|
|  |
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]
Re: st: RE: Robust regression for outliers in panel data
At 09:29 AM 7/20/2006, Nick Cox wrote:
In my view, although that naturally does not implicate
StataCorp, -rreg- is something of a historical
hangover, implementing as it does a particular
approach documented in one particular paper in
1985. Not that it does not seem smart, sensible
and so forth: the problem is that there are dozens
of ways of doing robust regression that also appear
smart, sensible and so forth, at least to outsiders
to robust statistics, and how is StataCorp supposed
to choose which ones should be dignified with
implementation within official Stata?
I suspect rreg came to be blessed partly because somebody else,
Lawrence Hamilton, wrote the original program! (I was surprised he
spent so much time on it in his Stata book until I made that
connection.) qreg has a lot more intuitive appeal to me. There was
a discussion of qreg versus rreg a few years ago in a thread that
somehow got broken up into these 2 pieces:
http://www.stata.com/statalist/archive/2004-01/msg00780.html
http://www.stata.com/statalist/archive/2004-01/msg00784.html
As a sidelight, I wonder how many esoteric techniques become popular
partly because somebody writes a program for them, and how many good
techniques fall by the wayside because nobody else can figure out how
to estimate them? I've reviewed a few papers that employed
user-written Stata routines. I always tell the authors that, if they
want their work to have some impact, it would be a good idea to get
their programs on SSC. I suspect that many people don't do so
because of the hassle of generalizing the code, doing error-checking
and writing a help file.
-------------------------------------------
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/
