[R] Re: Giving a first good impression of R to Social Scientists

[John Fox]

I apologize for joining this discussion late, but I was out of town and just
reviewed postings to R-help. I noticed that there were a couple of other
threads that also touched on similar issues.

Bob Andersen and I have written a paper on using R to teach social
statistics; a copy is at
<http://socserv.socsci.mcmaster.ca/jfox/Teaching-with-R.pdf>. It would be
tedious to rehash the paper in detail here, but I'll make a few brief
points:

(1) I believe that, except in special circumstances, GUIs to statistical
software are best for casual or infrequent use. Even users of SPSS who
intend to use the software seriously would be well advised to learn to write
commands. 

(2) The GUI in the Rcmdr package covers what's typically taught in first and
second social-statistics classes.

(3) The difficulty of moving from one statistical package (or computing
environment) to another shouldn't be confused with learning to use a
particular package (or environment) as a novice. I don't believe that it's
any more difficult to write R commands than, say, SAS or SPSS commands.

(4) As others have pointed out, the principal advantage of R (or another
statistical computing environment, as opposed to a statistical package) is
its programmability. Nevertheless, R is not difficult to teach and learn
even to those who will not (initially) be writing programs.

John

--------------------------------
John Fox
Department of Sociology
McMaster University
Hamilton, Ontario
Canada L8S 4M4
905-525-9140x23604
http://socserv.mcmaster.ca/jfox