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: OLS assumptions not met: transformation, gls, or glm as solutions?
From
Nick Cox <[email protected]>
To
[email protected]
Subject
Re: st: OLS assumptions not met: transformation, gls, or glm as solutions?
Date
Mon, 17 Dec 2012 16:41:10 +0000
We can converge by holding fast to the idea that regression is about
modelling the conditional mean. If you have a inappropriate model for
the conditional mean, wondering how well you can fit it is not a very
interesting or useful question.
Nick
On Mon, Dec 17, 2012 at 4:37 PM, JVerkuilen (Gmail)
<[email protected]> wrote:
> On Mon, Dec 17, 2012 at 10:27 AM, Maarten Buis <[email protected]> wrote:
>
> The whole point of robust standard errors is not that it
>> "solves" in some way for heteroskedasticity, it just makes that
>> "assumption" irrelevant. For more, see section 20.20 of the User's
>> Guide.
>
> I think I'm vehemently agreeing with you Maarten, but "sort of
> irrelevant" is probably a better characterization. Heteroscedasticity
> is often the sign of a very bad model. There's a temptation to "white
> wash" a bad model by using remedial measures such as one of the
> various heteroscedasticity-consistent VCE estimates, bootstrap,
> jackknife, etc., but if the model is not appropriate all the remedial
> measures in the world won't truly fix it up.
> *
> * For searches and help try:
> * http://www.stata.com/help.cgi?search
> * http://www.stata.com/support/faqs/resources/statalist-faq/
> * http://www.ats.ucla.edu/stat/stata/
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/faqs/resources/statalist-faq/
* http://www.ats.ucla.edu/stat/stata/