Table 1.
Average Bias, RMSE, and Correlation Values for MAR.
| Method | Test length | Missing proportion (%) | Bias | RMSE | Correlation |
|---|---|---|---|---|---|
| FIML | 20 | 30 | 0.000 | 0.339 | .916 |
| MICE-CART | 20 | 30 | 0.000 | 0.366 | .902 |
| MICE-RFI2 | 20 | 30 | 0.000 | 0.367 | .900 |
| MICE-CART2 | 20 | 30 | 0.000 | 0.373 | .897 |
| MICE-RFI | 20 | 30 | 0.000 | 0.373 | .896 |
| Zero replacement | 20 | 30 | 0.000 | 0.443 | .854 |
| FIML | 40 | 30 | 0.000 | 0.282 | .951 |
| MICE-CART | 40 | 30 | 0.000 | 0.304 | .942 |
| MICE-RFI2 | 40 | 30 | 0.000 | 0.311 | .940 |
| MICE-CART2 | 40 | 30 | 0.000 | 0.314 | .938 |
| MICE-RFI | 40 | 30 | 0.000 | 0.314 | .938 |
| Zero replacement | 40 | 30 | 0.000 | 0.384 | .908 |
| FIML | 60 | 30 | 0.000 | 0.250 | .964 |
| MICE-CART | 60 | 30 | 0.000 | 0.268 | .958 |
| MICE-RFI2 | 60 | 30 | 0.000 | 0.278 | .955 |
| MICE-RFI | 60 | 30 | 0.000 | 0.280 | .954 |
| MICE-CART2 | 60 | 30 | 0.000 | 0.283 | .953 |
| Zero replacement | 60 | 30 | 0.000 | 0.348 | .929 |
| FIML | 20 | 40 | 0.000 | 0.406 | .877 |
| MICE-CART2 | 20 | 40 | 0.000 | 0.437 | .857 |
| MICE-CART | 20 | 40 | 0.000 | 0.443 | .853 |
| MICE-RFI2 | 20 | 40 | 0.000 | 0.445 | .850 |
| Zero replacement | 20 | 40 | 0.000 | 0.465 | .845 |
| MICE-RFI | 20 | 40 | 0.000 | 0.470 | .830 |
| FIML | 40 | 40 | 0.000 | 0.348 | .924 |
| MICE-CART | 40 | 40 | 0.000 | 0.373 | .912 |
| MICE-CART2 | 40 | 40 | 0.000 | 0.378 | .910 |
| MICE-RFI2 | 40 | 40 | 0.000 | 0.382 | .908 |
| MICE-RFI | 40 | 40 | 0.000 | 0.401 | .898 |
| Zero replacement | 40 | 40 | 0.000 | 0.422 | .891 |
| FIML | 60 | 40 | 0.000 | 0.311 | .943 |
| MICE-CART | 60 | 40 | 0.000 | 0.332 | .935 |
| MICE-CART2 | 60 | 40 | 0.000 | 0.341 | .931 |
| MICE-RFI2 | 60 | 40 | 0.000 | 0.344 | .930 |
| MICE-RFI | 60 | 40 | 0.000 | 0.358 | .924 |
| Zero replacement | 60 | 40 | 0.000 | 0.400 | .907 |
| Zero Replacement | 20 | 70 | 0.000 | 0.585 | .738 |
| FIML | 20 | 70 | 0.000 | 0.658 | .642 |
| MICE-CART2 | 20 | 70 | 0.000 | 0.681 | .607 |
| MICE-RFI2 | 20 | 70 | 0.000 | 0.691 | .572 |
| MICE-CART | 20 | 70 | 0.000 | 0.705 | .580 |
| MICE-RFI | 20 | 70 | 0.000 | 0.747 | .469 |
| Zero replacement | 40 | 70 | 0.000 | 0.543 | .812 |
| FIML | 40 | 70 | 0.000 | 0.579 | .776 |
| MICE-CART2 | 40 | 70 | 0.000 | 0.604 | .749 |
| MICE-CART | 40 | 70 | −0.004 | 0.622 | .733 |
| MICE-RFI2 | 40 | 70 | −0.001 | 0.631 | .726 |
| MICE-RFI | 40 | 70 | −0.001 | 0.698 | .640 |
| Zero replacement | 60 | 70 | 0.000 | 0.523 | .838 |
| FIML | 60 | 70 | 0.000 | 0.539 | .820 |
| MICE-CART2 | 60 | 70 | −0.001 | 0.559 | .803 |
| MICE-CART | 60 | 70 | −0.005 | 0.568 | .796 |
| MICE-RFI2 | 60 | 70 | −0.001 | 0.585 | .787 |
| MICE-RFI | 60 | 70 | −0.001 | 0.660 | .709 |
Note. RMSE = root mean square error; MAR = missing at random; FIML = full-information maximum likelihood; MICE-CART = multiple imputation with chain equations utilizing classification and regression trees; MICE-RFI = multiple imputation with chain equations utilizing random forest imputation.