[R] Survey imputation

[David Winsemius]

See Lumley's text "Complex Surveys", chapter 9: "Missing Data". In addition to the survey package he also uses teh RSQLite and mitols packages in his worked examples. He doesn't use the NHIS data for the missing data chapter, but he does provide examples using NHIS for other tasks.

-- 
David.


On Jun 13, 2013, at 10:51 AM, Scott Raynaud wrote:

> I paln on using R to do the imputation once I figure out how to do it.
> 
>  
> ----- Original Message -----
> From: Bert Gunter <gunter.berton at gene.com>
> To: Scott Raynaud <scott.raynaud at yahoo.com>
> Cc: "r-help at r-project.org" <r-help at r-project.org>
> Sent: Thursday, June 13, 2013 12:45 PM
> Subject: Re: [R] Survey imputation
> 
> Is this an R question?
> 
> Seems like it belongs on a statistical or survey list, not r-help.
> 
> Cheers,
> Bert
> 
> On Thu, Jun 13, 2013 at 10:37 AM, Scott Raynaud <scott.raynaud at yahoo.com> wrote:
>> I'm working with NHIS survye data.  I'd like the to use muliple imputation
>> to cover the missing data for the variables in which I'm interested.  My
>> question concerns the use of certain variables in the imputation model.
>> For example, race would be an important predictor in the imputation
>> model, but it has been imputed (hot deck) as have other variables that
>> might be useful in the imputation model.  What is the wisdom of using
>> data that have been imputed in subsequent imputations models?
>> 


David Winsemius
Alameda, CA, USA