Hi all,
I am trying out multiple imputation in STATA and I noticed something wondering if people can give feedback on.
I had N=980, 770 of which had complete data-- for the 210 with missing, I used multiple imputation (tried both mvn and pmm, resulted the same problem) and stata imputed 110 of them.
So, I thought when I run mi estimate, I should see analyses done on close to 880 cases. However, when I run mi estimate I saw number of observations around 800.

So, looks like stata is not using all the imputed data during estimation. I wonder if people know where in the process stata might have dropped cases.

Thanks much in advance!