ROC Curve Analyses

[hoc]

Would ROC curve analyses work with 2 categorical variables measured at 2 time points (second time point is used as outcome measure) and a continuous variable (1-5) as a moderator with an N=36?