st: re: Anderson-Rubin test in ivreg2

[Kit Baum <baum@bc.edu>]

<>
Grant wrote

Using IVREG2 for estimation, one of the tests is the "Anderson-Rubin  
Test" of joint significance of endogenous regressors in main equation,  
which gives p-values for both F-stat as well as Chi-sq. My question  
whether it matters which one is picked.


Grant should heed the Statalist FAQ and use informative words in his  
subject line.

It is up to you whether you want to report an F-statistic or a Chi- 
squared statistic for this test. The latter is the numerator of the  
former. Just as one can choose between small-sample t stats (with the  
'small' option) and large-sample z stats, and in using -test- can get  
an F-statistic after -regress- but a Chi-squared statistic after - 
logit-, you can report either form of the A-R statistic. They will  
have similar p-values.

Kit Baum, Boston College Economics and DIW Berlin
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/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/