# What are the assumptions of ordinal mixed effects logistic regression?

Specifically, what are the assumptions of ordinal mixed effect logistic regression performed with the "ordinal" package in R? I just got knobbled by a reviewer because these weren't clearly stated in a paper. The same reviewer asked me how I could "account for repeated measures". How would you explain the ordinal(ity) assumptions and the ME/ hierarchical model in a relatively parsimonious way?

-
If you fit a mixed effects model then how have you not "accounted for repeated measures"? – Macro Jun 18 '12 at 22:33
I know that and u know that, my question is how to make him/her understand that. Do I just go around handing out copies of gelman & hill?? – rosser Jun 19 '12 at 23:22
You could write out, in your response, how the random effects model correlation between observations made on the same person. In particular, you could write out the value of the intraclass correlation coefficient: $$\frac{ \sigma^2 }{ \sigma^2 + \pi^2/3}$$ on the latent logistic scale, where $\sigma^2$ is the random effect variance. This shows that observations made on the same individual are not independent. – Macro Jun 19 '12 at 23:29
The reviewer is not a ststistician and will definitely not get that. – rosser Jun 21 '12 at 11:20