I am using glm.cluster from miceadds package to develop a cross-sectional logit model clustered by "states". I get the exact same odds ratios in the clustered model as with an unclustered model. However the standard errors change between two models, and hence the CIs. I want to develop the following for the clustered model: 1) an average marginal effect curve for one of the continuous predictors stratified by gender; and a predicted probability model for the same continuous predictor stratified by gender. I have used the margins package to do that for unclustered model. It seems this package does not support glm.cluster. Please help. I am new to R.
Asked
Active
Viewed 100 times
