People frequently ask for help with problems they encounter with the mimrgns command (SSC) either by mailing me privately or here on Statalist. As a general advice on problems with mimrgns, I encourage you to answer this one question in a first step.
Is the problem specific to mimrgns?
Whether you get an error message or unexpected results from mimrgns, I recommend to go back to the original non-imputed dataset and run your analysis with official Stata’s margins command. If you are facing the same problem, it is probably not specific to mimrgns.
If the problem is not specific to mimrgns, post your question to Statalist, ignoring the imputed nature of your data. Better yet, include as a last paragraph that your ultimate goal is to use multiple imputation and the mimrgns command but do not focus on the mimrgns command and do not use it in the title of your post. This way you increase your chances for valuable comments from members that do not use mimrgns themselves and/or would not necessarily answer questions related to multiple imputation. When you have worked out a solution for the non-imputed case, go back to MI.
If the problem turns out to be specific to mimrgns, i.e., you are not getting the same error or unexpected results from margins, you can either post to Statalist or contact me personally via mail. I have a strong preference for public discussion unless it is a clearly a bug in the software, in which case there is not too much harm in informing me privately. Note, however, that even in the case of a (suspected) bug report you still lose the opportunity to get valuable comments from members of Statalist on other aspects of your research.
Best
Daniel
Is the problem specific to mimrgns?
Whether you get an error message or unexpected results from mimrgns, I recommend to go back to the original non-imputed dataset and run your analysis with official Stata’s margins command. If you are facing the same problem, it is probably not specific to mimrgns.
If the problem is not specific to mimrgns, post your question to Statalist, ignoring the imputed nature of your data. Better yet, include as a last paragraph that your ultimate goal is to use multiple imputation and the mimrgns command but do not focus on the mimrgns command and do not use it in the title of your post. This way you increase your chances for valuable comments from members that do not use mimrgns themselves and/or would not necessarily answer questions related to multiple imputation. When you have worked out a solution for the non-imputed case, go back to MI.
If the problem turns out to be specific to mimrgns, i.e., you are not getting the same error or unexpected results from margins, you can either post to Statalist or contact me personally via mail. I have a strong preference for public discussion unless it is a clearly a bug in the software, in which case there is not too much harm in informing me privately. Note, however, that even in the case of a (suspected) bug report you still lose the opportunity to get valuable comments from members of Statalist on other aspects of your research.
Best
Daniel
