Think statistically. What are the reasons to prefer one test over the
other? It they are testing at the same level, then the only reason to
prefer one of them is because one of them has a greater power. Is there a
uniformly most powerful heteroskedasticity test? Oh Gosh, I doubt that it
is possible to construct a UMPT for anything but simple situations of
one-sided testing when a sample is taken from a normal distribution -- you
implicitly use this knowledge in the t-test (nobody ever discussed why
t-test is preferred to anything else, right?)
So if heteroskedasticity is related to groups of observations, GQ test is
likely to be the most powerful once you guessed the groups correctly. If
it is related to a certain variable, then approaches like Cook-Weisberg
(Stata's -hettest-) would be likely to be the most powerful. White test is
likely to be less powerful IN THOSE SPECIFIC situations, as it is a very
general test. But if you have no idea as regarding the form of your
heteroskedasticity, W might be the thing of your choice as other tests
are simply misspecified: the heteroskedasticity related to omitted
variables is not within their capabilities, they won't detect it. Which
means, low power. So... no good answer to your question.
--- Stas Kolenikov
-- Ph.D. student in Statistics at UNC-Chapel Hill
- http://www.komkon.org/~tacik/ -- [email protected]
* This e-mail and all attachments to it are not intended to provide any
* reasonable point of view and was transmitted to you in error. It
* should be immediately deleted by all recipients unless they really
* enjoy communicating with the author :). Other restrictions apply.
*
* 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/
