Statalist The Stata Listserver


[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

st: RE: Why not always specify robust standard errors?


From   "Schaffer, Mark E" <[email protected]>
To   <[email protected]>
Subject   st: RE: Why not always specify robust standard errors?
Date   Sun, 18 Feb 2007 23:07:05 -0000

Another one that didn't go through the first time.  Again, I hope it isn't too late to be useful to Richard or somebody out there.

> -----Original Message-----
> From: Schaffer, Mark E 
> Sent: 14 February 2007 08:45
> To: [email protected]; [email protected]
> Subject: AW: Why not always specify robust standard errors?
> 
> Richard,
> 
> To answer your second question, White's general test for 
> heteroskedasticity - based on the levels, squares and 
> crossproducts of the indep vars etc. - is, intuitively, a 
> comparison of the elements of the standard, non-robust VCV, 
> and the robust VCV.  The squares correspond to the diagonals 
> of the two VCVs, etc.  If you look at just the squares, you 
> are looking for a particular kind of heteroskedasticity - the 
> kind that affects the SEs.  So your informal observation is a 
> kind of test for heteroskedasticity, and your observation is 
> that you don't often find heteroskedasticity that is so 
> severe that you can see it in an informal comparison of the 
> two kinds of SEs.
> 
> I *think* this is right - I am far away from my books at the 
> moment....
> 
> Cheers,
> Mark
> 
> 
> -----Urspr�ngliche Nachricht-----
> Von: [email protected] im Auftrag von 
> Richard Williams
> Gesendet: Di 2/13/2007 4:22
> An: [email protected]
> Betreff: st: Why not always specify robust standard errors?
>  
> A student asked me a question the other day that I couldn't 
> think of a definitive answer for: Why not always specify 
> -robust- when using OLS regression?  My initial reaction is 
> to say that you shouldn't relax restrictions unnecessarily; 
> and there are various post-estimation commands where Stata 
> will at least whine at you if you've used robust standard 
> errors (e.g. -lrtest-).  But in practice, your model is 
> probably at least a little mis-specified and/or there may be 
> some degree of heteroskedasticity, so maybe robust is a good 
> idea.  Any thoughts on the matter?
> 
> Incidentally, my own experience is that robust standard 
> errors usually aren't all that different from non-robust 
> standard errors. Is that what other people have found as 
> well, or is just me?
> 
> 
> 
> -------------------------------------------
> Richard Williams, Notre Dame Dept of Sociology
> OFFICE: (574)631-6668, (574)631-6463
> FAX:    (574)288-4373
> HOME:   (574)289-5227
> EMAIL:  [email protected]
> WWW (personal):    http://www.nd.edu/~rwilliam
> WWW (department):    http://www.nd.edu/~soc 
> 
> *
> *   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/
> 
> 

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