On Fri, Mar 1, 2013 at 9:33 AM, Maarten Buis <maartenlbuis@gmail.com> wrote:
>
> A zero inflated negative binomial tries to extract quite a lot of
> information from quite subtle features in the data and the same is
> true for panel models. I would not be surprised that combining the two
> would lead to a model that would just be too fragile to be of
> practical use.
Yeah, that was my thought. I suspect you could get it to work if you
had good a priori information predicting zero inflation, but without
that it would be very difficult given how much zero inflation and
negative binomial mixing are likely to trade off.
--
JVVerkuilen, PhD
jvverkuilen@gmail.com
"It is like a finger pointing away to the moon. Do not concentrate on
the finger or you will miss all that heavenly glory." --Bruce Lee,
Enter the Dragon (1973)
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/faqs/resources/statalist-faq/
* http://www.ats.ucla.edu/stat/stata/
