[Limdep Nlogit List] Controlling for number of parameters in model selection tests

[Mikołaj Czajkowski]

William Greene wrote:
> Mikołaj. They should not be counted as parameters. They are simple
> functions of other parameters already in the model. The counterpart
> would be counting the precision = 1/sigma in an ordinary regression
> as an additional parameter. Obviously, it's not.
> /Bill Greene

Dear prof. Greene,
Thank you very much for your help (again).

-- 
     Mikołaj Czajkowski

     Warsaw Ecological Economics Center
     Warsaw University
     http://www.woee.pl/