Re: st: Negative R-squareds using xtivreg2

[John Antonakis <John.Antonakis@unil.ch>]

See this:

http://www.stata.com/support/faqs/stat/2sls.html

A negative r-square is not necessarily problematic (though it is likely 
that the model is misspecified).  You might want to use another kind of 
estimator for binary dependent variables.

Best,
J.

__________________________________________

Prof. John Antonakis
Faculty of Business and Economics
Department of Organizational Behavior
University of Lausanne
Internef #618
CH-1015 Lausanne-Dorigny
Switzerland
Tel ++41 (0)21 692-3438
Fax ++41 (0)21 692-3305
http://www.hec.unil.ch/people/jantonakis

Associate Editor
The Leadership Quarterly
__________________________________________


On 25.03.2012 12:53, Nick Cox wrote:
> Away from plain or vanilla regression, there are numerous different
> definitions of R-squared that need not give the same numeric answer.
> The help documents the flavo[u]rs reported and their different
> definitions. R-sq < 0 usually implies a lousy model in some sense;
> such results rarely make it to the literature.
>
> If your dependent variable is binary, it is not clear why you think
> this is a good command to use.
>
> Nick
>
> 2012/3/25 Rüdiger Vollmeier <ruediger.vollmeier@googlemail.com>:
>
>> I am running an IV regression accounting for country fixed effects
>> using the -xtivreg2- command. Stata reports R-squares. However, they
>> are surprisingly negative(!). Does anybody know why this is the
>> case?!?
>>
>> Should they be adjusted somehow? Is the fact that my dependent
>> variable is binary problematic?
> *
> *   For searches and help try:
> *   http://www.stata.com/help.cgi?search
> *   http://www.stata.com/support/statalist/faq
> *   http://www.ats.ucla.edu/stat/stata/

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/