R-Squared Measures for Multilevel Models
Generates both total- and level-specific R-squared measures from
Rights and Sterba’s (2019) framework of R-squared measures for multilevel
models with random intercepts and/or slopes, which is based on a completely
full decomposition of variance. Additionally generates graphical
representations of these R-squared measures to allow visualizing and
interpreting all measures in the framework together as an integrated set.
This framework subsumes 10 previously-developed R-squared measures for
multilevel models as special cases of 5 measures from the framework, and it
also includes several newly-developed measures. Measures in the framework
can be used to compute R-squared differences when comparing multilevel
models (following procedures in Rights & Sterba (2020) ).