Bookmark and Share

Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down on April 23, and its replacement, statalist.org is already up and running.


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: st: Using mi estimate with xtmixed and accounting for sampling weights


From   Stas Kolenikov <skolenik@gmail.com>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: Using mi estimate with xtmixed and accounting for sampling weights
Date   Sat, 14 Jul 2012 13:49:53 -0500

To do this statistically properly, you need to follow Shao & Sitter
(1996; http://www.citeulike.org/user/ctacmo/article/1269394). Namely:

1. pull a bootstrap sample;
2. run the imputation procedure once on the imputed sample;
3. compute the point estimates of interest;
4. repeat 1-3 a large number of times (bootstrap-large, hundreds, not
the MI-large, which is five)
5. combine the results using the bootstrap rules: point estimate =
average of the results, variance estimate = variance over the
bootstrap samples.

1-3 require a custom-written estimation command, but it should not be
too difficult to compose.

On Fri, Jul 13, 2012 at 1:35 PM, Justina Kamiel Grayman
<justina.grayman@nyu.edu> wrote:
> Hi everyone,
>
> I am attempting to estimate a two-level model with a random intercept.
> I am using xtmixed. My data was also imputed, using MI ICE. I now have
> 10 imputed datasets. Lastly, I am trying to include sampling weights
> at both levels of the model. So basically I need to include sampling
> weights in a two-level model estimated on imputed data.
>
> The problem: When I perform the analysis (syntax below), it runs fine
> on the original data.  It also runs when I use an if statement to run
> the analysis on only one of the imputed datasets. However, when I try
> to use mi estimate as a prefix, the model does not run. Is there some
> incompatibility between mi estimate and using pweights in xtmixed?
> Does anyone know why this is occurring?
>
> Thanks!
>
>
> Syntax That Does Not Work:
>
> mi estimate: ///
> xtmixed studeff ///
> classclimmeanc classclimvarc ///
> pctnonwhitec ///
> educc female usborn ///
> black latino asian other multi ///
> btgtaugc wrkshpcntc someba  ///
> bcna9c urban rural northeast southeast west privateschl [pw = totwgt] ///
> || idschool: , covariance(unstructured) pweight(schwgt) , mle pwscale(size)
>
>
> Syntax That Does Work
>
> xtmixed studeff ///
> classclimmeanc classclimvarc ///
> pctnonwhitec ///
> educc female usborn ///
> black latino asian other multi ///
> btgtaugc wrkshpcntc someba  ///
> bcna9c urban rural northeast southeast west privateschl [pw = totwgt] ///
> if  _mi_m==1 ///
> || idschool:  , covariance(unstructured) pweight(schwgt) , mle pwscale(size)
> --
> Justina Kamiel Grayman
> Doctoral Student, Psychology & Social Intervention
> New York University
> *
> *   For searches and help try:
> *   http://www.stata.com/help.cgi?search
> *   http://www.stata.com/support/statalist/faq
> *   http://www.ats.ucla.edu/stat/stata/



-- 
---- Stas Kolenikov
-- http://stas.kolenikov.name
---- Senior Survey Statistician, Abt SRBI
-- Opinions stated in this email are mine only, and do not reflect the
position of my employer
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/


© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   Site index