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: Chi-squared test after multiple imputation with ice


From   Stas Kolenikov <skolenik@gmail.com>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: Chi-squared test after multiple imputation with ice
Date   Wed, 6 Oct 2010 15:35:24 -0500

On Wed, Oct 6, 2010 at 12:04 PM, Jodie Katon <jkaton2@yahoo.com> wrote:
> I am trying to run chi-sq or anova to test the differences in means or prop=
> ortions across a categorical variable w/ 3 categories. Can you do this u=
> sing mim? The only way I have been able to do it is by first running a
> regression and then using testparm. For example:
>
> xi:mim:regress maternage i.diaboutcome
> mim:testparm _Idiaboutco_1 _Idiaboutco_2

That's the only way to do it justified within the MI context. It only
gives you a better set of parameter estimates and standard errors, it
cannot give you ANY test statistics. Of course you can mechanically
compute the test statistics on the completed data sets, but they are
not meaningful, in the sense that they do not have any theoretical
properties even with respect to the randomness in multiple
imputations. I suspect that they are also biased towards the null, but
that is only a gut feeling. So Wald tests (which is what you have been
using) is the only test procedure that is appropriate with MI.

-- 
Stas Kolenikov, also found at http://stas.kolenikov.name
Small print: I use this email account for mailing lists only.
*
*   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