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

Re: st: 2SLS and Instrumental Variables


From   "Dana Shills" <[email protected]>
To   [email protected]
Subject   Re: st: 2SLS and Instrumental Variables
Date   Fri, 10 Dec 2004 14:29:35 +0000

Thank you Prof. Schaffer and Prof Williams.

I understand that the instruments might be collinear but I thought to be good instruments, there should be no direct paths from the instruments to the dependent variable Y1, only indirect paths through the endogenous regressor Y2. So if the instruments do come out significant, then are they not valid instruments? What am i missing here?

Thanks again

Dana


Perhaps because of the loss of degrees of freedom and/or
multicollinearity between the Zs?  Try a test of the joint
significance of Z1-Z4.

That is my first inclination too. To take a more extreme example, suppose you had a hundred yes/no questions. Their sum (i.e. the number of yes answers) might well have a significant effect. That does not mean that if you entered all 100 items in a regression that any one of them individually would have a significant effect.
_________________________________________________________________
On the road to retirement? Check out MSN Life Events for advice on how to get there! http://lifeevents.msn.com/category.aspx?cid=Retirement

*
* 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