Hi,

I am trying to estimate the group residuals (RE) of the following simple model:

HTML Code:
Y_{ij} = f( X_{ij}b + a_{j} + e_{ij})
Where the Y is a categorical ordered variable.

May idea is running a ordered logit re regression, estimating the predicted posterior means (a_{j}) and then use simulations to get
HTML Code:
E[a_{j} | Y_{ij}, X_{ij}b].
Any idea on how to do that?

Thanks,
Luis