Can I obtain one estimate of imputed figures from multiple imputation for analyses, or must I embed each test in the MI combine command

[Jacob Thomas]

I have succeeded in doing a multiple imputation without interaction variables in Stata.
However I have a question: Must I embed all operations (e.g. reg, anova, sem) within the micombine command to analyze a multiply imputed data set as written in the code below?
I ask this because this operation gives me a data set of 10 imputed values (as I requested). I would like if possible to use such data and I was wondering if it is possible to obtain just one line of imputed values (perhaps an average of all the imputed values?) for which I can just conduct further analyses with?
Of course I could this by extracting the numbers and averaging them in Stata or Excel, but I presume Stata should have some operation that does this.
My code is below (though it is adapted from Don Treiman's textbook and probably there may be something more recent and versatile):
ice visa income lnincome pop inequality polfree using visafree_mobility_nearest_imputation.dta , replace seed(12345) m(20) boot ///
cmd(polfree : ologit) ///
eq(lnincome : pop inequality visa, ///
polfree : inequality visa, ///

[daniel klein]

To answer your question: it is, technically, possible to use one (averaged) value, but it defeats the hole purpose of multiple imputation, so you do not want to do it.

You should probably use Stata's mi suit of commands, not the user-written ice command (though it sometimes seems to be the better choice). When posting code, use the respective environment (see FAQs on this). You also want to explain where user-written commands, like ice, come from (so to make sure we all talk about the same thing). See the FAQs on this, too.

You claim that you request m=10 datasets, but the (part of the) command you show requests m=20.

Best
Daniel

[Jacob Thomas]

Thanks Daniel, I figured it out!