[S] RE: Publishing statistical abuses

Gunter, Bert (bert_gunter@merck.com)
Thu, 30 Apr 1998 08:35:44 -0400


I do not understand your publication proposal. To what end is this directed?
It seems to me that it's effect would be to

1) Further alienate practicioners from academics ("It's those ?!!x?
arrogant, theoretical statisticians again
who don't understand the practical realities of science and engineering and
really believe their idealized theories"). Subjecting others to a form of
public ridicule does not seem constructive to me.

2) Further demonstrate how badly statistics is taught -- as mathematical
niceties disconnected from
sceince and engineering -- and thus further erode support for having it
taught (maybe not such a bad thing. As George Box has remarked, when
teaching engineers who have had no prior statistics training he starts on
ground 0; when teaching those who have, he's below ground).

I believe a much better idea is for statisticians to get involved with
scientists and engineers working on substantive problems and develop/adapt
useful statistical methodoogy that is published both in statistical journals
(to maintain the tie to and scrutiny of the statistical mainstream) and the
subject matter literature (to influence the methodology used).
One problem: Most statistical journals are not interested in publishing
applications that do not involve "innovative" methodology. Our academic
orientation again?

Bert Gunter
Biometrics Research
Merck Research Labs
P.O. Box 2000
Rahway, NJ 07065-0900

This message was distributed by s-news@wubios.wustl.edu. To unsubscribe
send e-mail to s-news-request@wubios.wustl.edu with the BODY of the
message: unsubscribe s-news