Dear list members, I am using a cross validation of a generalised linear model (glm). The cv.glm function (from boot package) returns an error as so-called ?delta? value. I would like to get to a (cross-validated) squared q, because I want to directly compare it to the squared correlation coefficient r. I tried to find an an equation for the raw and/or adjusted cross-validation estimate of prediction error. But actually I did not get a clear answer up to now. So it would be very kind, if someone could help me with this. Thanks in advance, Markus