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: which -cmp- option to use for poisson model with count data?

From   Nick Cox <>
Subject   Re: st: which -cmp- option to use for poisson model with count data?
Date   Wed, 2 May 2012 17:06:38 +0100

By "binomial" I don't mean "negative binomial". I mean binomial.

David Roodman already I think signalled clearly that -cmp- is out of
the question for your problem.

When people respond to questions, it is sometimes relatively easy to
spot mistakes and misunderstandings and to correct them. Knowing what
somebody should best do, however, often requires enormously more
information to advise well.


On Wed, May 2, 2012 at 4:44 PM, Laura R. <> wrote:
> Ok, so I leave the variable ranging from 0 to 5, and 0 would be
> included as an outcome in -poisson-
> However, as you (all) suggested, I should neither use -poisson- nor
> -ologit-. Which model would you suggest?
> Nick, do you mean I should use a negative binomial model? Why? And how
> do I do that in Stata? And in -cmp-?
> By the way, there is an example in Statalist archives, where the
> number of kids is used in the -cmp- model, and it is treated as
> -oprobit- (if I am right that the variable "kids" is the number of
> kids).
> Isn't my number of experts similar to the number of kids? It cannot be
> less than 0, but it is usually not more than 8 or so.
> (this is the example:
*   For searches and help try:

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