Statalist


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: st: Multiple Endogenous Variables in Count Data Model (Poisson or Negative Binomial).


From   "Austin Nichols" <austinnichols@gmail.com>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: Multiple Endogenous Variables in Count Data Model (Poisson or Negative Binomial).
Date   Wed, 12 Nov 2008 14:54:28 -0500

ippab <ippab.statalist@gmail.com> :
Short answer--yes, include two residuals.  Or
ssc install ivpois, replace
help ivpois

On Wed, Nov 12, 2008 at 2:17 PM, ippab ippab <ippab.statalist@gmail.com> wrote:
> Hi All,
>
> I have a negative binomial model with 2 endogenous variables.  I also
> have two instruments.  However, I am not totally sure how we can test
> for endogeneity in this situation.
>
> I know that Wooldridge (2002) in Section 19.5.1 outlines a two-step
> approach for testing endogeneity with one endogenous variable.
> According to that approach, we need to calculate the residual from OLS
> in the first stage and include it in the original model.  Can we
> extend it to two endogenous variables -- calculate two residuals
> (include all exogenous variables and two instruments in each OLS) and
> estimate the original Poisson/negative binomial model with the
> residuals.
*
*   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