Re: [Gretl-users] intreg doubt
On Sat, 5 Mar 2011, Allin Cottrell wrote:
On Sat, 5 Mar 2011, Maria Dolores Montoya Diaz wrote:
> I?d like to compare results of censored model (tobit) with OLS regression.
> For this, I?ve prepared the following hypothetical data, that I?ve copied
> below. The censoring point for y is 1000 [...]
>
> And I?d like to estimate intreg with upper limit of 1000 (equivalent to
> Stata?s command: tobit YCENS X, ul(1000)).
This doesn't sound like a case for interval regression. But you
can easily get tobit to do the job; just transform Y so it's
censored at 0 instead of at 1000.
It must be said that, in fact, the Tobit model is a special case of the
interval regression model. Try this:
<script>
nulldata 100
x = normal()
e = normal()
ystar = 1 + x + e
y = ystar>0 ? ystar : 0
tobit y const x
y_lo = ystar>0 ? ystar : NA
y_hi = ystar>0 ? ystar : 0
intreg y_lo y_hi const x
</script>
The above should be rather easy to generalise.
Riccardo (Jack) Lucchetti
Dipartimento di Economia
Università Politecnica delle Marche
r.lucchetti(a)univpm.it
http://www.econ.univpm.it/lucchetti