Stata The Stata listserver
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

SV: st: treatreg


From   "Kjerstad Egil (SNF)" <[email protected]>
To   <[email protected]>
Subject   SV: st: treatreg
Date   Tue, 31 Aug 2004 11:32:46 +0200

Mark, 

 Thank you so much for recommending a possible alternative estimation method. We are in the process of trying it out.

Best regards
Egil 
> ----------
> Fra: 	Mark Schaffer[SMTP:[email protected]]
> Svar til: 	[email protected]
> Sendt: 	23. august 2004 17:37
> Til: 	[email protected]
> Emne: 	Re: st: treatreg
> 
> Egil,
> 
> Subject:        	st: treatreg
> Date sent:      	Thu, 19 Aug 2004 12:14:54 +0200
> From:           	"Kjerstad Egil (SNF)" <[email protected]>
> To:             	<[email protected]>
> Send reply to:  	[email protected]
> 
> > Hi,
> > 
> > I am a 7.0 user and seek to find a way to expand the scope for the
> > treatreg command. First challenge: The treatment effects model
> > estimates the effect of an endogeous binary treatment on another
> > endogenous, continuous variable, conditional on two sets of
> > independent  variables. How can the model be altered to allow for,
> > say, three treatment categories?
> 
> I don't think the code for a modified -treatreg- is out there, so you 
> might want to consider the following alternative, discussed in 
> Wooldridge's 2002 book in chapter 18.
> 
> (a) Estimate the three "first-stage" probits by hand, i.e., with the 
> treatment categories as dependent variables in each of the probits.
> 
> (b) For each probit, obtained the predicted probability for the 
> binary treatment.  This gives you 3 sets of fitted values.
> 
> (c) Estimate your treatment equation using instrumental variables, 
> where the 3 sets of fitted values are instruments (NOT regressors!) 
> for the three treatment variables.
> 
> This procedure is more efficient than using simple IV.
> 
> What's very convenient about this is that the standard errors 
> generated by this IV procedure are correct - no need for messy 
> corrections of the var-cov matrix.
> 
> You'll probably want to have at least 3 different exclusion 
> restrictions so that the equation is reasonably well identified.
> 
> Hope this helps.
> 
> --Mark
> 
> > Second challenge: What if the
> > endogenous, continuous variable is a covariate in the expression for
> > another endogenous, continuous variable? 
> > 
> > Do such amendments to the treatreg command already exist 'out there'?
> > If so, are they available?  
> > 
> > Best regards
> > 
> > Egil
> > 
> > *
> > *   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/
> 
> Prof. Mark E. Schaffer
> Director
> Centre for Economic Reform and Transformation
> Department of Economics
> School of Management & Languages
> Heriot-Watt University, Edinburgh EH14 4AS  UK
> 44-131-451-3494 direct
> 44-131-451-3008 fax
> 44-131-451-3485 CERT administrator
> http://www.som.hw.ac.uk/cert
> 
> *
> *   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 1996–2024 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index