Table 1.

Simulation results for case 1: Correct model specification

Variance = 2.45	LOD = 0.2 (16% under)			LOD = 0.25 (20% under)			LOD = 0.45 (30% under)			LOD = 0.95 (50% under)
	β1	SE	MC SE	β1	SE	MC SE	β1	SE	MC SE	β1	SE	MC SE
True exposure	0.95	0.06	0.06	0.95	0.06	0.06	0.95	0.06	0.06	0.95	0.06	0.06
Maximum likelihood	0.95	0.05	0.05	0.95	0.05	0.05	0.95	0.05	0.05	0.95	0.06	0.06
Multiple imputation	0.95	0.06	0.06	0.95	0.06	0.06	0.93	0.06	0.06	0.82	0.07	0.06
Cox regression	0.84	0.05	0.05	0.85	0.05	0.05	0.89	0.05	0.06	1.02	0.06	0.07
Complete case analysis	0.95	0.06	0.06	0.95	0.07	0.07	0.95	0.08	0.08	0.96	0.11	0.11
Fill-in LOD/√2	0.99	0.06	0.06	1.01	0.06	0.06	1.106	0.07	0.07	1.33	0.09	0.09
Missing indicator	0.95	0.06	0.06	0.95	0.07	0.07	0.95	0.08	0.08	0.96	0.11	0.11
Variance = 1	LOD = 0.2 (6% under)			LOD = 0.25 (8% under)			LOD = 0.45 (20% under)			LOD = 0.95 (47% under)
	β1	SE	MC SE	β1	SE	MC SE	β1	SE	MC SE	β1	SE	MC SE
True exposure	0.95	0.05	0.05	0.95	0.05	0.05	0.95	0.05	0.05	0.95	0.05	0.05
Maximum likelihood	0.95	0.05	0.05	0.95	0.05	0.05	0.95	0.05	0.05	0.95	0.05	0.05
Multiple imputation	0.95	0.05	0.05	0.94	0.05	0.05	0.91	0.05	0.05	0.81	0.06	0.05
Cox regression	1.25	0.05	0.06	1.26	0.05	0.06	1.31	0.05	0.06	1.45	0.06	0.06
Complete case analysis	0.95	0.05	0.05	0.95	0.06	0.06	0.96	0.07	0.06	0.96	0.09	0.09
Fill-in LOD/√2	1.05	0.05	0.05	1.07	0.05	0.05	1.17	0.06	0.06	1.36	0.07	0.08
Missing indicator	0.95	0.05	0.05	0.95	0.06	0.06	0.96	0.07	0.06	0.96	0.09	0.09

Coefficients represent the average β₁ over the 1,000 datasets, the SE corresponds to the average SE over the 1,000 datasets, and MC SE corresponds to the standard deviation of β₁ over the 1,000 datasets.

MC SE indicates Monte Carlo standard error; SE, standard error.