Re: st: When to use Poisson or Negative Binomial

[Maarten Buis <maartenlbuis@gmail.com>]

On Fri, May 27, 2011 at 1:34 PM, Querze, Alana Renee <arq@ku.edu> wrote:
> @Maarten: what else do I need to know besides the marginal mean and variance to choose the right command? I chose between random and fixed effects by using the Hausman test... is there a similar test in STATA that would help me know which model better fits the data?

One thing you can do look at the logic underlying each model. In this
case they would imply different stories for what is happening at the 0
counts. Compare those stories with what you know about your problem
and decide which one makes more sense for your situation.

Another thing you can do is look at the residuals.

I do not think that statistical tests are the right way for selecting
models, the logics underlying statistical testing and model selection
are just too different. When it comes to model selection we want our
model to be useful simplification of reality, which paradoxically
means that true models are either bad models (they don't simplify
enough) or useless (the situation does not require simplification).
The tradeoff is all about simplifying enough so that we understand the
results and not simplifying too much so we won't make too many errors.
Statistical testing is all about the probability of rejecting a true
statement. The tradeoff is all about the probability of rejecting a
true statement and the probability of not rejecting a false statement.
It is hard to see how one tradeoff can say something useful about the
other.

Hope this helps,
Maarten

--------------------------
Maarten L. Buis
Institut fuer Soziologie
Universitaet Tuebingen
Wilhelmstrasse 36
72074 Tuebingen
Germany


http://www.maartenbuis.nl
--------------------------

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/