Stata The Stata listserver
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

RE: st: question on chow test


From   "Daniel Schneider" <daniel.schneider@stanford.edu>
To   <statalist@hsphsun2.harvard.edu>
Subject   RE: st: question on chow test
Date   Sat, 19 Feb 2005 23:07:42 -0800

I've got a follow-up on this one:

My model looks like this:

regress y x1 g2 g3 g4 g2x1 g3x1 g4x1 g2x2 g3x2 g4x2 x3 x4 

I have more than one group and I have more than one variable that differ
and some control variables that should not vary across groups.

I know that with

testparm g2 g3 g4 g2x1 g3x1 g4x1 g2x2 g3x2 g4x2

I can test if the variables all together do vary across the groups.

But how can I test if, for example, the subgroup of x1 is differing
across groups. Is

testparm g2x1 g3x1 g4x1

Is this the correct way to do the test?

Secondly, can I use the differences between 

g2x1 <-> g3x1 <-> g4x1

to describe the "strength" of the differences (amount of difference, in
the same meaning as effect strength compared to statistical
significance)? 

Daniel

> -----Original Message-----
> From: owner-statalist@hsphsun2.harvard.edu 
> [mailto:owner-statalist@hsphsun2.harvard.edu] On Behalf Of 
> Richard Williams
> Sent: Saturday, February 19, 2005 9:27 PM
> To: statalist@hsphsun2.harvard.edu
> Subject: Re: st: question on chow test
> 
> 
> At 11:27 PM 2/19/2005 -0800, jiang.107 wrote:
> >Hi all, any expert in chow test? I've got a question and would 
> >appreciate
> >any insight--
> >
> >This is from the STATA FAQ:
> >to estimating the separate models
> >     y = a1 + b1*x1 + c1*x2 + u         for group == 1
> >     y = a2 + b2*x1 + c2*x2 + u         for group == 2
> >
> >     . gen g2 = (group==2)
> >     . gen g2x1 = g2*x1
> >     . gen g2x2 = g2*x2
> >     . regress y x1 x2 g2 g2x1 g2x2
> >
> >
> >My problem is that I've got lots of control variables on the 
> right hand
> >side(x1,x2....x15). I'm only interested in x1 and I wanna 
> test if the 
> >coefficient of x1 is significantly different for the two groups.
> > From the STATA FAQ, it looks like I need to interact each 
> variable with 
> > the group dummy, which is awkward with too many control 
> variables in my 
> > case when I only wanna test for x1.
> 
> Are you sure you want to let the effects of all the Xs differ across 
> groups?  If not, you could just do
> 
> regress y x1 g2 g2x1 x2 x3 x4 etc.
> 
> The T value for g2x1 will tell you whether the slopes differ across 
> groups.  Use
> 
> test g2 g2x1
> 
> if you want to see whether both the slopes and intercept differ.
> 
> If you want to let the effects of all variables differ across 
> groups, the 
> -xi- command will save you a little typing.
> 
> *
> *   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/



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