Maj Ruter said
Thanks to everyone who responded. What I am trying to do is perform a
specification test to verify that I have correctly modeled a binary
choice probit. The probit is modeling a cost overrun as a function of
the number of times a contract budget has changed, contract length,
industry concentration, change in R&D budget, change in procurement
budget, inflation, and the technology readiness of the commodity.
My thesis advisor has asked me to calculate the D-W statistic for my
model, but, after everything I've read and the responses I received
here, it does not seem like an appropriate test for my case.
What you are looking for (and perhaps your advisor was thinking of)
is a test of the independence of the errors. For a binomial probit
model, you are assuming that the errors are properly modeled as
Normal iid. Tests for normality of the errors -- for instance, for
the 3d and 4th moments (Jarque-Bera and the like) would seem
appropriate, as well as general tests of the distribution (Q-Q plots?
Kolmogorov-Smirnov tests?) as well as tests for heteroskedasticity.
Estimation with hetprob might be useful.
Kit Baum, Boston College Economics
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/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
