Statalist


[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

RE: st: Issue with multiple imputation -ICE-


From   Maarten buis <maartenbuis@yahoo.co.uk>
To   statalist@hsphsun2.harvard.edu
Subject   RE: st: Issue with multiple imputation -ICE-
Date   Wed, 13 Feb 2008 17:02:53 +0000 (GMT)

--- René Wevers <rene@ricardis.tudelft.nl> wrote:
> The absolute number of employees does not have any zero values, as
> would be expected. Thus it is still a complete mystery to me why
> negative values appear after imputation.
> 
> I am now applying the -impute- command from Stata, which seems to
> perform better. However, there are also dummy variables which need to
> be imputed. This becomes a bit messy when applying the -impute-
>  command since no estimation technique can be selected (-logit- 
> instead of -regress-).
> I am now thinking of applying -impute- and then rounding the results
> for  the dummies, since these dummy variables are not very important
> for my analysis (only for control). I know it isn't very pretty, but
> I don't see any other (simple) solutions.

I sympathise with your despair: Multiple Imputation sounds easy but
when you try to apply it to real data it pretty soon gets very messy.
My guess is still that there is a problem with your data which you need
to solve first. One thing you might want to look at is this:

gen mis = missing(emptot_fte)
twoway kdensity emptot_abs if mis == 1 ||   ///
       kdensity emptot_abs if mis == 0 ,    ///
       xtitle("number of employees")        ///
       legend(order( 1 "missing fte"        ///
                     2 "not missing fte"))

This way you can see if there is anything "weird" going on with the
distribution of emptot_abs for the companies with missing values on
emptot_fte.  

I would definately not use -impute-. The logic against -impute-
explained here: 
http://www.stata.com/statalist/archive/2007-12/msg00507.html

IMHO -impute- is one of the few offical Stata commands that StataCorp
should consider withdrawing. 

Hope this helps,
Maarten

-----------------------------------------
Maarten L. Buis
Department of Social Research Methodology
Vrije Universiteit Amsterdam
Boelelaan 1081
1081 HV Amsterdam
The Netherlands

visiting address:
Buitenveldertselaan 3 (Metropolitan), room Z434

+31 20 5986715

http://home.fsw.vu.nl/m.buis/
-----------------------------------------


      __________________________________________________________
Sent from Yahoo! Mail - a smarter inbox http://uk.mail.yahoo.com

*
*   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