 |
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: st: RE: AW: Tobit, negative predictesd values
I have a similar problem, but in my case I think the reason is that I
was using predict xb, now I am using predict ystar(0,.) , which should
provide the values of the y observed and not of the latent y (which
can be negative): is that correct?
On Tue, Nov 10, 2009 at 3:46 PM, Solorzano Mosquera, Jenniffer
<[email protected]> wrote:
> Yes that is the paper. He uses a tobit censoring almost 50% of the sample! However he uses a tobit. I'm trying to replicate the same exercise he did but with the same kind of data but for Latinamerica.
>
>
> -----Mensaje original-----
>
>
> De: [email protected] [mailto:[email protected]] En nombre de Martin Weiss
> Enviado el: Tuesday, November 10, 2009 7:53 AM
> Para: [email protected]
> Asunto: st: AW: Tobit, negative predictesd values
>
>
>
> " a published paper by Zwick in 2002(Centre for European Economic Research)"
>
>
>
> You are probably referring to ftp://ftp.zew.de/pub/zew-docs/dp/dp0250.pdf ?
> Note the onus is on you to provide the reference...
>
>
>
>
> HTH
> Martin
>
>
> -----Ursprüngliche Nachricht-----
> Von: [email protected]
> [mailto:[email protected]] Im Auftrag von Solorzano Mosquera, Jenniffer
> Gesendet: Montag, 9. November 2009 23:59
> An: '[email protected]'
> Betreff: st: Tobit, negative predictesd values
>
> Hi everyone,
> I estimated a tobit model having intensity labor training as dependent variable and a group of firm characteristics which are presumpted as determinants of that intensity. I need to predict the estimated intensity to use it in a second stage model to know what is the impact in firm productivity. I'm using the whole sample, those firms offering training and those who are not. I'm censoring to the interval 0,1. However I obtain negative predicted variables from the tobit.
>
> I have 0's for those firms who decided not to offer training, following a published paper by Zwick in 2002(Centre for European Economic Research).
> That way allows not to lose observations from the firms who have missing values in the intensity variable because they decided not to offer training.
> However I've been looking and I found that heavy censoring causes these kind of problems on predicted values and even worse when high proportion of censored cases is the situation.
>
> Any suggestion to obtain tobit predicted values non-negatives or is it can be interpreted as a expected result?
>
> Thanks
>
>
>
> Jenniffer Solorzano Mosquera
> Labor Markets Division - Inter-American Development Bank Washington, DC - USA
> (202) 623-3659
>
>
> *
> * 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/
>
>
>
> *
> * 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/