https://communities.sas.com/t5/SAS-Health-and-Life-Sciences/dependent-working-correlation-structure/m-p/5003#M423 How can I get the choice of dependent working correlation structure (specified by TYPE= in the REPEATED statement) for multinomial models? Tue, 09 Oct 2007 11:04:22 GMT deleted_user 2007-10-09T11:04:22Z https://communities.sas.com/t5/SAS-Health-and-Life-Sciences/dependent-working-correlation-structure/m-p/5003#M423 How can I get the choice of dependent working correlation structure (specified by TYPE= in the REPEATED statement) for multinomial models? Tue, 09 Oct 2007 11:04:22 GMT https://communities.sas.com/t5/SAS-Health-and-Life-Sciences/dependent-working-correlation-structure/m-p/5003#M423 deleted_user 2007-10-09T11:04:22Z https://communities.sas.com/t5/SAS-Health-and-Life-Sciences/dependent-working-correlation-structure/m-p/5004#M424 Are you talking about a GEE model or the "upcoming" GEMNOD mod? I tried searching support.sas.com for your title and came up with<BR /> <BR /> <A href="http://support.sas.com/rnd/app/papers/gee.pdf" target="_blank">http://support.sas.com/rnd/app/papers/gee.pdf</A><BR /> <BR /> It looks like you can get a fair ways with GEE now. Perhaps we'll see an experimental version off GENMOD with the feature in 9.2. Tue, 09 Oct 2007 18:12:15 GMT https://communities.sas.com/t5/SAS-Health-and-Life-Sciences/dependent-working-correlation-structure/m-p/5004#M424 Doc_Duke 2007-10-09T18:12:15Z