RE: st: Instrumental Variable Method and Probit

[Roger Harbord <roger.harbord@bristol.ac.uk>]

--On 20 January 2005 Roger Harbord <roger.harbord@bristol.ac.uk> wrote to 
Statalist:


> > Allowing for clustering should be possible with -qvf- (see -findit qvf-).
> > This does IV for generalized linear models (including probit) and has a
> > cluster() option.
--On 21 January 2005 Scott Merryman <smerryman@kc.rr.com> wrote to Statalist:

> Roger,
>
> In July 2003, I had made similar suggestion of using -qvf- to correct for
> endogeneity and Vince Wiggins kindly pointed out to me that IV for
> measurement error is different than IV for endogeneity and employing -qvf-
> when you have an endogenous regressor will yield inconsistent estimates.
I've found Scott's message from July 2003 on the Statalist archive:
<http://www.stata.com/statalist/archive/2003-07/msg00434.html>

However the abstract of the article by Hardin, Schmiediche and Carroll about 
-qvf- in Stata Journal 3(4) p351-360 appears to differ:

"One of the most notable features of the qvf command is its ability to 
include instrumental variables. This functionality was added specifically to 
address measurement error but may be utilized by the user for other purposes."
<http://www.stata-journal.com/abstracts/st0049.pdf>

Could someone please clarify whether qvf can estimate models where 
instrumental variables are used to deal with endogeneity? If not i've got 
some serious corrections and apologies to make sharpish!

Many thanks,
Roger.

--
Roger Harbord
Department of Social Medicine, University of Bristol
http://www.epi.bris.ac.uk

*
*   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/