Weights following multiple imputation

[Hannah Jongsma]

Dear all,

I am using a case-control data set (n = roughly 2,600) with some missing data.

The added complication is that we have data on the population-at-risk, and know that controls aren't entirely representative.

Ideally, we would therefore like to allow for this using weights.

In a complete case analysis, I've been able to fit this as follows:

meqrlogit case i.ethn [pweight=wt1],or || setting:

However, when I try to fit this following imputation as follows:

mi estimate, eform: meqrlogit case i.ethn[pweight=wt1] || setting:

I get the following error:

weights not allowed.

I'm using Stata 14.

Does anyone know if it is at all possible to run weighted post-imputation analysis?

I can elaborate if need be, but I thought I'd ask the question as simple as possible!

Best wishes,

Hannah