Problems with Regression with Multiply-Imputed Data

[Erez Garnai]

Hello!
I have been trying to run regressions with multiply-imputed data (using the mi estimate command), and it all went well, until I added one specific variable that's been causing me problems.
It's a categorical variable that's essentially an interaction between two variables (which were themselves imputed), which makes wonder whether that's the source of my problems.
When I try to run the model, I get the following error message (below).

I then tried creating the variable following the multiple imputation process (rather than have the variable imputed to begin with), using the "mi passive" prefix; that didn't help either (same error message).
Any idea what's going on here, and how to fix it?
I'd be happy to share more specific information if that would help.
Thank you!

[Rich Goldstein]

the -mi passive- route should have solved the problem at least as you described it; however, there is an option ( esampvaryok) that can be used; I would be very careful using this given what you said but it can be used; my guess is that there is something else going on that you have not told us about but data examples (read the FAQ on how to post these) are needed for more