-mi estimate- within sub-samples that are defined by imputed variables

[Mark Marshall]

Dear Stata-listers,

I have an imputed a dataset, and I would like to -mi estimate- in subsamples defined by imputed variables, to confirm interactions that I detected in the complete case analysis (albeit in MAR data with possibly biased estimates)

Obviously, the subsample for estimation varies from imputation to imputation, depending on the value for the variable that was imputed.

Obviously also, Stata notices that the estimation sample varies between imputations, and stops the procedure.

What is the strategy to whereby one can estimate in subsamples defined by imputed variables?

(Stata MP 13.2 iOSX Yosemite)

Thanks,

MM

[Richard Williams]

See -help mi estimate-. Here is one of the options:

esampvaryok allows estimation to continue even if the estimation sample varies across imputations. mi estimate stops if the estimation sample varies. If esampvaryok is specified, results from all imputations are used to compute MI estimates and a warning message is displayed at the bottom of the table. Also see the esample() option. See Potential problems that can arise when using mi estimate for more information.

[Mark Marshall]

Thanks, I can breathe again!