Hello,
I have a multilevel logistic regression model, and I'm trying to do cross-level interaction with the dummy variable "men" (level 1) and the continuous variable "gini_average" (level 2).
I use Stata 13 so my command contains "xi" (the advice of a member from this forum).

This is the command:
Code:
xi: xtmelogit racist i.men i.academic_education age i.religion2 religiosity_level  i.difficult_on_present_income i.ethnic_minority gini_average GDP_average i.men##c.gini_average || cntry_numeric: i.men , var mle or
After I run the command I get an error message:
interactions not allowed
and if I put the interaction variables in different order like this:
Code:
xi: xtmelogit racist i.men i.academic_education age i.religion2 religiosity_level  i.difficult_on_present_income i.ethnic_minority gini_average GDP_average c.gini_average##i.men || cntry_numeric: i.men , var mle or
I get this error message:
must use R. when specifying factor variables in random-effects equations
what is the right way to write the interaction variables? and how can I solve this problem?

I want to try and explain the variance in the men variable (level 1) slope by Gini variable (level 2).

I would most appreciate your help.
Thank you,
Shir