[Rd] default for 'signif.stars'
Lenth, Russell V
ru@@e||-|enth @end|ng |rom u|ow@@edu
Sun May 10 22:23:01 CEST 2026
Hello,
I'm following up on a thread from 7 years ago (Mar 2019) where there are several arguments -- some of them passionate -- for and against having the 'show.signif.stars' option default to TRUE. You can look back in the archives (https://stat.ethz.ch/pipermail/r-devel/) under "default for 'signif.stars'", starting 27 March 2019 to see what was said. You may also want to look at the "Regression stars" thread initiated by Frank Harrell starting 7 Feb 2013.
Clearly, there is a long-standing debate on this issue. I'm writing to suggest a more modest response than changing the default for 'show.signif.stars': Instead, consider making 'show.signif.stars' a configuration option along with 'papersize' and others that some users will want to change. The default .Rprofile script starts with "# Things you might want to change" - and it could be added there. I think this file is initialized with some modal dialogs in a new installation of R, and it would be good to add it there too.
Why do this? Because there are a lot of users out there who strongly object to significance stars, just like there are a lot of people who use letter-sized paper. They are your colleagues, even though you may disagree with them; and it gives them a more visible way of not having those asterisks cluttering their results.
Thanks
Russ Lenth
Professor Emeritus
University of Iowa, USA
More information about the R-devel
mailing list