TABLE 4.

Simulation Study: Bias, MSE, and Nominal 95% NCP under Low Separation Linear Model

	β0			β1			β2			α1			α2
	Bias	MSE	NCP (%)	Bias	MSE	NCP (%)	Bias	MSE	NCP (%)	Bias	MSE	NCP (%)	Bias	MSE	NCP (%)
n = 500
Two-stage	0.038	0.012	9.0	0.000	0.001	5.5	−0.001	0.008	3.0	0.015	0.018	10	−0.13	0.047	20
Imputation	0.039	0.011	9.0	0.000	0.001	5.0	−0.001	0.008	4.0	0.019	0.018	10	−0.16	0.052	27
EM	0.016	0.010	8.0	0.000	0.001	5.0	0.000	0.008	3.5	0.016	0.022	15	−0.067	0.042	16
Bayesian	0.002	0.008	2.0	0.001	0.001	2.0	0.000	0.007	1.0	0.013	0.017	8.0	0.003	0.018	4.0
n = 200
Two-stage	0.069	0.038	14	0.005	0.002	6.0	0.003	0.019	4.0	−0.003	0.055	11	−0.24	0.15	25
Imputation	0.065	0.036	13	0.005	0.002	6.0	0.003	0.019	5.0	0.001	0.055	11	−0.24	0.15	26
EM	0.036	0.041	16	0.005	0.002	5.0	0.004	0.020	4.5	0.011	0.074	17	−0.19	0.16	28
Bayesian	0.006	0.023	7.5	0.006	0.002	6.5	0.001	0.019	4.5	0.021	0.048	6.0	−0.038	0.067	7.0
n = 100
Two-stage	0.087	0.068	10	0.000	0.004	6.0	0.001	0.043	6.0	0.062	0.13	14	−0.35	0.26	34
Imputation	0.094	0.072	12	0.001	0.004	6.0	0.001	0.043	6.0	0.058	0.13	14	−0.37	0.29	34
EM	0.075	0.084	14	0.000	0.004	6.0	−0.001	0.047	7.0	0.057	0.14	18	−0.30	0.27	28
Bayesian	0.033	0.075	11	0.008	0.004	5.5	0.011	0.045	7.0	0.060	0.16	16	−0.16	0.24	14

MSE indicates mean square error; NCP, noncoverage probability.