No, it isn't that much different: you make a prediction using the
vector of 1's times the coefficient from the first stage, and you
don't have a constant in your second regression. So you just need to
have _one's and nocons sorted out properly.
On 6/15/07, [email protected] <[email protected]> wrote:
Dear statalisters,
I red both the Hardin's and Hole's articles, but
I
need an help for the following two step model:
First step:
y=alpha+beta*x+e
Second step:
z=gamma*alphahat+b'x'+e'
where the
parameter of interest is gamma. If I use this 2-step model it'll
result
in an underestimation of the errors, so I need to use the
Murphy and
Topel estimation for the standard errors. This is
different from what
I saw in the article, so may you help me please?
Thank you
M.
*
* 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/
--
Stas Kolenikov
http://stas.kolenikov.name
*
* 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/
