Dear Jayesh
I have asked myself the same question just a week ago and i think i
can give you an interpretation.
I runned the same model in Stata and in LIMDEP because the
differences found in the R squared. In LIMDEP the R squared was
bigger but coeficientes and t ratios were the same!!!
The R-squared provided by STATA with xtreg responds only to the
effects of the independient variables, nor the fixed effects. The
overall R-squared reported by STATA correponds to the R-squared (no
adjusted)in LIMDEP in other models.
I renember to read a text on the official STATA website explaining
the differences in the xtreg and areg R squared, but i cant find it.
In the construction of the R squared of xtreg fixed effects impact is
droped.
So, what is the most suitable R squared???? I think that it depends
on your results. In my model i am gonna use the areg R squared and
explain the composition of it.
Hope it helps.
Alfredo L�pez
Universidad de Zaragoza
Zaragoza (Spain)
--- In [email protected], JAYESH KUMAR <jayesh@i...> wrote:
> Dear Statalisters,
> I have posted this mail a week before, but didn't get any response.
> I am sending it again in the hope someone may help me.
> -Jayesh
>
> ---------- Forwarded message ----------
> Date: Fri, 17 Oct 2003 23:20:49 +0530 (IST)
> From: JAYESH KUMAR <jayesh@i...>
> To: statalist@h...
> Subject: help with R-square of xtreg,fe and areg
>
> Dear Users,
>
> I have a basic question regarding the difference in xtreg,fe and
areg.
> Which R-square should I report in my results? R-sq: within, between
or
> overall obtained from xtreg,fe or shall I report R-squared or
Adjusted
> R-squared obtained from the areg.
>
> I am inclosing the output from the two commands. As you can clearly
see
> that both results are same in terms of coeff, p-value, F-stats,
etc. The
> only difference is with different R-squares. I am bit puzzled here,
which
> R-sq should I report in my final tables. I am using Stata 7.
>
> Any suggestion would be of help. Is there any reference, in which I
can
> find the differences in R-squares, in terms of interpretation,
rather than
> derivation?
>
> TIA,
>
> -Jayesh Kumar
>
**********************************************************************
**
> JAYESH KUMAR,
> Research Scholar,
> Indira Gandhi Institute Of Development Research (IGIDR),
> Gen. Arun Kumar Vaidya Marg,
> Santosh Nagar, Goregaon (East), Mumbai-400065, INDIA.
> Tel # + 91 (22) 2840 0919/0920/0921 Extn. 591(Office) 263
(Residence)
> Fax # + 91 (22) 2840 2752/2026
> visit: www.igidr.ac.in/~jayesh
> SSRN Papers on the web at:
> http://papers.ssrn.com/sol3/cf_dev/AbsByAuth.cfm?per_id=333715
>
**********************************************************************
**
> When I don't know what I'm doing I'm doing Research!
>
>
>
>
> OUT PUT:
>
======================================================================
==========
>
>
> . xtreg inc age pq_a,fe
>
> Fixed-effects (within) regression Number of obs
= 5132
> Group variable (i) : ind Number of groups
= 2524
>
> R-sq: within = 0.0485 Obs per group: min
= 1
> between = 0.0045 avg
= 2.0
> overall = 0.0054 max
= 7
>
> F(2,2606)
= 66.35
> corr(u_i, Xb) = -0.8909 Prob > F
= 0.0000
>
> --------------------------------------------------------------------
----------
> inc | Coef. Std. Err. t P>|t| [95%
Conf. Interval]
> -------------+------------------------------------------------------
----------
> age | -.0107904 .0009367 -11.52 0.000 -
.0126272 -.0089536
> pq_a | 9.67e-06 .0000694 0.14 0.889 -
.0001265 .0001458
> _cons | .3441254 .0208627 16.49
0.000 .3032161 .3850346
> -------------+------------------------------------------------------
----------
> sigma_u | .26712239
> sigma_e | .07998241
> rho | .91772273 (fraction of variance due to u_i)
> --------------------------------------------------------------------
----------
> F test that all u_i=0: F(2523, 2606) = 4.40 Prob >
F = 0.0000
>
>
>
>
> . areg inc age pq_a,absorb(ind)
>
> Number of
obs = 5132
> F( 2,
2606) = 66.35
> Prob >
F = 0.0000
> R-
squared = 0.8109
> Adj R-
squared = 0.6277
> Root
MSE = .07998
>
> --------------------------------------------------------------------
----------
> inc | Coef. Std. Err. t P>|t| [95%
Conf. Interval]
> -------------+------------------------------------------------------
----------
> age | -.0107904 .0009367 -11.52 0.000 -
.0126272 -.0089536
> pq_a | 9.67e-06 .0000694 0.14 0.889 -
.0001265 .0001458
> _cons | .3441254 .0208627 16.49
0.000 .3032161 .3850346
> -------------+------------------------------------------------------
----------
> ind | F(2523, 2606) = 4.399 0.000 (2524
categories)
>
>
>
>
>
> *
> * 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/
