Re: [Gretl-devel] gretl bug: ordered probit model
On Mon, 9 Oct 2017, Artur T. wrote:
Ok, I can't see any problem with this approach. However, apart
from "series x
= const(-p)" (for p>0) we should also consider the possible case
<hansl>
list L = const x y z
list Llag = L(-p)
# as well as
list Lllag = L(-p to -q) # p<q
</hansl>
Hm, that's right. Which brings us back to the idea that const(-k) should
be a no-op.
-------------------------------------------------------
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
-------------------------------------------------------