The relative efficiency (RE) of using the finite m imputation estimator, rather than using an infinite number for the fully efficient imputation, in units of variance, is approximately a function of m and (Rubin 1987, p. 114):
where mis the number of imputations and is the fraction of missing information.
Table 75.7 shows relative efficiencies with different values of m and .
Table 75.7: Relative Efficiencies
|
||||||
---|---|---|---|---|---|---|
m |
10% |
20% |
30% |
50% |
70% |
|
3 |
0.9677 |
0.9375 |
0.9091 |
0.8571 |
0.8108 |
|
5 |
0.9804 |
0.9615 |
0.9434 |
0.9091 |
0.8772 |
|
10 |
0.9901 |
0.9804 |
0.9709 |
0.9524 |
0.9346 |
|
20 |
0.9950 |
0.9901 |
0.9852 |
0.9756 |
0.9662 |
The table shows that if the fraction of missing information is modest, only a small number of imputations are needed. For example, if , only three imputations are needed to have a 91% efficiency and five imputations are needed to have a 94% efficiency.