Statalist


[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

RE: st: difference between -treatreg- and -ivreg-


From   "Schaffer, Mark E" <M.E.Schaffer@hw.ac.uk>
To   <statalist@hsphsun2.harvard.edu>
Subject   RE: st: difference between -treatreg- and -ivreg-
Date   Tue, 7 Aug 2007 21:11:07 +0100

Kyle et al.,

> -----Original Message-----
> From: owner-statalist@hsphsun2.harvard.edu 
> [mailto:owner-statalist@hsphsun2.harvard.edu] On Behalf Of Kyle Hood
> Sent: 07 August 2007 20:54
> To: statalist@hsphsun2.harvard.edu
> Subject: Re: st: difference between -treatreg- and -ivreg-
> 
> I think that ivreg and treatreg make different assumptions 
> about how the endogenous RHS variables are related to the 
> instruments.  In the former, the instruments are related 
> linearly to a continuous, perfectly observed endogenous 
> variable.  In the latter, instruments are linearly related to 
> an unobserved, latent continuous variable; if this variable 
> is above some threshold, the binary endogenous variable is 
> observed as a 1 and is 0 otherwise.

This part of the picture.  treatreg tries to get efficiency gains by
modelling the equation with the binary endogenous variable as a probit;
it is, in effect, a system estimator. ivreg (or ivreg2) is a
single-equation estimator.  The usual robustness-vs-efficiency tradeoff
applies: if the probit eqn is misspecified, treatreg will be
inconsistent, including the main equation; if it's correctly specified,
it's more efficient than IV.  But IV will be consistent in both cases.

For more details and some other options, see my other posting.

Cheers,
Mark

> On Tue, 7 Aug 2007, Prabal Kr. De wrote:
> 
> > Thanks, but let me be more specific:
> > what is the difference between -treatreg- and -ivreg?
> > Are they same in case the endogenous variable is binary?
> > Thanks again,
> > Prabal
> > --- Kyle Hood <kyle.hood@yale.edu> wrote:
> >
> >> Of the three commands you have asked about, only one allows for a 
> >> dependent variable that is discrete (ivprobit).  I believe 
> this would 
> >> be the appropriate choice.
> >>
> >> On Tue, 7 Aug 2007, Prabal Kr. De wrote:
> >>
> >>> hi!
> >>>  I wanted to know what is the difference between
> >>> -treatreg- and -ivreg- or -ivprobit- command. For 
> instance if I have 
> >>> a dummy dependent variable and
> >> a
> >>> dummy endogenous variable and a continuous
> >> instrument,
> >>> which one is more appropriate.
> >>> Thanks in advance,
> >>> Prabal
> >>>
> >>>
> >>>
> >>
> > 
> ______________________________________________________________________
> > ______________
> >>> Fussy? Opinionated? Impossible to please? Perfect.
> >>  Join Yahoo!'s user panel and lay it on us.
> >>
> > http://surveylink.yahoo.com/gmrs/yahoo_panel_invite.asp?a=7
> >>>
> >>> *
> >>> *   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/
> >>
> >
> >
> >
> >
> > 
> ______________________________________________________________________
> > ______________ Building a website is a piece of cake. Yahoo! Small 
> > Business gives you all the tools to get online.
> > http://smallbusiness.yahoo.com/webhosting
> > *
> > *   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/
> 

*
*   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/



© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index