Re: st: non-proportional odds and multilevel models
Date
Tue, 13 Nov 2012 07:50:31 -0500 (EST)
When the assumption of proportional odds is violated for ordinal dependent variables, you could use generalized ordered logistic regression (gologit2: http://www.nd.edu/~rwilliam/gologit2/gologit2.pdf). Here is another link with some resources: http://www.nd.edu/~rwilliam/gologit2/
I'm not sure whether or how it deals with multilevel models, but perhaps these links describe a solution.
----- Original Message -----
From: "Linda Kampata" <Linda.Kampata@student.uib.no>
To: statalist@hsphsun2.harvard.edu
Sent: Tuesday, November 13, 2012 2:52:48 AM
Subject: st: non-proportional odds and multilevel models
Hi,
Does stata have a model one can use for multilevel modelling for an
ordinal dependant variable which does not hold for non-proportional
assumptions. The non-proportional test results are significant
(<0.001) under the omodel command. I have read on statalist as well as
the Multilevel and Longitudinal Modeling Using Stata book and it does
not spell it out clearly for me to understand.
What options do I have?
Can someone please recommend or help me.
Regards,
Linda
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/faqs/resources/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/faqs/resources/statalist-faq/
* http://www.ats.ucla.edu/stat/stata/
