Statalist


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

Re: st: testing for the validity of instruments in probit model estimated via ML


From   "Austin Nichols" <austinnichols@gmail.com>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: testing for the validity of instruments in probit model estimated via ML
Date   Tue, 5 Aug 2008 11:38:52 -0400

Alejandro--
I think you would want to report the output from a linear probability
model estimated with -ivreg2- alongside your -ivprobit- output (not
-ivtobit- right?) to reassure readers that the instruments pass all
the usual tests for the linear case, even if equivalent tests do not
yet exist for nonlinear models.

On 8/4/08, Alejandro Delafuente <alejandro.delafuente@sant.ox.ac.uk> wrote:
> Dear statalisters,
>
> Am estimating a probit model with one continuous endogenous regressor and four
> instruments via maximum likelihood estimation (i.e. ivtobit command).
>
> Am reporting the Wald statistic that appears at the end of the routine to test
> for exogeneity, but was wondering if this would suffice or do I need to report
> any other test statistic?
> I know that ivreg2 reports several tests. Yet, what would be an adequate test
> to report with ML estimation, if necessary? And also, what would such
> statistically significant test statistic would indicate with respect to the
> validity of the instruments used?
>
> Many thanks,
>
> Alejandro
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/



© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index