# Estimating fixed effects for ordinal regression

I am running a model when the response variable (behaviour) is an ordinal factor (levels: 0,1,2,3,4) and I am testing the relationship with two fixed effect factors (year: A,B,C,D; colony; X1,X2).

I want to extract the estimates of the effect of year and colony on behavoiur. In a normal GLM, I understand that this is simple, but for an ordinal model, I am not sure how to incorporate the cut points for the ordinal levels with the fixed effect estimates. From reading about ordinal models, it seems that maybe you get a matrix of probabilities such that you get for colony a 2 by 5 matrix and for year a 4 by 5. Am I understanding this correctly? And if so, it does not seem possible to extract any kind of residual value for behaviour controlling for fixed effects?

Thanks

Sam

-
add comment