Thanks so much for such a wonderful package -- it makes these analyses so much better.
Looking at the documentation for analyzing outcomes in MatchThem, svglm() with quasibinomial is recommended for multiply imputed data with a binomial outcome. But I also noted that the WeightIt package has the function glm_weightit() with provides robust variances and accounts for the creation of the weights. This isn't mentioned in the vignette -- is there a situation where one would be preferred over the other? Either way, perhaps this could be addressed in a future version of the vignette?
Thanks so much for such a wonderful package -- it makes these analyses so much better.
Looking at the documentation for analyzing outcomes in MatchThem, svglm() with quasibinomial is recommended for multiply imputed data with a binomial outcome. But I also noted that the WeightIt package has the function glm_weightit() with provides robust variances and accounts for the creation of the weights. This isn't mentioned in the vignette -- is there a situation where one would be preferred over the other? Either way, perhaps this could be addressed in a future version of the vignette?