Significance stars (was Re: R-beta: glm bug)
Martin Maechler
maechler at stat.math.ethz.ch
Thu Jun 18 19:55:12 CEST 1998
>>>>> "Jim" == Jim Robison-Cox <jimrc at mathfs.math.montana.edu> writes:
Jim> On Thu, 18 Jun 1998, Jim Lindsey wrote: (snip)
>> I might also say that I strongly object to the stars and
>> significance levels in summary (sorry, Martin, we already discussed
>> this). I consider fixed level testing to be bad statistics; at
>> worst, a P-value should be given if nothing better (confidence or
>> credibility intervals) is available. I do not want my students
>> seeing such stars and fixed levels. Make it an option if you like
>> but not the default display with summary(). Jim
Jim> I agree! Is there a good way to set an option which goes to a
Jim> print method (in this case print.summary.lm)? Would it have to be
Jim> a global option?
R> str(print.summary.glm)
function (x, digits = max(3, .Options$digits - 3), formatfun = format, na.print = "", symbolic.cor = p > 4, signif.stars = TRUE, ...)
R> str(print.summary.lm)
function (x, digits = max(3, .Options$digits - 3), symbolic.cor = p > 4, signif.stars = TRUE, ...)
Which shows you there IS already an argument to turn off the signif.stars:
print(summary([g]lm(....)), signif.stars = FALSE)
Jim Lindsey wanted a *global* option so that he can turn it off for all
his students when they just use
summary([g]lm(....))
Any proposals for a NAME of the global option?
JL even looked for a possibility to turn off the P values..
Do we want/need this as well?
---------------------------------------------------------------
A few afterthoughts:
Note that I introduced those both to print.summary.lm() and ...glm().
We (Statistics ETH Zurich) have been using these stars for about seven
years now (in S-plus, not R..),
both for teaching and (serious) data analysis.
Most of my teaching (and researching) colleagues, many students
and consulting customers have made explicit *positive* comments about the
stars, (and asked for our S-plus replacement code in order to have it also
on their own machines...) and I haven't got any negative ones till both Jim
Lindsey's.
NOTE: We do *NOT* advocate believing in P-values,
(some of you may know how little we believe in normally distributed
data here ..),
the main point of the sig.stars is a
``graphical summary'' of which coefficients might be significant
~~~~~~~~~~~~~~~~~
Martin Maechler <maechler at stat.math.ethz.ch> <><
Seminar fuer Statistik, ETH-Zentrum SOL G1; Sonneggstr.33
ETH (Federal Inst. Technology) 8092 Zurich SWITZERLAND
phone: x-41-1-632-3408 fax: ...-1086
http://www.stat.math.ethz.ch/~maechler/
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
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 at stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._
More information about the R-help
mailing list