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

Table 3 Estimates of the linear regression coefficient (β₁) 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)	0.498
Complete Case Analysis	0.339
Multiple Imputation under MAR	0.339	0.340	0.339	0.339	0.339	0.339
Sensitivity Analysis under MNAR	0.440	0.470	0.532	0.557	0.611	0.633

Note: The empirical Monte Carlo standard errors were all around 0.004 for MI and 0.005 for sensitivity analysis

ISSN: 1471-2288