I conducted a series of generalized estimating equation (GEE) models in R using the gee and geepack packages. These models have data from two waves. We are looking to see if a Time 1 variable predicts change in a variable measured at both Time 1 and Time 2. Several covariates are included in the model. Some of these models have an additional Time 1 predictor included in an interaction term.
We reported the regression coefficients, but a reviewer is requesting effect sizes specifically. After reading about this for a while, I don't see any consensus on the best measure of effect size for GEE. My understanding was regression coefficients are a measure of the effect. Do you all know alternate options for GEE effect sizes? And if you have resources for calculating those in R, I would appreciate it. Thanks!
