Research Article

Evaluation of Four Multiple Imputation Methods for Handling Missing Binary Outcome Data in the Presence of an Interaction between a Dummy and a Continuous Variable

Table 5

Comparison of the performance of imputation methods based on the empirical SE of regression coefficients over 1000 simulations.

VariableMethodMARMCAR
10%20%30%40%50%10%20%30%40%50%

(BMIreal)MCR0.0690.1020.1350.1600.2070.0720.1020.1270.1550.181
MI0.0630.0890.1220.1410.1870.0650.0930.1170.1490.172
MRF0.0600.0820.1010.1100.1240.0640.0830.0950.1110.122
MS0.0790.0940.1090.1100.1130.0780.0940.1000.1070.110
(Betotal)MCR0.0550.0770.1000.1200.1460.0540.0730.0930.1140.136
MI0.0500.0670.0880.1030.1300.0500.0680.0830.1050.132
MRF0.0500.0620.0720.0800.0930.0490.0610.0710.0820.092
MS0.0580.0680.0730.0780.0790.0580.0680.0720.0780.076
(PACS4)MCR0.0550.0740.0930.1160.1400.0530.0710.0880.1060.127
MI0.0510.0660.0830.1050.1220.0480.0670.0790.0950.121
MRF0.0510.0600.0690.0790.0870.0470.0600.0670.0740.083
MS0.0590.0670.0690.0740.0700.0560.0630.0690.0710.069
(PSPS)MCR0.0550.0770.0940.1130.1410.0510.0700.0870.1100.126
MI0.0480.0680.0820.1020.1220.0460.0640.0790.0980.118
MRF0.0490.0610.0690.0790.0870.0450.0580.0660.0770.081
MS0.0590.0660.0670.0710.0690.0540.0650.0690.0710.069
(Gender)MCR0.0770.1130.1450.1720.1940.0830.1140.1520.1720.205
MI0.0720.1050.1360.1680.2080.0730.1080.1430.1710.217
MRF0.0700.0950.1110.1220.1350.0710.0940.1180.1200.135
MS0.0890.1110.1250.1340.1260.0910.1140.1220.1210.134
(BMIreal ∗Gender)MCR0.0960.1390.1820.2160.2480.1020.1430.1750.2170.242
MI0.0860.1260.1720.2140.2650.0890.1310.1730.2190.258
MRF0.0850.1150.1320.1430.1470.0880.1160.1310.1430.143
MS0.1270.1440.1560.1510.1490.1230.1420.1480.1420.143