We are using the mixed command in our analysis with some post estimations, but we also need to use weights. When we add weights, all standard errors become robust, and some post-estimation options become unavailable.
Here is the code we use without weights (this works perfectly fine).
mixed dv time || school_ed: time || student_id:
predict time_eb school_eb student_eb, reffects reses(time_eb_se school_eb_se student_eb_se)
The moment we add weights, we can no longer generate SE using "reses"
Does anybody know another option to produce SE for this type of model?
Here is the code we use without weights (this works perfectly fine).
mixed dv time || school_ed: time || student_id:
predict time_eb school_eb student_eb, reffects reses(time_eb_se school_eb_se student_eb_se)
The moment we add weights, we can no longer generate SE using "reses"
Does anybody know another option to produce SE for this type of model?
