Date:         Thu, 17 Aug 2006 19:38:35 -0400
Reply-To:     Roger Johnson <roger_johnson555@YAHOO.COM>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Roger Johnson <roger_johnson555@YAHOO.COM>
Subject:      How to fit logistic regression model with muliple responses

I know how to use PROC GLM to fit MANOVA. However, right now I receive a
case which contains multiple responses. Those responses are not time-
variated, but some different indicators from each participant. I think GEE
it not a good way to deal with this case because this is not truely
repeated measurement. I googled some papers about this topic, but most of
them using the term "multivariate logistic regression model" do not use
multiple responses, but multipe dependent variables. PROC GENMOD can't use
multiple responses in its model statement. Does any expert have good ideas
about how to fit this kind of model?