Re: st: Comparing nested regression models

[Richard Williams <Richard.A.Williams.5@ND.edu>]

At 09:04 AM 7/2/2008, Tam Lin wrote:
> Thank you for this Maarten.
> I have already seen that help page.
> From what I understand, the Wald test that it recommends is
> designed to test the contribution of a single variable in a
> single model.
> What I want to do is compare one regression model with another.
> Will the Wald test do this?
> Thank you
Suppose model 1 has x1, x2 and x3.  Then model 2 adds x4 and 
x5.  After running model 2, just say

test x4 x5

Or, suppose model 1 has x1, x2 and x3.  In the constrained model, the 
betas for x1 and x2 are equal.  Run model 1 and then give the command

test x1 = x2

In short, so long as the models are nested, you estimate the 
unconstrained model, and then use the test command to see whether the 
unconstrained model is an improvement over the constrained model.

If in doubt as to whether you are imposing the constraints correctly, 
you can add the -coef- option to test, e.g.

test x1 = x2, coef


Weesie's -linest- command is also sometimes helpful.  Use -findit- to 
get the version stored at http://www.fss.uu.nl/soc/iscore/stata.

Here is some code that will illustrate my points.  The last lijne 
requires that linest be installed.

webuse nhanes2f
svy: ologit health female black rural weight
test female=rural, coef
constraint 1 female = rural
linest, c(1)


-------------------------------------------
Richard Williams, Notre Dame Dept of Sociology
OFFICE: (574)631-6668, (574)631-6463
HOME:   (574)289-5227
EMAIL:  Richard.A.Williams.5@ND.Edu
WWW:    http://www.nd.edu/~rwilliam

*
*   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/