I am trying to use a ROC curve for the first time using SPSS. I just want to verify that I am putting the correct type of variables in the right spots. I am looking at the continuous vars of body mass index (BMI) and waist circumference as the test vars with a dichotomous var (at risk lipids, 0 or 1) for the state var. Can someone tell me if that is correct? Does the test var have to be continuous and state var have to be dichotomous? Also, is there a way to then also get SPSS to calculate the liklihood ratios in order to determine cut points? Finally, if anyone could point me to good references explaining more on ROC curves, that would be great! Thanks.

Maria