| Strategy | Symbol | Parameter bias | Average SE (as compared to complete data) | Standard error bias | Coverage rate of confidence intervals | Power (as compared to complete data) |
| Complete data | □ |
|
|
|
|
|
| Recall | | Negative bias up to 78% | Up to 33% smaller |
| Can be as small as 0.025 | Loss of power |
| Case deletion | |
|
| Negative bias up to 22% | Can be as small as 0.87 |
|
| Period deletion | |
| Up to 110% larger |
|
| Loss of power |
| Assume event non-occurrence | | Negative bias up to 22% |
|
|
| Loss of power |
| Assume event occurrence | | Negative bias up to 90% | Up to 39% smaller |
| Can be as small as 0 | Loss of power |
| Single imputation | + |
|
| Negative bias up to 23% | Can be as small as 0.87 |
|
| Multiple imputation | |
|
| Positive bias up to 11% |
| Loss of power |