Statalist


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

st: RE: Test of intercepts across OLS regressions


From   "Lachenbruch, Peter" <Peter.Lachenbruch@oregonstate.edu>
To   <statalist@hsphsun2.harvard.edu>
Subject   st: RE: Test of intercepts across OLS regressions
Date   Thu, 6 Nov 2008 08:09:53 -0800

At the risk of answering something that has already been answered, this
sounds a lot like analysis of covariance:  If you can assume the slopes
are the same across the regressions, you have
xi: anova y X i.group where group indicates the regressions.  The test
of i.group=0 tests the common intercept.  You can also try noconstant to
get the individual intercepts.  For non-common slopes you would want
i.group*X in the model

Tony

Peter A. Lachenbruch
Department of Public Health
Oregon State University
Corvallis, OR 97330
Phone: 541-737-3832
FAX: 541-737-4001

-----Original Message-----
From: owner-statalist@hsphsun2.harvard.edu
[mailto:owner-statalist@hsphsun2.harvard.edu] On Behalf Of Jared Delisle
Sent: Wednesday, November 05, 2008 7:41 PM
To: statalist@hsphsun2.harvard.edu
Subject: st: Test of intercepts across OLS regressions

I asked this question a few weeks ago, with no response.  So I'll give
it another try and be more explicit, since I was vague in the first
post.  Here goes...  I have a set of regressions, let's say 10 separate
regressions, all over the same time period, all with the same
specification: Y(n) = a + bX, where n=1..10 (designating which vector of
Y will be the dependent variables), and X is the same vector of
regressors for all regressions.  I am interested in testing if the
intercepts, a1...a10, are jointly equal to zero.  I can do this with a
-sureg- and -test _cons- which gives me a Wald Chi-square statistic.  I
need to perform a Gibbons, Ross, & Shanken (Econometrica 1989) test,
which yields an F-statistic, and is essentially (if I interpret it
correctly) a Hotelling's T-squared statistic times (T-N-1)/[N(T-2)]
where T is the number of time series observations and N is the number of
intercepts (10 in my case).  Is anyone familiar with performing such an
analysis?  Is there a .do file floating around that calculates this GRS
F-statistic, or any suggestions on how to write one up?

Thanks for your assistance,
Jared

*
*   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/



© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index