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

Re: st: test for equality of covariances (not covariance matrices)


From   Joseph Coveney <jcoveney@bigplanet.com>
To   Statalist <statalist@hsphsun2.harvard.edu>
Subject   Re: st: test for equality of covariances (not covariance matrices)
Date   Sat, 17 Sep 2005 17:43:06 +0900

Jerry Hamster wrote:

I would like to test the equality of covariances (not covariance matrices)
for two independent samples.  The only tests I have found are for covariance
matrices.  Can anyone steer me in the right direction (offer references,
etc)?  Many thanks

--------------------------------------------------------------------------------

Couldn't Fisher's z-transform be one approach to the question, "Is the
correlation coefficient (covariance) between A and B equal to that between C
and D, given that the dataset pairs AB and CD are independent samples?"  Use
the transform to construct confidence intervals for each of the correlation
coefficients and determine whether they overlap.  (Assumes "large" samples.)

www.fon.hum.uva.nl/praat/manual/Correlation__Confidence_intervals___.html
recommends Ruben's approximation for confidence interval construction over
Fisher's.

If you were able to model patterned covariance structures for residual
errors with -xtmixed-, then that would be another approach, using a
likelihood ratio test.

Joseph Coveney

*
*   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