Statalist


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

Re: st: 2SLS with probit in the first stage regression


From   [email protected]
To   [email protected]
Subject   Re: st: 2SLS with probit in the first stage regression
Date   Sat, 16 Feb 2008 12:09:44 +0100

-ivreg2- and thus -xtivreg2- both from SSC have a lot of tests. Have a look at them...

P.S. "grouped" makes me think about a cluster option rather than a panel command
P.P.S. don't forget to have a lot at -cmp- (from ssc, too), which lets you combine several commands (and thus e.g. to allow different instruments for different endogenous variables) in order to make some 2SLS-like estimations

Nicola

At 02.33 15/02/2008 -0500, Renuka Metcalfe wrote:
>Hi
>
>I would like to find out if training which is one of
>my RHS variable is an endogenous variable. The
>original regression was a pay equation as follows:
>
>.xtreg y training meanworkplacetraining x2 x3 x4 
>
>I used xtreg as the data are grouped across workplaces
>and considered to be a less biased OLS estimator.
>Training is a binary variable, dummy=1 if they have
>had training and 0 otherwise and the data is
>cross-section data. I left out the meanworkplace
>training for now. I want to deal with one endogenous
>variable at a time.
>
>I did:
>
>.probit training x2 x3 x4 z
>
>where z = my instrument
>
>I then did:
>
>predict ghat
>
>I then issued:
>
>.ivreg pay x2 (training = ghat) x3 x4 
>
>Part of what I got is:
>
>Number of obs =   14321
>      
> F( 58, 14262) =  115.93
> Prob > F      =  0.0000
> R-squared     =  0.3104
> Adj R-squared =  0.3076
> Root MSE      =  .49362
>
>         Coef.   Std. Err.   t  P>|t|   
>- -------------+----------------------------------------
>training |.2904799 .116244 2.50  0.012
>
>Can I take interpret that the training variable is not
>endogenous? I would be grateful if anyone would tell
>me if I have done it correctly to find out if my
>training variable is endogenous and if it is
>endogenous. If I have done it incorrectly what would
>be the correct way to go about it.

*
*   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