Statalist The Stata Listserver

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

RE: st: RE: RE: Count data regression

From   "Rajesh Tharyan" <>
To   <>
Subject   RE: st: RE: RE: Count data regression
Date   Thu, 23 Mar 2006 17:40:48 -0000

Hi scott,

Thanks for the clarification... I think I have understood it now.. I will go
and have a look at the book..


-----Original Message-----
[] On Behalf Of Scott Cunningham
Sent: 23 March 2006 17:13
Subject: Re: st: RE: RE: Count data regression

On Mar 23, 2006, at 12:00 PM, Rajesh Tharyan wrote:

> Hi scott,
> Re your reply to hugh..
> 2. For the Poisson MLE, valid inference requires equality of the  
> conditional
> mean and variance (equidispersion) - it does not require that the  
> dependent
> variable have a Poisson distribution.
> I didn't get point two..  my impression was that it is the  
> distribution of
> the underlying variable which sort of dictated whether you used a  
> poisson
> regression or a nb regression..etc. but of course with a poission
> distributed variable you always have mean=variance..

(Different Scott talking now).

When you don't have equidispersion, you lose precision in your  
estimates, but you don't lose consistency.  So the Poisson is still  
consistent even if the conditional mean does not equal the  
conditional variance.  Thankfully, there are ways to correct the  
standard errors for overdispersion.  See the -poisson, robust- option  
and the -xtpoisson, vce(boot)-.  I don't have my Cameron and Trivedi  
MICROECONOMETRICS handy, but if you look in there, they talk about  
this correct procedure, and I think -poisson, robust- does this. -sc

*   For searches and help try:

*   For searches and help try:

© Copyright 1996–2020 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index