mimrgns: inconsistent estimation sample

[Andrew Perrin]

Greetings-

I am trying to use the SSC command mimrgns to compute predicted probabilities for eprobit models. The models are ERMs using endogenous treatment effects, if that matters, and are estimated using:

mi estimate, saving(filename) esample(sampname) cmdok: svy: eprobit ...

I then use:

mimrgns using filename, esample(sampname) dydx(*) predict(pr)

This works for some of the models, but in several cases I get the following error:

inconsistent estimation sample levels 0 and 1 of factor VARNAME
an error occurred when mi estimate executed mimrgns_estimate on m=2
r(459);

Any advice will be most appreciated. Thanks.

[daniel klein]

Andrew,

thanks for your interest in the mimrgns command. Please read and follow the advice I give in this post.

From what you report here, I suspect you might get the same error in the non-imputed case, meaning that the error is not specific to mimrgns. The one thing that I find a bit unusual is that the error occurred for the second imputed dataset, not the first one. Do you need the cmdok option to get eprobit to run?

Best
Daniel

[Andrew Perrin]

Thank you Daniel. I will follow those directions and respond. I appreciate the quick advice! These are protected data that live on a secure machine so it may take some time before I can work on this diagnosis.

[Andrew Perrin]

Moving to a new topic here after having replicated in a single non-imputed case.