# st: Re: Re: Re:Re:xttobit

 From "Marina.Balboa" <[email protected]> To <[email protected]> Subject st: Re: Re: Re:Re:xttobit Date Thu, 17 Jul 2003 20:28:34 +0200

```Dear statalist,

sorry for my last mail, but it seems that it is not possible to see the
formulas I have written (mathtype). I rewrite the mail:

Now,
I would like to be sure about just one thing. In the Tobit model, you could
be interested in the following marginal effects:

1) d(y*|x)/dx = beta

2) d(y|x)/dx = beta*cumulative distribution(x*beta/sigma)

3) d(y|x,y>0)/dx =
beta*[1-(xbeta/sigma)Landa(xbeta/sigma)-Landa^2(xbeta/sigma)]

where y* is the latent variable and y is the censored variable. I just have
two questions:

i) Are the coefficients that Stata gives in xttobit  the ones numbered as 1?
(the marginal effect of the latent variable). I suppose it is this way. If
this were true, I could calculate 2 and 3 from 1 by hand.

ii) Is there any other way to obtain 2 and 3 directly without doing by had?
(mfx compute?)

Thank you to all and sorry for the aspect of the formulas.

>
> > Thanks Scott!
>
>
> > > Hello!
> > >
> > > I would be so happy if someone could help me. I am estimating a tobit
> > panel data model and I would like to know if it is possible to calculate
> > marginal effects in xttobit.
> > >
> > > Thanks in advance.
> > >
> > > Paula Garcia
> > >
> >
> > Yes,  take a look at -mfx compute-
> >
> > With it you can calculate the margin effects of the probability of being
> > uncensored, conditional on being uncensored, and the unconditional
> expected
> > value.
> >
> >
> > Hope this helps,
> > Scott
> >
>

*
*   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/
```