The way I understand it, as of now:
Within subset dv2==1 (select variable),
dv1==1 always.
Question: Why is dv1 different from dv2
within this subset?
My thoughts:
I know you are trying to find out whether the
variables that impact selection also impact
scale (of dispute). Shouldn't you have a dv1 which
says something like
dv1 =1 if disputes escalate
dv1 =0 otherwise.
Otherwise, there is no way for the specf'n to
distinguish between the two variables (dv1 and dv2)....
Or is there something here that I am not seeing?
Jyotsna (Jo) Puri
Cell and Voicemail: 201-805-1690 [email protected]
----Original Message Follows----
From: Michael Horowitz <[email protected]>
Reply-To: [email protected]
To: [email protected]
Subject: RE: st: question about heckprob
Date: Wed, 18 May 2005 17:30:51 -0400 (EDT)
Ok. dv2 is a subset of dv1. All times dv2 is 1, dv1 is 1, but many of
the times where dv1 is 1, dv2 is not 1. Does that make sense? Should the
DV be set up differently?
On Wed, 18 May 2005, jyotsna puri wrote:
> Michael,
> Can you check/send a cross-tab of dv1 and
> dv2, i.e.
> table dv2 dv1
> Is dv1 varying within dv2==1? The way you've explained your
> variable construction below, it may not be.
> Considered using a Tobit?
> -Jo
>
> Jyotsna (Jo) Puri
> Cell and Voicemail: 201-805-1690
> [email protected]
>
>
>
>
> ----Original Message Follows----
> From: Michael Horowitz <[email protected]>
> Reply-To: [email protected]
> To: [email protected]
> Subject: st: question about heckprob
> Date: Wed, 18 May 2005 16:34:40 -0400 (EDT)
>
> To whom it may concern:
>
> I am trying to run a censored probit model in STATA but I am not sure
what
> the structure of the 2 dependent variables for the heckprob command needs
> to look like. The way my dependent variable is currently set up is that
> there is a 0,1 variable of whether or not states enter into disputes and
> then a 0,1 variable that only has values for those states that are in
> disputes and codes whether they escalated them (what I want to know is
> whether the same variables that influence whether or not states select
> into disputes also predict whether states will escalate disputes).
>
> But every time I try to use heckprob I get the following error message:
>
> Fitting probit model:
>
> outcome does not vary; remember:
> 0 = negative outcome,
> all other nonmissing values = positive outcome
> r(2000);
>
>
> Could someone please help me and tell me what the structure of a
> successful heckprob command is (I looked at the help file already) and
> more importantly, how the DVs for it are coded. The way my code is set
up
> is:
>
> heckprob dv1 iv1 iv2 iv3 iv4 iv5, sel(dv2 = iv4 iv6 iv7)
>
> Mike
>
>
*******************************************************************************
>
> Michael Horowitz
> 78 Kirkland Street, #1
> Cambridge, MA 02138
>
> *
> * 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/
>
