Bookmark and Share

Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.


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

Re: st: Comparing weighted and unweighted distributions via Chi2 test


From   Austin Nichols <[email protected]>
To   [email protected]
Subject   Re: st: Comparing weighted and unweighted distributions via Chi2 test
Date   Thu, 9 Jun 2011 11:18:02 -0400

Stas--
I think a generalization of the Hausman (1978) test (findit xtoverid)
is appropriate for comparing weighted and unweighted regression, but
Duru wants to compare vectors of means.  Here is a simple
generalization of -hotelling- (1931) instead:

webuse nhanes2, clear
isid sampl
expand 2
bys sampl: g wtd=_n-1
replace finalwgt=1 if wtd==0
reg wtd age height weight bpsystol bpdiast tcresult tgresult hdresult
[pw=finalwgt]

where the F test is the test of the null than the vectors of means are
the same in both samples (wtd and not); compare e.g.

webuse nhanes2, clear
reg sex age height weight bpsystol bpdiast tcresult tgresult hdresult
hotelling age height weight bpsystol bpdiast tcresult tgresult hdresult, by(sex)

though probably some kind of permutation/randomization test is preferable.

On Thu, Jun 9, 2011 at 10:56 AM, Stas Kolenikov <[email protected]> wrote:
> If you kind find this paper, it will be quite useful:
> http://www.citeulike.org/user/ctacmo/article/1036965. Pfeffermann is
> one of few statisticians who is aware of Hausman (1978) test that just
> celebrated a third of a century... think about the time for good ideas
> to diffuse from one discipline to another. This is the test that's
> directly applicable to your case, but Stata is reluctant to run it
> with weighted data, so you'd have to find ways to force it to perform
> it.
>
> On Wed, Jun 8, 2011 at 4:05 PM, Duru <[email protected]> wrote:
>> Dear all,
>>
>> In order to test if my nonresponse weights change my survey outcomes
>> substantially, I want to conduct Chi2 tests or t-tests between
>> weighted and unweighted distributions/means for a number of variables.
>> Since, I dont know a practical way to do this on Stata, I have to
>> insert weighted and unweighted frequency tables or calculate t-values
>> manually from weighted and unweighted mean and variance estimates. Any
>> ideas on how to do it more easily? (using Stata 10.1)
>>
>> Best,
>>
>> Duru

*
*   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 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index