On 13.11.2019 00:14, Allin Cottrell wrote:
On Tue, 12 Nov 2019, Riccardo (Jack) Lucchetti wrote:
> Apart for my personal dislike for df corrections, which would make me
> prefer 0 for the default, I'd say it's perfect.
Thanks, Jack! I'm not too keen on df corrections myself, but I guess
n-1 is what most users would expect per default.
So, my three cents: very often I use list together with 'smpl' command
and sometimes I need mean and/or std for selected subsample but
sometimes for the whole sample. So I think that cdmean would have an
option for that.
Marcin
--
Marcin Błażejowski