st: RE: RE: ivreg2

["Schaffer, Mark E" <M.E.Schaffer@hw.ac.uk>]

Marta,

> From: owner-statalist@hsphsun2.harvard.edu
> [mailto:owner-statalist@hsphsun2.harvard.edu] On Behalf Of M.Sanchez-
> Sache@lse.ac.uk
> Sent: 22 August 2007 18:51
> To: statalist@hsphsun2.harvard.edu
> Subject: RE: RE: ivreg2 
>
> Mark,
>
> Thanks for your answer. However, something stills puzzles me. Why
would we
>  interested in uncentered statistics if we do have a constant in our
model?
> When STATA computes uncentered statistics, is it not taking into
account the
> constant we explicitly introduce in our model? I thought that
uncentered
> statistics were only used when a constant was lacking in the model.
>
> Thanks again.
>
> Marta

I can think of two reasons.  First, sometimes users might estimate a
model with the -nocons- option but the model actually does have a
constant.  (For example, if the regression is on individuals and there
is a gender dummy, the user might exclude a constant and include both a
male and a female dummy.)  From a programming point of view, it was
easiest just to report both centered and uncentered R-sqs and let the
user decide which s/he wants to use.  Second, we don't always use
regressions to estimate models; sometimes we do tests based on
artificial regressions.  There are N*R-sq tests which require the
uncentered R-squared.

Cheers,
Mark

> -----Original Message-----
> From: owner-statalist@hsphsun2.harvard.edu 
> [mailto:owner-statalist@hsphsun2.harvard.edu] On Behalf Of 
> Schaffer, Mark E
> Sent: 22 August 2007 18:35
> To: statalist@hsphsun2.harvard.edu
> Subject: st: RE: ivreg2 
> 
> Marta,
> 
> > -----Original Message-----
> > From: owner-statalist@hsphsun2.harvard.edu
> > [mailto:owner-statalist@hsphsun2.harvard.edu] On Behalf Of 
> > M.Sanchez-Sache@lse.ac.uk
> > Sent: 22 August 2007 18:09
> > To: statalist@hsphsun2.harvard.edu
> > Subject: st: ivreg2
> > 
> > I'm using ivreg2 command. It doesn't report an ANOVA table and, 
> > instead, it reports uncentered and centered TSS and r-squared. Why 
> > does it report uncentered TSS and r-squared if I'm including a 
> > constant in my regression and, therefore, centered TSS and r-squared

> > are valid? Why does it report both?
> > 
> > Thanks a lot in advance.
> > 
> > Marta.
> 
> It reports both because sometimes people want the centered 
> R-sq and sometimes they want the uncentered R-sq.  Seemed the 
> easiest way to program it and still give users statistics 
> they might find useful.
> 
> Cheers,
> Mark (co-author of ivreg2 along with Kit Baum and Steve Stillman)
> 
> Prof. Mark Schaffer
> Director, CERT
> Department of Economics
> School of Management & Languages
> Heriot-Watt University, Edinburgh EH14 4AS tel 
> +44-131-451-3494 / fax +44-131-451-3296
> email: m.e.schaffer@hw.ac.uk
> web: http://www.sml.hw.ac.uk/ecomes
> 
> 
> > 
> > Please access the attached hyperlink for an important electronic 
> > communications disclaimer:
> > http://www.lse.ac.uk/collections/secretariat/legal/disclaimer.htm
> > 
> > *
> > *   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/