[Limdep Nlogit List] Model selection Criteria

Fred Feinberg feinf at umich.edu
Wed Mar 5 05:32:32 EST 2008

Good point. I believe this would give the same outcome as using AIC or BIC, because
each of those penalizes the likelihood by a function of the number of parameters
and the amount of data, which remain constant here across models.


Noel Roy wrote:

> I would add here the Pollock-Wales paper (JEcon 47 (1991)) on the
> likelihood dominance criterion. On this criterion, if two non-nested
> models have the same number of parameters and same dependent variable
> (or, if the dependent variables are functionally related, the
> appropriate Jacobian is incorporated into the likelihood function),
> the model with the highest log-likelihood dominates.

More information about the Limdep mailing list