Table 1.

Predictive Ability of Imbalance Measurements for Power, N=300, X ~ Bimodal, (β̃_tx is treatment effect in equation (2) corresponding to 80% power)

Scenario	Measure	Model p-value	AIC	Hosmer-Lemeshow p-value	D
	t	<2E-16	1167.9	0.675	0.063
βx= 0.63β̃tx	WRS	8.34E-16	1178.8	0.499	0.054
βtx = 0	KS	3.55E-12	1187.1	0.016	0.047
	sAUC	2.36E-16	1168.7	0.056	0.062
	t	<2E-16	5401.4	0.011	0.051
βx = −0.6β̃tx	WRS	<2E-16	5429.1	0.744	0.047
βtx = β̃tx	KS	<2E-16	5465.2	0.028	0.040
	sAUC	<2E-16	5410.1	0.006	0.050
	t	<2E-16	1066.5	0.400	0.102
βx = β̃tx	WRS	<2E-16	1090.8	0.861	0.082
βtx = 0	KS	1.16E-15	1097.7	0.394	0.076
	sAUC	<2E-16	1077.0	0.054	0.094
	t	<2E-16	5928.3	0.060	0.109
βx = −β̃tx	WRS	<2E-16	6021.9	0.509	0.095
βtx = β̃tx	KS	<2E-16	6158.0	<0.001	0.074
	sAUC	<2E-16	5948.1	0.063	0.106
	t	<2E-16	977.5	0.679	0.229
βx = 1.5β̃tx	WRS	<2E-16	1019.6	0.924	0.196
βtx = 0	KS	<2E-16	1052.4	0.003	0.170
	sAUC	<2E-16	984.5	0.862	0.224
	t	<2E-16	5503.1	0.619	0.194
βx = −1.5β̃tx	WRS	<2E-16	5722.5	0.843	0.161
βtx = β̃tx	KS	<2E-16	59850	<0.001	0.123
	sAUC	<2E-16	5576.4	<0.001	0.183