Multiply-imputed complex survey data: post-estimation command for design effect (deff)

[Ayalew Astatkie]

I am working with a multiply-imputed complex survey data. I want to determine the design effect of the sampling design on estimates. However, for multiply-imputed data the post-estimation command estat effects, deff doesn't work. I would appreciate if anyone suggests me a way for estimating deff for multiply-imputed survey data in Stata.

[Steve Samuels]

Sorry, but I missed this in January. mi extract one of the mi datasets; svyset it (not mi: svyset); do your analysis runestat effects, deff. The design effect is not a precise descriptive parameter, so one (or two) extracts should be enough to estimate it.