Hi all - I would really appreciate some thoughts/advice. I'm working with survey data containing individual responses nested within a number of institutions. My DV is an unordered categorical variable and, given the hierarchical nature of the data, I'd like to use a multilevel approach. I generally use Stata, and I understand this is possible using the gsem commands within that program (as in Stata example 41g).

That said, I need to address missing data in the data set. I am familiar with multiple imputation, but from what I gather I can't use Stata's -mi- command syntax in concert with -gsem-. Thus, my question... Is there a way to use MI then estimate a multilevel multinomial logistic regression model in either Stata (that I haven't found) or another program? I have easy access to SPSS and SAS, though I'm not terribly familiar with either one.

Thanks in advance for your help, and please let me know if my question isn't clear or more information would be helpful.