Table 3.

Efficiency of other estimators (using data from the target population or based on the combined-data analysis) relative to the constrained maximum likelihood estimator calculated using data only from the target population. Standard errors of entries do not exceed 3% of the value.

	Population 1					Population 2
	Pop-Specific		Combined-data			Pop-Specific		Combined-data
	EML	PSL	EML	CML	PSL	EML	PSL	EML	CML	PSL
β
β0 = −3.8	0.98	0.94	1.81	1.89	1.77	0.98	1.00	2.24	2.34	2.20
β1 = 1.2	0.99	0.92	1.81	1.87	1.73	0.99	0.99	2.27	2.35	2.18
β2 = 1.5	0.99	0.91	1.80	1.85	1.70	0.99	0.98	2.35	2.42	2.23
ROC(t)
t = .1	0.96	0.95	1.75	1.84	1.76	0.99	1.01	2.16	2.28	2.18
t = .3	0.99	0.98	1.77	1.81	1.77	1.00	1.01	2.25	2.31	2.25
t = .5	0.99	0.98	1.71	1.75	1.71	0.99	0.99	2.33	2.38	2.33
t = .7	0.98	0.95	1.66	1.71	1.62	0.98	0.97	2.33	2.39	2.27
t = .9	0.97	0.92	1.66	1.72	1.59	0.98	0.96	2.31	2.39	2.21
CDFRx(p)
p = (.22, .12)★	0.99	1.01	1.14	1.15	1.17	1.00	1.02	1.28	1.29	1.32
p = (.30, .16)	0.98	0.95	1.31	1.35	1.26	0.99	0.97	1.39	1.43	1.36
p = (.43, .23)	0.96	0.96	1.20	1.24	1.19	0.99	0.98	1.35	1.40	1.34
p = (.56, .35)	0.98	0.96	1.37	1.38	1.35	0.99	0.98	1.44	1.46	1.42
p = (.67, .48)	0.98	1.00	1.24	1.25	1.25	0.99	1.01	1.26	1.28	1.26
Risk(y|x)
y = (−1.1, −1.2)★	1.00	0.96	1.51	1.53	1.47	0.99	0.97	1.98	2.00	1.91
y = (−.24, −.36)	0.99	0.95	1.32	1.32	1.29	1.00	1.00	1.51	1.55	1.49
y = (.33, .20)	0.99	0.92	1.16	1.14	1.13	0.99	0.98	1.35	1.40	1.32
y = (.88, .75)	1.00	0.99	1.08	1.08	1.08	0.99	1.00	1.10	1.12	1.10
y = (1.6, 1.5)	1.00	0.93	1.40	1.45	1.34	0.99	0.98	1.53	1.53	1.50

^★the values separated by commas correspond to population 1 and population 2 respectively.