Evaluation of a weighting approach for performing sensitivity analysis after multiple imputation

Table 4 Estimates of the marginal proportion of the binary outcome variable under four analysis methods (n = 100, δ = 1); True value = 0.5

	Number of imputations (m)
	5	10	50	100	500	1000
Full dataset (before deletion)	0.497
Complete Case Analysis	0.639
Multiple Imputation under MAR	0.608	0.608	0.608	0.608	0.608	0.608
Sensitivity Analysis under MNAR	0.545	0.526	0.493	0.482	0.460	0.452

Note: The empirical Monte Carlo standard errors were all around 0.002 for MI and for sensitivity analysis

ISSN: 1471-2288