Statalist The Stata Listserver


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

st: RE: RE: adjusted r square


From   "Ranjita majumder Singh" <[email protected]>
To   <[email protected]>, <[email protected]>
Subject   st: RE: RE: adjusted r square
Date   Tue, 20 Feb 2007 09:02:27 -0500

Hi All,
thank you for your response. I had been told to use adjusted R2 because my R2 were decreasing when I was adding new variables. Hence I have to use adjusted R2 but I wasn't sure which one from xtreg (i.e, between, within or overall) seemed a way to compare the one with areg....
Would appreciate help on this
Regards
Ranjita
 
Ph.D candidate Strategy and Organization Theory
Rotman School of Management
University of Toronto
105 St. George St
Toronto, ON
M5S3E6

________________________________

From: [email protected] on behalf of Schaffer, Mark E
Sent: Tue 2/20/2007 8:39 AM
To: [email protected]
Subject: st: RE: adjusted r square



Hi all.

> -----Original Message-----
> From: [email protected]
> [mailto:[email protected]] On Behalf Of Kit Baum
> Sent: Tuesday, February 20, 2007 1:28 AM
> To: [email protected]
> Subject: st: adjusted r square
>
> Ranjita said
>
> I saw your note in Stata re: adjusted R square and that while
> -areg- calculates the adjusted R^2 statistic, but, it is not
> statistically equivalent to the FE model in -xtreg, fe-:
> Therefore, my question is if I wanted to stay with xtreg, fe
> how whould I calculate adjusted R square? Thank you for your help
>
> The model estimated by xtreg,fe is certainly 'statistically
> equivalent' to that estimated by areg:
>
> webuse grunfeld
> xtreg invest mvalue kstock,fe
> areg invest mvalue kstock,absorb(company)
>
> These models have identical coefficients, standard errors, F-
> statistics and RMSEs. What is not equivalent? These are just
> two different ways of implementing the LSDV (least squares dummy
> variable) approach, taking into account the loss of d.f.

Using Kit's example,

webuse grunfeld
qui areg invest mvalue kstock,absorb(company)
di e(r2)
di e(ar2)
qui xi: reg invest mvalue kstock i.company
di e(r2)
di e(r2_a)
qui xtreg invest mvalue kstock, fe
di e(r2)
di e(r2_a)

the R2 and adjusted R2 for areg is identical to that for the LSDV
approach, and much higher (96%) than that for the FE model (75-76%).

I suppose what is happening is that the areg R2 and adj R2 refer to the
total variation in the dependent variable, whereas the FE R2 and adj R2
refer to the variation in the demeaned dependent variable.

So the question for Ranjita is, which R2 do you want?

--Mark

> Kit Baum, Boston College Economics
> http://ideas.repec.org/e/pba1.html
> An Introduction to Modern Econometrics Using Stata:
> http://www.stata-press.com/books/imeus.html
>
>
> *
> *   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/



*
*   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