Multilevel models

Advantages and problems with dependent observations. Summary measures (e.g. Area under the curve [AUC], slope coefficient, min/max values). Covariance and correlation. Variance components. Linear mixed effects models. Logistic mixed effects models. Use of relevant software (Stata)

After completing the course, the student should be able to:

• Understand the nature of dependency in clustered measurements, and how this dependency alters the approach to statistical analysis and modeling; as well as the consequences of ignoring it;
• Identify clusters and potential dependency by inspecting the design and/or viewing the resulting data set;
• Understand the principles of experimental design in which experimental factors vary both between and within clusters;
• Perform simple, descriptive analyses such as obtaining sample covariance and correlation, and corresponding scatter plots to illustrate key features of data with clustered observations;
• Estimate and interpret variance components;
• Estimate and interpret linear mixed regression models and logistic mixed effects models with random intercept;
• Estimate and interpret linear mixed effects models with random intercept and random slopes for covariates;
• Use of relevant software

Forelesninger og øvinger med veiledning