Dear Brian, I looked on the internet for this warning message "Warning message: 'newdata' had X rows but variable(s) found have Y rows" that happens when using some kind of fitting and prediting. I still can't figure out what is going wrong by reading the documentation or the posts in de forum. So I made a small example pos = rep(1,times=10) neg = rep(0,times=10) y = c(pos,neg) x = c(pos*30,neg*2) result1 <- glm(y~x, family=binomial) p = predict.glm(result1,as.data.frame(x),type='respons') This works fine. But when I change the argument of the last row into as.data.frame(x[1:10]) (or some new data) I get that warning again. Why? Can you help me out? Best regards Herman van Haagen (Netherlands) [[alternative HTML version deleted]]