[R] zerinfl() vs. Stata's zinb

[Achim Zeileis]

On Mon, 12 Apr 2010, David Fortunato wrote:

> Hello,
>
> I am working with zero inflated models for a current project and I am
> getting wildly different results from R's zeroinfl(y ~ x, dist="negbin")
> command and Stata's zinb command. Does anyone know why this may be? I find
> it odd considering that zeroinfl(y ~ x, dist="poisson") gives identical to
> output to Stata's zip function.

It's hard to say anything useful about this without further information 
(and without having Stata available). Please provide a good example, e.g., 
the bioChemists data in "pscl" (which might also be available in Stata, 
it's a standard example in the literature). Check that ZIP is in fact 
identical and then check the differences between zinb and zeroinfl(). Post 
the code to replicate both along with the results.

I have checked zeroinfl against some of Stata's results but only printed 
in papers, manuals, etc. because I haven't access to Stata myself.

One difference might be that there are different versions of negative 
binomial distributions. The one implemented in zeroinfl() is also referred 
to as NB2.
Z

> Thanks,
>
> --david
>
> 	[[alternative HTML version deleted]]
>
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
>