418.

All course material will be available online. There will weekly be two hours lecturing and two hours for exercises including computing exercises, mostly practical data analysis challenges. The format will depend on the number of students participating, but student involvement is to be expected.