For what it's worth, the only things I am aware of that one can do
easily in STATA is the box-cox transformation (-boxcox-) and/or try
adding polynomial terms and testing their joint significance. The most
robust test, again that I am aware of, would entail estimating a
parametric model and something like a semiparametric partially linear
regression model and doing a specification test to see if you reject the
parametric model (see, e.g., the paper on my web cite with List and
Stengos, forthcoming in Rev of Eco and Statistics).
Dann
----------------------------------------------------------
Daniel L. Millimet, Assistant Professor
Department of Economics
Box 0496
SMU
Dallas, TX 75275-0496
Phone: 214.768.3269
Fax: 214.768.1821
http://faculty.smu.edu/millimet
----------------------------------------------------------
> -----Original Message-----
> From: Mwale, McDonald [mailto:[email protected]]
> Sent: Thursday, May 01, 2003 6:16 AM
> To: [email protected]
> Subject: st: testing non-linearity
>
> Dear statalisters
>
> I am Using Stata 7. How do I test whether the relationship between
the
> dependent variable and the independent variable is linear or
non-linear. I
> stand to be corrected but I think "testnl" in stata tests
relationships
> between the independent variables only. One idea that I have is to
include
> the square of the independent variable in the model specification and
see
> if
> it takes a different sign from its level. The questions: (a) can this
be
> reliable (b) is there any robust test. Thanks for reading.
>
> Macdonald
>
> *
> * 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/
