Table 3.

Simulation results: Comparison of the ITR to the non-personalized universal rule. The proportion of rejecting the null that the ITR has the same benefit as the universal rule^∗ are reported for the overall sample and by subgroups.

	Setting 1. Four region means = (1, 0.5, −1, −0.5).
	Overall	W < −0.5	W ∈ [−0.5, 0.5]	W > 0.5
N = 800
PM	0.22	0	0.09	0.33
Q-learning	0.37	0.02	0.20	0.40
O-learning	0.39	0.02	0.20	0.43
ABLO	0.86	0.07	0.47	0.78
N = 1600
PM	0.76	0.02	0.38	0.83
Q-learning	0.92	0.05	0.59	0.90
O-learning	0.95	0.06	0.67	0.94
ABLO	0.99	0.08	0.79	0.98
	Setting 2. Four region means = (1, 0.3, −1, −0.3).
N = 800
PM	0.18	0.01	0.07	0.27
Q-learning	0.35	0.03	0.17	0.37
O-learning	0.31	0.03	0.17	0.35
ABLO	0.82	0.07	0.43	0.74
N = 1600
PM	0.72	0.03	0.38	0.75
Q-learning	0.88	0.05	0.57	0.86
O-learning	0.90	0.07	0.59	0.86
ABLO	0.99	0.12	0.77	0.97

^*For Setting 1, the mean difference (sd) of the universal rule is 0.09(0.08) for N = 800 and 0.07(0.05) for N = 1600.

For Setting 2, the mean difference (sd) of the universal rule is 0.11(0.08) for N = 800 and 0.08(0.05) for N = 1600.