Hello,

This question refers to Daniel Klein's excellent mimrgns program. I am attempting to obtain predicted probabilities from a GEE model with a logit link / binomial family. I would use a regular logit, but there is some clustering in the sample. The data is multiply imputed. I have read Daniel's useful helpfile, but continue to get an error. This is the code I am attempting to run.

Code:
mi est, eform: xtgee y i.x1##c.x2##c.x3, family(binomial) link(logit) corr(exch)
mimrgns, dydx(x2) at(x1=(1) x3=(-5(.5)1.5) predict(pr) post
I get an error that says:
Code:
option pr not allowed
an error occurred when mi estimate executed mimrgns_estimate on m=1

I am able to run this without the "predict(pr)" and get the linear predictions, as is the default. However, because the model is nonlinear, I'd like to be able to see the predicted probabilities. I would appreciate any advice!

Robbie Dembo