[R] Hornet's Nests and Parallel Universes

John Maindonald john.maindonald at anu.edu.au
Sun May 19 03:36:51 CEST 2002

Fisher is not the only person that it may be necessary
to read 4 or 5 times.  The same may be the case for
side comments that Bill Venables is wont to make.

Now to the parallel universe that I have in mind.  I wonder
whether the time is opportune for a list that focuses on
"Statistical Methodology for R Users".  The difference from
other statistical methodology lists is that it will be
assumed that list members have contact with, and an interest
in, R.  So the appearance of R code, and references to R
documentation, would be entirely in place.

One of the great things about R, and this email list, is
that it has allowed methodological specialists from a
professional statistical tradition to encounter application
area specialists from areas that have developed their own
relatively separate statistical tradition.  Here we aim,
primarily, to discuss R.  I would like to see a place where
those from the R community who wish discuss methodology,
with R a secondary focus.  We might, for example, talk about
other things than R where there could be useful co-operative
endeavour and exchange of ideas.


The best account I know of the separation of applied statistics
into separate traditions, at least with respect to education
and psychology, is in
Gigerenzer et al. 1999.  The Empire of Chance.  Cambridge
Univ Press.

There is a lot more that could be added to what Gigerenzer at al
say, but that book is a good start.  There are books waiting to
be written: 1) on the different traditions of multi-level
modeling, generalizability theory etc, that have grown up in
different application areas.
2) on the different notational conventions that have grown up
in different areas.  

In reading Fisher, modern readers have the problem that he
does not follow conventions that had not been invented (!)
when he wrote.  The diifficulty goes deeper.  It is not all
that obvious how the language, notational conventions and
intellectual framework would change if for example his "The
Design of Experiments" was to be completely rewritten for a
modern audience.  It would make quite a difference whether
that audience was made up of field biologists or
pyschologists.  Professional statisticians, who were not the
book's primary audience, are another matter again.

John Maindonald                     email : john.maindonald at anu.edu.au
Centre for Bioinformation Science,  phone : (6125)3473        
c/o SMS,                            fax   : (6125)5549 
John Dedman Mathematical Sciences Building (Building 27)
Australian National University
Canberra ACT 0200
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