Re: [Gretl-users] some problems... OK report + 4 simple questions
Dear Riccardo,
As I could guess LIMDEP do it
Individual effects consistent for logit, though biased.
Even for probit where they are inconsistent
they can be calculated. Greene gives so
simple algorithm for estimating fe probit
I did it in R several months ago in less
than 2 hours.
Oleh
P.S.
I have temporary access to stata 12
so I'll look what the can do
3 вересня 2015, 20:20:47, від "Riccardo (Jack) Lucchetti"
<r.lucchetti(a)univpm.it>:
On Thu, 3 Sep 2015, Allin Cottrell wrote:
>> 4) Is it possible to extract fixed effects stats from felogit.
>> Again, with mfx in mind
>
> Maybe Jack can answer that.
Uhhh, I don't think there's a strightforward way to do this: fixed-effects
logit uses a clever trick to zap the individual effects away from the
estimation problem (which, btw, uses a very special feature of the
logistic distribution). I guess it _should_ be possible to derive an
unbiased estimator for the individual effects, but I've never seen
anything like that used in practice. And besides, it wouldn't be
conceptually possible for those units for which the dependent variable is
all zeros or all ones. Maybe Claudia has more info.
-------------------------------------------------------
Riccardo (Jack) Lucchetti
Dipartimento di Scienze Economiche e Sociali (DiSES)
Università Politecnica delle Marche
(formerly known as Università di Ancona)
r.lucchetti(a)univpm.it
http://www2.econ.univpm.it/servizi/hpp/lucchetti
-------------------------------------------------------
_______________________________________________
Gretl-users mailing list
Gretl-users(a)lists.wfu.edu
http://lists.wfu.edu/mailman/listinfo/gretl-users