Table 4.

Bias, 95% Confidence Interval Coverage, and Mean Squared Error for Simulation Studies^a Under 9 Scenarios for Nondifferential Misclassification

Sensitivity	Specificity	Sample Size	Validation Percentb	Naive			Validation			Direct Maximum Likelihood			Multiple Imputation
				Biasc	Coverd	MSEe	Bias	Cover	MSE	Bias	Cover	MSE	Bias	Cover	MSE
0.9	0.9	1,000	10	24	62	8	−5	96	47	−4	96	33	2	97	27
		1,000	30	24	62	8	−1	95	10	−1	95	6	−1	95	6
		300	30	24	85	13	−5	96	47	−4	96	42	2	97	27
0.6	0.9	1,000	10	35	40	15	−5	96	47	−5	96	48	−3	96	29
		1,000	30	35	40	15	−1	95	10	−1	95	8	−1	95	8
		300	30	35	76	21	−5	96	47	−4	96	48	−3	96	33
0.3	0.9	1,000	10	51	20	30	−5	96	47	−5	96	54	−4	95	35
		1,000	30	51	20	30	−1	95	10	−1	95	9	−1	95	9
		300	30	51	66	38	−5	96	47	−5	96	58	−3	96	37

Abbreviation: MSE, mean squared error.

^a Results are summarized over 10,000 simulations.

^b Percent of all participants included in the validation subgroup.

^c Bias was defined as 100 times the difference between the average estimated log odds ratio and the true log odds ratio.

^d Confidence interval coverage was calculated as the percentage of simulations in which the estimated 95% Wald-type confidence limits included the true value.

^e MSE was calculated as the sum of the bias squared and the variance.