Hallo! Does anybody know a way to calculate the sample size for comparing AUC of ROC curves against 'by chance' with AUC=0.5 (and/or against anothe AUC)? Thanks! Karl