PDA

View Full Version : Re: What criteria should we use for interpreting epidemiology?



fbuczek60
02-20-2008, 12:47 AM
Chris and others,

I can think of nothing that would more quickly dilute the truth, than
research by consensus. Even small scientific teams, composed of a
senior scientist and a handful of post docs, working on a limited
budget, can find the truth if they use good experimental techniques. If
their sample size is considered by some to be small, then others can
replicate the study with a larger sample to confirm or refute the
findings. This is, after all, the way science corrects itself over
time.

I can recall an exchange between two men I very much respect, at the
2003 ASB conference in Toledo, Ohio, USA:

Dr. McNeil Alexander had just finished giving a wonderful Borelli Award
Lecture. After a few questions from the audience, Dr. Peter Cavanagh
asked Dr. Alexander if he had any advice for the young people in the
audience who were just beginning their careers. After thinking for a
moment, Dr. Alexander replied, "Don't follow the herd." He went on to
say that important discoveries in science seem to happen more often when
people allow themselves to think independently.

It is difficult for me to imagine this happening as easily within the
consortium described by Dr. Ioannidis, as quoted in an earlier BIOMCH-L
submission.

[These are my opinions, and not necessarily those of NIOSH.]

Best regards,
Frank

Frank L Buczek Jr, PhD
Branch Chief, HELD/ECTB
National Institute for Occupational
Safety and Health (NIOSH)
1095 Willowdale Road MS 2027
Morgantown, WV 26505
304-285-5966 voice, 304-285-6265 fax
fbuczek@cdc.gov


-----Original Message-----
From: * Biomechanics and Movement Science listserver
[mailto:BIOMCH-L@NIC.SURFNET.NL] On Behalf Of Chris Kirtley
Sent: Wednesday, February 20, 2008 7:11 AM
To: BIOMCH-L@NIC.SURFNET.NL
Subject: Re: [BIOMCH-L] What criteria should we use for interpreting
epidemiology?

Very interesting, Adrian!

For those interested, the original article is at:
http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=1182327

I have often thought that research papers have following weaknesses:

1. In the introduction or background (which should build the case for
the
question being answered), reference is often made to statements by
others in
a similar section of their paper, or speculative comments in the
discussion.
In my view, reference should only be made to the conclusions previous
papers.

2. Particularly in biomechanics (and even more sports biomechanics) I
suspect, scientists often rely on their intuitive feelings about what is
true, refusing to believe results which have been confirmed over and
over
again.

3. The numbers of subjects and tests are usually so small as to make the
application of statistics absurd.

Great food for discussion, Adrian & David!

Chris