I am using the MI suite to impute individual data across 3 waves of data
collection. My sample size is about 18,000. Ideally I am hoping to
impute 25 datasets, but STATA will produce an error or not reach
concavity somewhere in the 3rd dataset. So I am imputing 2 datasets at a
time. Although, sometimes it will only impute 1 dataset and then produce
an error in the 2nd. I assumed that each imputation is unaffected by the
previous dataset imputed and the subsequent dataset imputed, so I was
going to use 'mi add' to put all the mini datasets together to reach 25.
Is there something horribly wrong with my imputation, or something I'm
missing I should be doing? Any advice would be most helpful!
