what is the right method and procedure for analyzing non-proportional longitudinal ordinal outcome?

[Tiny_Kane]

Hi there,

I have a statistical question. I am trying to run mixed effects model on repeatedly measured ordinal outcome, but the data looks not meed the proportional odds assumption. Is there any procedure can take care of this issue? Thank you very much in advance. 

[Norman21]

You might want to try a non-parametric approach; a collection of SAS macros can be found here:

http://www.ams.med.uni-goettingen.de/sasmakr-de.shtml

Although the site is in German, the explanatory PDFs can be Google-translated into understandable English.

[Ksharp]

Better post it at Stat forum and calling @StatDave