Table 2.

Comparison among the optimal robust ROC curve method (OPT-ROC), logistic regression with backward selection (Mul LOG-REG, Dom LOG-REG, and Rec LOG-REG), and classification tree (CLA-TREE)

Cases:Controls	250:250			500:500			1000:1000
	BIAS	SD	MSE	BIAS	SD	MSE	BIAS	SD	MSE
OPT-ROC	0.0158	0.0338	0.00139	0.0010	0.0205	0.00042	0.0007	0.0154	0.00024
Mul LOG-REG	0.0418	0.0250	0.00237	0.0210	0.0170	0.00073	0.0119	0.0122	0.00029
Dom LOG-REG	0.0433	0.0247	0.00249	0.0190	0.0174	0.00067	0.0123	0.0121	0.00030
Rec LOG-REG	−0.0546	0.0242	0.00357	−0.0717	0.0176	0.00546	−0.1046	0.0108	0.01106
CLA-TREE	−0.0195	0.0362	0.00169	−0.0229	0.0220	0.00101	−0.0225	0.0153	0.00074