Analysis of longitudinal data with hierarchical linear models

Background: In clinical research, repeated-measures analysis of variance (RM-ANOVA) is the most commonly used tool for the analysis of multiple occasions. In contrast, multilevel models for longitudinal data make less restrictive assumptions and are more flexible. Objective: flow can multilevel models be used for analyzing longitudinal data in psychotherapy research? Method: Based on a previous article by Keller (2003) and a study on the outcomes of a cognitive-behavioral therapy, problems of classical methods for repeated measures will be discussed and compared to the advantages of multilevel models. Results: Treatment effects, the fitting of discontinuous and non-linear change trajectories and the addition of a further level of analysis will be discussed in detail. Conclusions: Multilevel models provide a powerful framework for studies in the field of clinical research.