Re: st: Intro social science stats book

[David Airey <david.airey@vanderbilt.edu>]

.

I rarely see the same texts get used in intro stats course on  
Vanderbilt's campus. I guess that's why everyone thinks they can  
write one (and do so!).

Vanderbilt is all over the map in terms of software promotion. In  
Psychology, SAS got a note by Tomarken, the new Chair, in a recent  
talk on test multiplicity. In Peabody Quantitative Psychology, R  
shows up on their web pages. In our Medical Center, Dr. Harrell  
pushes R pretty hard, but acknowledges Stata as "next best".

http://biostat.mc.vanderbilt.edu/twiki/bin/view/Main/SoftwareRecs

However, that's just Frank's opinion, and I think he's got his order  
a little mixed up! To teach introductory statistics, he lets Stata be  
used, not R. I think that is telling. I personally cannot see R being  
used in an introductory stats course, even though I've keenly  
followed its development, occasionally used it, and spent money on  
"introductory" R books (because R manuals stink).

There is no doubt about it, R is here to stay and it has clearly  
gained visibility: I have now seen job ads that would once only ask  
for SAS experience, acknowledge "R" as an alternative skill set. Why  
not Stata?! R's GUI is nowhere. R is not multiprocessor capable like  
Stata MP and has no short term plans to become so. Stata is  
consistent, good for reproducible research, has excellent flexible  
graphics (my opinion), lots of models, has Mata, is on every major OS  
platform, is very programmable (even its GUI is so), etc., etc...

Maybe Stata needs more review press. It has had major additions since  
version 7, the last time I saw it reviewed in magazines. Maybe Stata  
needs to hire someone to fill the evangelist position?! I'll send my  
resume.

;)

-Dave




> I would argue that your sample is too small to make general
> statements about the use of Stata.
> Unless Suzy comes from the same school where I teach now, I would add
> my convenience sample to hers. Stata is pretty much unheard of in
> mainstream statistics. We have a STAT 7xxx class on statistical
> software that teaches SAS and S-plus/R, as all the medical folks are
> using SAS, and all the top-notch statistical research is usually
> implemented in R... or at least nobody will call Matlab or Fortran a
> purely statistical software. So students see SAS as a must when
> applying for industry jobs, and R is necessary to talk to our
> professors because the latter use R themselves.
>
> I personally think it is time for Stata Corp. to stop chickening out
> of statstics market. Stata has all the programming tools statisticians
> find in R (a matrix language, object-oriented tools, compiled code for
> faster implementations, even interface with C which, to my memory, was
> only used my James Hardin for the suite of measurement error GLMs).
> And  the more I work with SAS, the more I like Stata :))
>
> --
> Stas Kolenikov
> http://stas.kolenikov.name
> *
--
David C. Airey, Ph.D.
Research Assistant Professor

Department of Pharmacology
School of Medicine
Vanderbilt University
8148-A Medical Research Building 3
465 21st Avenue South
Nashville, TN 37232-8548

TEL   (615) 936-1510
FAX   (615) 936-3747

*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/