R help - Dec 2014 - Missing Data Imputation for Complex Survey Data

Dear all,
I've got a bit of a challenge on my hands. I've got survey data produced
by
a government agency for which I want to use the person-weights in my
analyses. This is best accomplished by specifying weights in {survey} and
then calculating descriptive statistics/models through functions in that
package.

However, there is also missingness in this data that I'd like to handle
with imputation via {mi}. To properly use imputed datasets in regression,
they need to be pooled using the lm.mi function in {mi}. However, I can't
figure out how to carry out a regression on data that is properly weighted
that has also had its missing values imputed, because both packages use
their own mutually incompatible data objects. Does anyone have any thoughts
on this? I've done a lot of reading and I'm not really seeing anything
on
point.

Thanks in advance!

	[[alternative HTML version deleted]]

the mitools package is compatible with the survey package..  asdfree.com
has complete step-by-step R code examples to work with govt microdata.
here are the ones with multiply imputed survey data.  :)

national health interview survey
national survey of children's health
consumer expenditure survey
program of international student assessment
survey of consumer finances
survey of business owners
program for the international assessment of adult competencies

once you have the survey design constructed properly, you can just execute
the svyglm like this:

https://github.com/ajdamico/usgsd/blob/d300884bd63dd05c61e8a6fa76ed7293adae55c2/Consumer
Expenditure Survey/2011 fmly intrvw - analysis examples.R#659




On Fri, Dec 12, 2014 at 7:14 PM, N F <arjunamusic at gmail.com>
wrote:
>
> Dear all,
> I've got a bit of a challenge on my hands. I've got survey data
produced by
> a government agency for which I want to use the person-weights in my
> analyses. This is best accomplished by specifying weights in {survey} and
> then calculating descriptive statistics/models through functions in that
> package.
>
> However, there is also missingness in this data that I'd like to handle
> with imputation via {mi}. To properly use imputed datasets in regression,
> they need to be pooled using the lm.mi function in {mi}. However, I
can't
> figure out how to carry out a regression on data that is properly weighted
> that has also had its missing values imputed, because both packages use
> their own mutually incompatible data objects. Does anyone have any thoughts
> on this? I've done a lot of reading and I'm not really seeing
anything on
> point.
>
> Thanks in advance!
>
>         [[alternative HTML version deleted]]
>
> ______________________________________________
> R-help at r-project.org mailing list -- To UNSUBSCRIBE and more, see
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
> http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
>
	[[alternative HTML version deleted]]