Re: [Gretl-users] panel estimation based on restricted sample