Re: st: RE: RE: RE: proportion as a dependent variable

[Todd Wagner <twagner@stanford.edu>]

I think you are referring to the fractional logit (flogit).  These models 
are discussed (briefly) in Wooldridge's book.  Also there is a talk by John 
Mullahy, see http://lhcwww.soph.uab.edu/presentation.pdf.

It is a relatively simple model and can be run with Stata.

glm <dv>  <iv's> , link(logit) family(binomial) robust

As Mullahy's talk shows, it is critical to use the robust option.

Todd



At 04:56 PM 7/14/2003 +0100, Nick Cox wrote:

> Joao Pedro W. de Azevedo
> >
> > I've seen papers in which the authors used a tobit model
> > censored at 0 and 1
> > to model a proportion.
> > Would this be an acceptable approach?
> >
>
> Without knowing about the data or the underlying science,
> it is difficult to say. But it does not sound like a
> general solution to me.
>
> Nick
> n.j.cox@durham.ac.uk
> *
> *   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/