I am using lda to predict case (success / no success) using about 5 discrete variables. If z is my model, I can get the coefficients of lda using z[[4]] or I can just output the full summary for the model by typing 'z' into the command line, but I don't see where it reports the intercept. In other words, if I want to calculated the predicted value myself instead of using 'predict', my numbers are off by the value of the intercept term. I can use this knowledge to back into the intercept term, but I am surprised its not reported in the model output. Second question, I am also comparing lda to glm with family binomial. Are there any other methods I should be looking at to linear regression on case type variables from discrete variables? Thanks in advance for any help, Roger [[alternative HTML version deleted]]