New Effect Size Measures for Random Slope Effect in Multilevel Models

• In Event: Effect Sizes, Multilevel Models, and Causal inference

Abstract

Random slopes can be estimated within multilevel models in order to ascertain if, and how much, the relationship between a predictor and outcome varies by group membership. Despite the importance of reporting effect size measures for quantitative studies, it is unclear what measure applied researchers should report to provide evidence of the scope of randomly varying slopes. The purpose of the present study is to suggest two new measures representing variance explained and a standardized version of the variance of varying slopes, respectively. These measures are evaluated through a simulation study and results indicate that they are both suitable for reporting. Implications and future research are described.

Author

• Julie Ann Lorah, Indiana University - Bloomington