My source on these issues is the very nice paper by Winship and Mare in
ASR 1984 in which they suggest one identify the probit model by setting
the variance of Y to equal 1. Why this is not a simple switch in every
program *still* escapes me (--or is it now a simple switch, and it's
introduction escaped me?) At any rate, with this normalization lots of
challenges with comparing models (e.g., "R^2", coefficients) are greatly
eased or reduced. I have it programmed somewhere "after the fact" of
estimation, if memory serves it is simple to do.
Winship, Christopher, and Robert D. Mare. 1984. "Regression Models with
Ordinal Variables." *American Sociological Review* 49: 512-525.
Sam
On Tue, 23 Jan 2007, Herb Smith wrote:
> And a second thought.... the comparison of coefficients across logit
> equations can be tricky for reasons having nothing to do with getting the
> right standard errors. Even if the dropped variables are uncorrelated
> with the regressor that you are interested in, the value of the
> coefficient for that regressor may change since there is no set variance
> for the latent variable being mapped onto the observed dichotomy (and
> identification is typically achieved by setting the error variance at 1 in
> both equations...) So even if you in some sense reject the hypothesis
> that a coefficient for a given variable is the same in two different
> equations, it is tough to rule out the possibility that the coefficient
> values might not be the same under alternative normalizations. Cameron
> and Heckman 1998 hammer on this point from a slightly different angle, and
> see the recent papers by Hauser and Mare in *Soc Meth 2006*
>
> Professor of Sociology and
> Director, Population Studies Center
> 230 McNeil Building
> 3718 Locust Walk CR
> University of Pennsylvania
> Philadelphia, PA 19104-6298
>
> [email protected]
>
> 215.898.7768 (office)
> 215.898.2124 (fax)
>
> On Tue, 23 Jan 2007, Herb Smith wrote:
>
> > Wendy,
> >
> > Ah, takes me back to a discussion with Cliff in 1989 about this.
> > He posed the problem of a test. I said, "Why not bootstrap it?" But
> > Cliff was a real statistician, and wanted a formal solution....
> >
> > Alas, Cliff is no longer with us and can't defend himself, but I
> > still bootstrap in these situations. That is, estimate both equations,
> > calculate the difference between the coefficients... and then do it over
> > and over again, with new samples. Too fogged right now to think through
> > the three or four lines necessary, but if no one comes up with it by
> > tomorrow I will find where I have done it before or make up a new one.
> >
> > I would be interested, too, however, if anyone has programmed the
> > ideas in the *AmJrnlSoc* paper. These were expanded on by Sobel in *Soc
> > Meth 1996* in an encomium to Clogg....
> >
> > Best,
> >
> > --Herb
> >
> > Professor of Sociology and
> > Director, Population Studies Center
> > 230 McNeil Building
> > 3718 Locust Walk CR
> > University of Pennsylvania
> > Philadelphia, PA 19104-6298
> >
> > [email protected]
> >
> > 215.898.7768 (office)
> > 215.898.2124 (fax)
> >
> > On Tue, 23 Jan 2007, Wendy Manning wrote:
> >
> > > I would like to compare coefficients - on the same independent variable -
> > > between two nested logistic regression models using STATA. I would like to
> > > test whether the two coefficients are significantly different from one
> > > another. In SAS this can be done using the proc iml commands. I was
> > > wondering if anyone has worked up some STATA commands. Clogg et al. 1995 in
> > > AJS describe the statistics.
> > >
> > > *
> > > * 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/
> >
> *
> * 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/
