Re: st: Multinomial logistic regression with cross-classified data structure

["Michael I. Lichter" <MLichter@Buffalo.EDU>]

Vivien,

If you have multilevel data and want to fit a cross-classified random 
effects mlogit model, you might be able to do it with the Stata package 
gllamm (see http://www.stata.com/statalist/archive/2004-04/msg00762.html).

If you are just saying that there might be interactions between 
neighborhood characteristics and workplace characteristics, you can 
certainly use -mlogit- and use factor variables (in Stata 11) or -xi- 
(in earlier versions of Stata) to help you specify the interactions, and 
you can also specify the cluster() option if you need to take clustering 
at the neighborhood or workplace level into account.

Does that help?

Michael

On 4/4/2010 10:59 PM, Vivien W. Chen wrote:
> Dear all,
>
> I want to fit a multinomial logistic regression model in Stata. The
> dependent variable is a categorical variable that indicates 4
> occupations. The control covariates include factors of individual
> backgrounds, workplace features and characteristics of residential
> neighborhood. The selection among these 4 occupations might be related
> to workplace and neighborhood simultaneously. So I consider to fit a
> multinomial logistic cross-classified model. Could I fit this kind of
> model in Stata? If yet, what syntax would you suggest? Any idea will
> be deeply appreciated. Thanks!
>
> Best,
> Vivien Chen
> *
> *   For searches and help try:
> *   http://www.stata.com/help.cgi?search
> *   http://www.stata.com/support/statalist/faq
> *   http://www.ats.ucla.edu/stat/stata/
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/