Re: st: Re: heteroscedasticity in logit/ probit model

[Maarten buis <maartenbuis@yahoo.co.uk>]

--- On Wed, 24/11/10, rohaida wrote:
> I already use the hetprob command. But if i'm not mistaken the
> hetprob is a test for heteroscedasticity (correct me if I'm
> wrong). I get the hetprob result and confirm that my probit
> model suffers the endogeneity. So, what should I do next? Can
> I use the logit command with robust std. error to fix the 
> heteroscedasticity in my probit model?

No, heteroskedasticity in -probit-/-logit- models changes the 
scale of your dependent variable. This means that a regular
-logit- or -probit- will misspecify the means function so 
robust standard errors won't help as these assume a correctly
specified mean function. However, this assumes that you 
actually want to correct for heteroskedasticity. To answer that
you need to think very hard about what you want your dependent
variable to be, and what the dependent variable is in the 
different models. After that you can make your choice.

Hope this helps,
Maarten

--------------------------
Maarten L. Buis
Institut fuer Soziologie
Universitaet Tuebingen
Wilhelmstrasse 36
72074 Tuebingen
Germany

http://www.maartenbuis.nl
--------------------------


      

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/