R-beta: Re: Significance stars

Jim Lemon (bitwrit@ozemail.com.au)
Fri, 19 Jun 1998 11:57:40 +0000

Date: Fri, 19 Jun 1998 11:57:40 +0000
From: Jim Lemon <bitwrit@ozemail.com.au>
To: r-help@stat.math.ethz.ch
Subject: R-beta: Re: Significance stars

Just a comment from someone else who was taught to be suspicious of
'standardized' significance criteria.  I got the 'critical statistic'
viewpoint in one set of stats courses, and the 'parameter estimation'
viewpoint in another.

>From my experience, about the only thing most people can guardedly agree
upon is the probability of the obtained statistic.  This is only in the
context of a plausible model distribution, of course.

This said, I have spent years ignoring the 'standardized' significance
displays that are churned out by most statistics packages.  One thing
that endeared S (and recently R) to me was the minimization of such
advice from the program.

I suppose one could cobble up a generic show() function which could be
customized to the preferences of the user.


r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-help-request@stat.math.ethz.ch