There are two questions revolving PROC MI that I would like to raise:
(1) How can I know to what extent has the imputation completed (i.e., the number of imputation completed) during the imputation process? Multiple imputation is a computer-intensive process. Moreover, its theory tells us to perform as many imputations as possible and putting in as many variables as possible yields the best results, which makes the user feeds in an enoromous dataset when executing imputation, a practice that may cause the time needed to further increase. I often spend much time checking my computer to see if my imputation is done, only to find it is yet to be done. Sometimes the need for an imputed dataset is imminent, so this makes me anxious.
(2) How can I know the "culprit variable" when the errror warning "An imputed variable value is not in the specified range after XXX tries" appears in my log? At present, I have to go through the vast amount of continuous variables in the imputation model to see what variable(s) cause this error. This greatly wastes my time.
Thanks!
