Stata The Stata listserver
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

Re: st: Re: Estimate parameters from multiple datasets


From   "David A. Wagstaff" <daw22@psu.edu>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: Re: Estimate parameters from multiple datasets
Date   Fri, 25 Oct 2002 11:54:43 -0400

10/25/2002


John!


I would be interested in testing your package.  I am
using suest (regh) and test(nl) in combination.  When I
figure out how to get the vector and covariance matrix 
used by test, I would like to be able to combine the
estimates based on 10 multiply imputed datasets.

David, daw22@psu.edu

John Carlin wrote:
> 
> We have recently completed a suite of ado files that will do what
> Jennifer needs. They provide flexible data manipulation and combined
> inferences from multiply imputed datasets. The -mifit- inferences
> (should!) work for any Stata estimation command. This package has
> been submitted to the Stata Journal but I will send a copy separately
> to Jennifer and am happy for others to beta-test if interested. Would
> prefer not to post the software for general use until after review
> and changes from first round of testing.
> 
> John Carlin
> 
> >Date: Tue, 22 Oct 2002 07:36:52 -0500 (CDT)
> >From: Jennifer Nicoll Victor <jnvictor@artsci.wustl.edu>
> >Subject: st: Estimate parameters from multiple datasets
> >Stata Experts:
> >I am writing to ask if anyone has do- code or guidance for generating
> >estimates from multiple datasets. I have five datasets as generated
> >from a multiple imputation (MI) process to fill-in missing values. I am
> >obtaining poisson estimates and need to generate a single parameter
> >estimate from the five datasets.
> >King et al. (2001) provide guidance for the cofficients and their SEs.
> >The coefficients are easy enough, you just average them. But the SEs are
> >more tricky (see equation (3) in King et al. 2001 APSR).
> >Can anyone provide guidance in generating the code to compute a single SE
> >for each estimate from these five datasets?
> >Thanks very much!
> >Jennifer
> >- ----------------------------------------
> >Jennifer Nicoll Victor
> >Ph.D. Candidate in Political Science
> >Washington University in St. Louis
> >E-mail: jnvictor@artsci.wustl.edu
> >Homepage: www.artsci.wustl.edu/~jnvictor
> >- ----------------------------------------
> >
> >Date: Tue, 22 Oct 2002 15:05:11 +0200
> >From: Thomas Gschwend <Thomas.Gschwend@mzes.uni-mannheim.de>
> >Subject: Re: st: Estimate parameters from multiple datasets
> >clarify is your friend.
> >type
> >findit clarify
> >or download it from Kings homepage
> >Thomas
> >
> >
> >Date: Tue, 22 Oct 2002 08:06:13 -0500
> >From: Fred Wolfe <fwolfe@arthritis-research.org>
> >Subject: Re: st: Estimate parameters from multiple datasets
> >You want Gary King's -miest-
> >You should be able to find it at http://gking.harvard.edu
> >Fred Wolfe
> >
> >
> >Date: Tue, 22 Oct 2002 08:09:08 -0500 (CDT)
> >From: Jennifer Nicoll Victor <jnvictor@artsci.wustl.edu>
> >Subject: Re: st: Estimate parameters from multiple datasets
> >Thomas:
> >Thanks for the suggestion. Clarify is definitely my friend and I use it
> >all the time. However, in this case I have a model with robust standard
> >errors and a cluster option defined. Clarify will not estimate a model
> >with these specifications (well... it estimates it but does not use robust
> >SEs or clustering). Do you know of a way to use these options in
> >Clarify??? If so, I would be MOST grateful for the enlightenment.
> >Thanks a bunch!
> >- -Jennifer
> >
> >
> >
> >Date: Tue, 22 Oct 2002 09:17:19 -0400
> >From: frone@ria.buffalo.edu
> >Subject: Re: st: Estimate parameters from multiple datasets
> >- -miest- is no longer available. -clarify- has absorbed this function.
> >
> >
> >Date: Tue, 22 Oct 2002 15:35:57 +0200
> >From: Thomas Gschwend <Thomas.Gschwend@mzes.uni-mannheim.de>
> >Subject: Re: st: Estimate parameters from multiple datasets
> >correct! If I remember correctly than -miest- does not do the trick either,
> >since clarify is based on it.
> >I once had the same problem and ended up calculating the sdterrs "by hand",
> >i.e. I set up an excel sheet, pasted in the estimation results ect.
> >You can also save the estimation results in a seperate file (-parmest-) and
> >compute the MI-stderrs in Stata.
> >Thomas
> 
> John Carlin
> Professor, University of Melbourne Departments of Paediatrics and Public Health
> 
> Director, Clinical Epidemiology and Biostatistics Unit
> Murdoch Children's Research Institute
> Royal Children's Hospital
> Parkville,  Vic. 3052.  AUSTRALIA
> 
> E-mail: jbcarlin@unimelb.edu.au
> Tel. +61-3-9345-6363     Fax +61-3-9345-6000
> 
> *
> *   For searches and help try:
> *   http://www.stata.com/support/faqs/res/findit.html
> *   http://www.stata.com/support/statalist/faq
> *   http://www.ats.ucla.edu/stat/stata/
*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   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   |   What's new   |   Site index