in tobit with censoring at 0 you have 4 marginal effects:
(1) dE[y*|x]/dx
(2) dE[y|x,y>0]/dx
(3) dE[y|x]/dx
(4) dP[y>0|x]/dx
possible reference: greene (2003), pp.764-766
apparently you are familiar with (1) and (2), but the one i mean is number (3), which is provided by dtobit2 together with the other 3 (but not its s.e., which is different from the s.e. of 1 or 2)
> -----Ursprüngliche Nachricht-----
> Von: "Maarten buis" <[email protected]>
> Gesendet: 31.05.09 10:48:40
> An: [email protected]
> Betreff: Re: st: marg effects & s.e. in tobit
>
> --- On Sat, 30/5/09, John Bunge <[email protected]> wrote:
> > no, i mean the unconditional expected
> > value for y, not y*, which exists and is computed by
> > dtobit2, which however does not report the standard errors
> > of these coefficients
>
> Could you tell exactly what you want to condition on and what
> not, because in the terminology I am familiar with y is
> always conditional on being censored and y* is not.
>
> -- Maarten
>
> -----------------------------------------
> Maarten L. Buis
> Institut fuer Soziologie
> Universitaet Tuebingen
> Wilhelmstrasse 36
> 72074 Tuebingen
> Germany
>
> http://home.fsw.vu.nl/m.buis/
> -----------------------------------------
>
>
>
>
>
>
> *
> * 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/
>
*
* 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/
