Bookmark and Share

Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at

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

Re: st: IV estimation in a negative binomial framework

From   Austin Nichols <[email protected]>
To   [email protected]
Subject   Re: st: IV estimation in a negative binomial framework
Date   Wed, 10 Aug 2011 10:51:05 -0400

Arka Roy Chaudhuri <[email protected]>
-glm- with a log link does not need an assumption that the conditional
variance is equal to the conditional mean to be consistent, nor does
its IV analog.  See -ivpois- on SSC and

On Tue, Aug 9, 2011 at 5:03 PM, Arka Roy Chaudhuri <[email protected]> wrote:
> Hi,
> I am analyzing a model where my dependent variable is a count
> variable, specifically the number of riot cases in a district. Since
> the variance of the dependent variable is quite high compared to the
> mean I plan to use the negative binomial distribution. However my
> independent variable is endogenous and I would like to use a
> instrument variable analysis. Unfortunately I have been unsuccessful
> in finding any program which implements IV in a negative binomial
> context. Could somebody please advice me how to implement IV
> estimation in a negative binomial framework? I would really appreciate
> any help in this regard.
> Thanks,
> Arka
*   For searches and help try:

© Copyright 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index