To put it very briefly: I couldn't agree more!
As I am inclined towards writing too long postings, I shall try to refrain
myself from numerous other arguments in favour of Excel. So just a few notes
(please bare with the style - they're put as unemotionally as possible by an
Excel fanatic).
Of course, my statistical expertise is a couple of levels below that of the
people involved in this discussion so far (they all belong the sci.stat
expert community I learn from and admire so much!) but for what it's worth:
I am neither an amateur nor a non-specialist, and have a degree in
psychology so I should know some "theory" about teaching and learning and
the like, and after almost ten years of everyday tatistical analysis,
consulting and teaching I am convinced that spreadsheet literacy is
absolutely essential and enormously beneficial - not only for medical
students, but for every student, and not only for students and
professionals, but for just about every "white colar" job, from high-school
level on! - Of course, it's not necessarily Excel, but all in all it is
simply a league above any other spreadsheet (I've tried many to avoid bias),
especially if one considers user community support (e.g., via newsgroups)
and wealth of add-ins available, of which so many so capable ones are
publicly available! - As soon as, say, Gnumerics gets reasonably close, I'll
be the first one to use and teach it with even more zeal, just as I'm
currently learning and using R, while I'll remain an SPSS enthousiast,
especially as long as our university (BTW, one with well over 50.000
students) has a campus license. - Just think how many computers come with
pre-installed MS Office, or at least are entitled to a license, and how much
office "work" could be automated and saved with such negligible cost and
effort compared to, say, mayor RDBMS suites, or dedicated application
development ... - Sure, Excel is just a Swiss army knife, but a space-tech
one, and if the issue is eating with bare hands or a tiny SAK fork and
spoon, or breaking a log by hand or sawing it with a tiny SAK saw, etc., I
think there can be no doubt ... - I am even convinced that apart from the
Internet, spreadsheets are the most revolutionairy contribution of computers
to the ways of information processing, but let's leave that aside (again,
I've read and written at least a little on the matter).
Anyway, maybe concrete examples are better than any advocacy, so sceptics
(e.g., mathematicians convinced that it is a blasphemy to represent abstract
concepts as concretely as possible for "the masses" ...) can see and judge
for themselves:
- if the problem is accuracy:
http://digilander.libero.it/foxes/index.htm
- if one longs for simulations:
http://home.uchicago.edu/~rmyerson/addins.htm
http://www.cse.csiro.au/poptools/index.htm
http://www.tushar-mehta.com/ -> SIMULATE
- for statistical graphics enthousiasts:
http://www.prodomosua.it/ppage02.html
http://www.geocities.com/jonpeltier/Excel/index.html
- econometrics (robust regression, logit and probit) using Excel's Solver:
http://www.wabash.edu/EconExcel/home.htm
etc. etc. etc. The above are just some of the many gems.
And finally, even though it should probably come first, survival analysis
(whereby I'll skip the fine commercial statistical add-ins, e.g., WinSTAT
and XLSTAT):
- as for academic add-ins, a fine example is EcStat:
http://www.stat.mq.edu.au/units/stat170/index.htm (has K-M curves and
logrank test; stronger in, say, logistic regression)
- as for professors designing their own workbooks/macros, here's an amazing
Japanese example:
http://aoki2.si.gunma-u.ac.jp/lecture/stats-by-excel/vba/index.html
(the home page "Stats-by-Excel is at
http://aoki2.si.gunma-u.ac.jp/lecture/stats-by-excel/)
The author is certainly not unaware of Excel's limitations and flaws
(http://www.matsusaka-u.ac.jp/~okumura/software/excel/roundoff.html), but
the most important thing is that no. 21 (incidentally, the only item with
latin characters text on the page, so easily spotted) leads to, yes, K-M
curves! The macro charts them and also computes standard errors (there's
another one for the logrank test). And, for what it's worth, with my zero
knowledge of Japanese, a social science degree and very modest knwoledge of
VBA, I've managed to fully translate the workbook and the macro into English
in a day, so ...
Gaj Vidmar
University of Ljubljana, Institute of Biomedical Informatics
P.S. I believe that not only should accountants strive to learn to use a
stetoscope, but that everyone should strive to learn everything all the
time! For one sole reason: in order to be able to teach and help others! -
Not maybe, for sure I am an idealist, but hey, how come I'm still alive if
Thomas Hobbes was right?! And to be provocative: ever wondered why many
think that in the history of mankind there has been just one true succes
story of socialism - Slovenia? (If nothing else convinces you, double GDB
per capita to any other "socialist" country, not to mention that at least
from the late seventies on there was - unlike in the rest of Yugoslavia -
literally full freedom of speach [unlike, erm, khm, nowadays in ...], and
full-fledged market economy without private ownership of companies, and a
world-famous health care system, and more cars per capita than some EU
countries?!) And why many think it is the only success story of transition
(back) to capitalism? (In a few months we'll join the EU not being on the
receiving end of the aid to the underdeveloped ....) - This resourceless
unknown historyless pocket-world? - Ever read that by the beginning of this
century, there were NO illiterate inhabitants in Slovenia?! - Maybe
knowledge isn't and shouldn't be beneficial just for the elite in any sense
of the word?! - Knowledge and information, whether from GNU and Linux or
William Gates, or CIA or Arabic astronomers - doesn't matter!! - And maybe,
someday, we'll all ...
Till then, happy calculations, with whichever tool.
.
.
=================================================================
Instructions for joining and leaving this list, remarks about the
problem of INAPPROPRIATE MESSAGES, and archives are available at:
. http://jse.stat.ncsu.edu/ .
=================================================================
