[Gretl-devel] Re: RE probit and perfect prediction

On Sat, 7 Nov 2020, Allin Cottrell wrote: >> I wonder if it could really be as easy as this. How come we were flagging >> an error in that case? It feels as if I'm missing something. Allin, do you >> remember why we were setting err to E_DATA? > > Not really. What happens if we just proceed: do we use the reduced mod.list > or try the full list for the panel estimator? We do use the reduced list, as I think we should. Therefore, the fix would just be the one-liner I posted before. But again, it looks to good to be true.