Table 1.

Empirical bias, empirical and estimated variances, coverages and mean squared errors for the settings with one mismeasured covariate, when

			Ma & Yin method			Naive method			Simex method
		Estimate
200	0.1	Bias	-1.0	3.2	-0.2	4.3	-9.2	0.2	-0.4	1.5	-0.1
		Emp. var.	5.3	13.2	4.1	4.7	9.9	4.0	5.2	12.8	4.1
		Est. var.	5.3	13.0	3.6	4.6	9.8	3.6	5.2	12.4	3.6
		95% cv	94.6	94.8	93.6	93.6	93.4	93.8	94.2	94.2	93.8
		MSE	5.3	13.3	4.1	4.8	10.8	4.0	5.2	12.8	4.1
200	0.2	Bias	-3.0	8.5	-0.5	14.2	-31.5	.7	3.6	-7.6	0.1
		Emp. var.	6.9	22.6	4.4	4.1	7.5	4.0	5.4	14.3	4.2
		Est. var.	7.4	22.7	3.9	4.0	7.4	3.6	5.2	12.7	3.7
		95% cv	96.0	96.8	93.2	88.4	78.8	93.4	93.8	92.4	93.0
		MSE	7.0	23.4	4.4	6.1	17.5	4.0	5.5	14.8	4.2
300	0.1	Bias	-1.1	3.7	-1.4	4.2	-8.4	-1.1	-0.5	2.3	-1.4
		Emp. var.	3.9	9.4	2.6	3.4	7.1	2.5	3.8	9.0	2.6
		Est. var.	3.5	8.7	2.4	3.1	6.5	2.4	3.4	8.3	2.4
		95% cv	95.6	94.4	96.6	93.4	93.2	96.6	95.4	94.2	96.6
		MSE	3.9	9.5	2.6	3.6	7.8	2.5	3.8	9.1	2.6
300	0.2	Bias	-2.1	6.8	-1.8	14.2	-31.0	-0.6	3.6	-6.9	-1.3
		Emp. var.	4.9	14.5	2.7	3.0	5.1	2.5	3.9	9.7	2.6
		Est. var.	4.6	13.9	2.6	2.7	4.9	2.4	3.4	8.4	2.5
		95% cv	96.6	95.8	95.8	85.2	71.6	95.4	93.8	93.4	95.8
		MSE	5.0	14.9	2.7	5.0	14.7	2.5	4.1	10.2	2.6

Emp. var., empirical variance; Est. var., estimated variance; 95% cv, coverage probabilities of 95% confidence intervals computed based on the asymptotic normal distribution; MSE, mean squared error. All numbers were multiplied by 100.