[Gretl-users] Re: Gretl: means by categorical variable
On Sun, 26 May 2024, g s wrote:
If I were to include gretl in my review of free to use, easy to
use, menu based statistical software,
https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4105959
Would it be okay if I indicated that gretl only lists 5 digits in
the summary command?
Yes, since that's true. But I'd suggest being a tad more precise: it
prints the results to 5 significant digits.
In the case of the mean population of Australia-Oceania, as Sven
pointed out, gretl actually stores a value of 1557590.272727. By
rounding to 1.5576 million we're in effect printing a value that's
good to the nearest hundred people, not to the nearest 10. I'd be
very surprised if any population tallies were good to the nearest
10, other than for very small communities.
Still, I can see a case for considering alternative ways of
printing some of the statistics for series composed of large
integers.
Allin Cottrell