[R] R for psychology

[Prof Brian D Ripley]

On Sat, 26 Aug 2000, Jonathan Baron wrote:

> For those new to this list, I point out that Yuelin Li and
> I have drafted an introduction to R for psychology, trying
> to emphasize the things the psychologists usually do.  It
> is at
> http://www.psych.upenn.edu/~baron/rpsych.pdf
> and
> http://www.psych.upenn.edu/~baron/rpsych.htm
> Because it is a draft, with frequent changes, we have not
> yet asked for it to be placed in the CRAN page.  We welcome
> suggestions for changes.
> 
> One thing I want to add to it is a discussion of Error() in
> analysis of variance, once I understand it better myself.
> 
> On the topic of R for psychology, I should point out that
> the two biggest gaps I see in what R can do are manova and
> factor rotation.  Prof. Ripley pointed out recently that
> "the various multivariate test statistics" are "not difficult
> to write if you know the formulae and have a need," but this
> leaves out the fact that most psychologists would be hard
> pressed to find the formulae and even harder pressed to figure
> out how to write functions in R.  Sad but true, even though
> this sort of test is very common in some fields of psychology.

You will now (well, tomorrow perhaps) find manova() and friends in the
R-devel version .... I had hoped that someone else would write them, but
got down to it myself this morning.  All my comments needed was *one*
person (Jonathan?) who has a need (I don't) and knows enough.  If that's
not what you want, we need to know more precisely what that is.  (It does
what S and the multivariate books I consulted suggested.)

Some nice examples for the help page would be appreciated.

> (In their defense, most of these people studied these tests
> at some point and understand them, and have not bothered to
> learn the formulae because they know they can rely on SPSS,
> Systat, or SAS to do it for them.)  It is possible that the
> multilm package might do some of what is needed, but I haven't
> checked yet.

(Don't think so. That does a multivariate test on the *overall* fit
of a multiple linear model, a much less interesting test.)

> As for factor rotation, the same comments apply.

Those are a little harder to do.  My main difficulty (apart from a lack of
interest) is lacking a definitive reference.  As far as I can see no two
sources entirely agree on their definitions.

To distill things to their essence, R is cooperative enterprise and things
are missing in the main because no one has supplied them. (Large parts of R
are there because I as a user wrote them and submitted them to R-core.)

-- 
Brian D. Ripley,                  ripley at stats.ox.ac.uk
Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272860 (secr)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595

-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
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
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._