In the present article we illustrate a method for simultaneously examining between-leader (between-person) and within-leader (within-person) relationships in group therapy research. To do so, we introduce Hoffman and Stawski’s (2009); Curran and Bauer’s (2011), and Wang and Maxwell’s (2015) method for partitioning longitudinal data into between-person and within-person variance components, review the existing psychotherapy research that has used this method, and provide an example of the application of this method to group therapy data. Specifically, using an archival data set, we model between-leader and within-leader dominance and empathy ratings as predictors of group engagement. Additionally, we model time and how time interacts with the within- and between-leader effects, to examine the relationship between within-leader and between-leader empathy and dominance and group climate over time. We conclude by discussing the strengths and limitations of variance partitioning as well as future application in group therapy research. It is our hope that the present article familiarizes group psychotherapy researchers with an effective method for examining multilevel and complex group phenomenon.