02-16-2018 09:39 AM - edited 02-16-2018 12:16 PM
I hope that this is the right forum to ask this question. I was a bit in a hurry so I couldn't make sure of it and possible previous responses.
When the response is binary, glimmix does a very good job giving odds ratios of about an interaction in a logistic regression model.
But when the response is ordinal, I get the following error
ERROR: Least-squares means are not available for the multinomial distribution.
when I try a line like this
lsmeans treat*time / oddsratio ilink slicediff=time;
What can I do in such a situation to get the odds ratios in slices?