Dear all,
i hope anyone could help me. my short problem is the following:
I have clustered data (individual level and national level). I know i have
to run a multilevel model and i did it.
Anyway I want to produce OR's country-specific, so, one regression model for
each country.
In order to make fair comparison among countries, does it mean to inflate
95% CI?
can I use quasi-likelihood estimation?
thanks in advance
|