🚀 go-pugleaf

I know this is an old thread, but it did not seem to get talked up much.
I meant to say more at the time, but did not get the chance.

Ed Johnson tells us of difficulties he has had with other academics accepting
his inclusion of the use of statistics in a media research course. (!)
To some extent this can be dealt with by having a Statistics Dept, simply
because these concerns are raised in more lockerroom discussions, and there
are more people there who might actually know what they are talking about.
But this does not really get at the problem; even here are Virginia Tech,
with a very strong and active Statistics Dept (e.g., the Communications Dept
requires their undergrads to take our Stat for Social Sciences course) we
see the phenomenon Ed is referring to.  The ignorance, and insight, of the
mentioned bias is often summarized by Sam Clemens (a Missouri man) as "There
are three kinds of lies: lies, damned lies, and statistics."

This, I think, gets at the root of the problem, which is, not that all use of
statistical methodology is a lie, as Mr. Clemens no doubt meant to imply, but
that there are inscrupulous and ignorant people who try to argue their case
in a given context with statistics that are meaningless.  I recall hearing a
story on the news one time, that medical researchers at blahblahblah
university had finally proven that if you do not die of something else first,
you will die of cancer.  Imagine that!  Now, who got it wrong, the researchers
or the media?  Or both?

Another common problem in discussions involving statistics is the difference,
rarely noted, between the roles of inference and description.  Associated with
this is the difference between sample and population statistics.  Most people
only think of descriptive/population statistics, but ask the critics how they
would study the differences between, say, television viewing patterns of
different demographic segments of society?  This has obvious importance in the
marketing/research of the media but how can you possibly study this without
the use of statistics?

Finally, let me pass on a few points I often make to people who criticize
statistics.  First, yes liars can use statistics to continue lying, but that
does not imply that there is no such thing as an honest statistician.  Second,
statistics is, perhaps most concisely, the study of data analysis.  Ask your
skeptic if he has ever heard of these rare items we call data.  Finally, while
all branches of science think about "the scientific method" only
statisticians (and philosophers), to my knowledge, actually claim to study it
for itself.  As with any academic discussion, you are probably talking to
someone with a Ph.D., and therefore they are convinced that they are an
expert in everything that they know ... and they know that statistics as a
field is rubbish, and hence you, who think otherwise, are obviously wrong.
But occassionally I have these discussions with those rare misguided academics
who do not realize that they know everything, and they often say "hmmm, I
didn't really realize that statisticians thought about things like that."

(By the way, it is my experience that this is the conclusion of open-minded
people when they talk to anyone about what they do.  When people spend their
entire careers thinking about a given area, they tend to have deeper thoughts
about it than those who never thought about it before.)

Clark K. Gaylord                  ****************************
Dept. of Statistics               *** Thelonious Monk -    ***
Virginia Tech                     ***   a melodious thunk  ***
703/552-2868                      ****************************
e-mail: gaylord@vtvm2 or gaylord@vtvm2.cc.vt.edu