MI is attractive for a number of reasons. First, it works in conjunction
with standard complete-data methods and software. One the MI'S have been generated, the analyses can be carried out using procedures in SAS, LISREL, or virtually any other statistical package. Second, one set of m imputations may be used for a variety of analyses; there is often no need to reimpute when a new analysis is performed. Third, the inferences - standard errors, p-values, etc. - obtained from MI are generally valid because they incorporate uncertainty due to missing data. Finally, MI is attractive because it can be highly efficient even for small values of m. In many applications,
just 3-5 imputations are sufficient to obtain excellent results.