Assistant Professor, Health and Kinesiology University of Utah Salt Lake City, Utah
This combination lecture and hands-on workshop will introduce clinicians and researchers to longitudinal model building using linear mixed effects regression (LMER) in the R statistical environment. Rehabilitation outcomes are better suited to multi-level longitudinal modeling than to pre-post regression or repeated measures analysis of variance (RM-ANOVA). LMER has real advantages over RM-ANOVA in terms of flexibility and statistical power. Following the course, learners will be able to appraise existing data for suitability to LMER, discuss the relevance of such models with statistical consultants, and plan for data collection in future projects which are suited to modeling outcomes longitudinally.
Identify circumstances in which longitudinal mixed-effects regression (LMER) models are appropriate
Discuss measurement issues in rehabilitation related to LMER
Describe a process for examining longitudinal data and determining a best-fit linear trajectory
Describe a process for evaluating covariate associations with trajectory parameters
Build longitudinal linear two-level models of time nested within persons using the R statistical environment