this is for the person who asked me about prediction confidence intervals in a GLM because I lost your email. Below follows a simple example in CAR and the variance covariance of the beta coefficients is in the summary. So, I think, given that output, it should be pretty straightforward to do what you want. I also bet that code for what you want to do is in some function in John's effects package but I haven't looked at the code. If it's not possible to find it in John's code, ( I think prediction intervals have an extra term in the formula for the CI because of the error term variance ) that formula should be in any decent regression text book. Good luck. #======================================================================================= attach(Mroz) model.mroz<-glm(lfp ~ k5 + k618 + age + wc + hc + lwg + inc, family=binomial) print(model.mroz) modsumm<-summary(model.mroz) print(modsumm) print(str(modsumm))