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

Re: st: RE: Goldfeldt-Quandt versus -hettest-

From   Richard Williams <[email protected]>
To   [email protected]
Subject   Re: st: RE: Goldfeldt-Quandt versus -hettest-
Date   Thu, 05 Feb 2004 07:18:10 -0500

At 12:02 PM 2/4/2004 -0500, Stas Kolenikov wrote:
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?)
Part of what confuses me is that GQ and -hettest- both seem to test, or can test, similar hypotheses. In the most basic case, if I run

regress y x

and then do either a GQ test or -hettest-, I believe I am testing the hypothesis that as x goes up, the error variances go up (or else down). Or, if I have something more complicated,

regress y x1 x2 x3 x4
[commands to do a gq test using x1]
hettest x1

again it seems like I am testing similar, if not identical, hypotheses.

Now, obviously, the approaches are very different and perhaps the hypotheses are not quite so similar after all. But it remains unclear to me why I would ever prefer the more tedious GQ test (other than to see if my tests all agree with each other.)

* For searches and help try:

© Copyright 1996–2024 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index