Re: st: Heterocedasticity

 From Richard Williams To statalist@hsphsun2.harvard.edu Subject Re: st: Heterocedasticity Date Thu, 27 Jul 2006 18:26:43 -0500

```At 11:17 AM 7/27/2006, Armando Tarupi wrote:
```
```help me please, I need make a test for heterocedasticity for the probit
model, thanks
```
I'm not aware of a nice simple post-estimation command, although there might be one out there. If you have specific hypotheses about sources of heteroskedasticity, you could run nested models and do chi-square contrasts using either probit/hetprob or the user-written oglm (which is slower but more flexible and powerful). For example, to test for hetero with price using the auto data, we could do something like

. sysuse auto
(1978 Automobile Data)

. quietly probit foreign weight length price

. est store probit

. quietly hetprob foreign weight length price, het(price)

. est store hetprob

. lrtest probit hetprob, stats force

Likelihood-ratio test LR chi2(1) = 4.29
(Assumption: probit nested in hetprob) Prob > chi2 = 0.0384

------------------------------------------------------------------------------
Model | Obs ll(null) ll(model) df AIC BIC
-------------+----------------------------------------------------------------
probit | 74 -45.03321 -17.73612 4 43.47223 52.68849
hetprob | 74 . -15.59338 5 41.18676 52.70708
------------------------------------------------------------------------------

It might be tempting to just include all the explanatory vars in the hetero equation, but I've found that that often produces very unstable results, i.e. the models have trouble converging and different programs will give you different results.

Doing the same thing with oglm,

. sysuse auto
(1978 Automobile Data)

. quietly oglm foreign weight length price, link(p) store(probit)

. quietly oglm foreign weight length price, link(p) store(hetprob) het(price)

. lrtest probit hetprob, stats

Likelihood-ratio test LR chi2(1) = 4.29
(Assumption: probit nested in hetprob) Prob > chi2 = 0.0384

------------------------------------------------------------------------------
Model | Obs ll(null) ll(model) df AIC BIC
-------------+----------------------------------------------------------------
probit | 74 -45.03321 -17.73612 4 43.47223 52.68849
hetprob | 74 -45.03321 -15.59338 5 41.18676 52.70708
------------------------------------------------------------------------------

-------------------------------------------
Richard Williams, Notre Dame Dept of Sociology
OFFICE: (574)631-6668, (574)631-6463
FAX: (574)288-4373
HOME: (574)289-5227
EMAIL: Richard.A.Williams.5@ND.Edu
WWW (personal): http://www.nd.edu/~rwilliam
WWW (department): http://www.nd.edu/~soc
*
* 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/