Statalist


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

Re: st: variance of R2


From   "Stas Kolenikov" <skolenik@gmail.com>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: variance of R2
Date   Wed, 8 Aug 2007 09:44:16 -0500

On 8/8/07, agostino@unical.it <agostino@unical.it> wrote:
> Dear all,
> does anyone of you know how to compute the variance of the R2 (coefficient of
> determination) in STATA?

What would you need this for? Anything that deals with R^2 is ad hoc,
and standards of "decent" R^2 vary by discipline from 99% in physics
and engineering to 5% in economics.

R^2 is a nonlinear transformation of F-statistic, and you can look up
the variance of F (probably a non-central one) distribution in the
standard textbooks. Then you can apply the delta method to get
analytic asymptotic variance. The bootstrap is faster these days, of
course, but it's always better to have answers from two completely
different methods to see if they match. And if you thought the
bootstrap is a universal remedy for every situation, you must read
this: http://www.citeulike.org/user/ctacmo/article/575126 (and Stata
Corp. might consider putting warnings arising from it in the biggest
print possible in [R] bootstrap in the next printing :))

-- 
Stas Kolenikov, also found at http://stas.kolenikov.name

Small print: Please do not reply to my Gmail address as I don't check
it regularly.
*
*   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 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index