[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
RE: Absence of Proof...
Answer to Jim Dukelow's mailing dated April 21st
commenting on our "Absence of Proof.." posting.
By Fritz A. Seiler and Joseph L. Alvarez
Dear Jim,
In our first post on "Absence of Proof..." we were referring
to you simply as the originator of the letter containing all the
EMF data with which you answered Robert Godfrey's mailing.
Our post did not really constitute an answer to your letter, nor
was it addressed that way. Also, we sent it to RISKANAL
and to RADSAFE because we are members of both lists and
know that the subject would interest readers on both lists. In
fact, we got a few more responses and comments from members
of RADSAFE than from the RISKANAL mailing list.
So we were a little bit baffled by your letter to RADSAFE
accompanying your original answer to Robert Godfrey. Partly
we understand your reaction since you seem to think that our
post was referring to yours. Well, to say it again, it was not.
It was our commentary to many things we have read, heard and
talked about since we all went on the EMF spree of last summer.
(By the way, for every physicist and every non-physicist who
has ‘suffered' through Physics 101, "emf" means electromotive
force!)
So, we did not put up any "Straw Men"; we were addressing
real views which are out there. Indeed we got responses on
private and list mail which implied or averred outright that the
author thought that science was capable of ‘proof'. We know
that this attitude is based on the fact that some people define a
"scientific truth", and go on from there. That kind of ‘truth',
however, is a very squishy notion which depends too much on
the size of the relative errors. "Truth" as a function of its
uncertainty? We think not. "Experimental observations" or
even "experimental facts" as a function of uncertainty? You
bet!
We have already pointed out in our "Letter to Vicki Bier"
what our opinion is on the freedom of selecting a null hypothesis.
To say it again in mental shorthand: In pure abstract statistics
you may think that you have such a freedom; for the uses in
science and its applications you do not, because the experiment
always asks a clear-cut question which provides the proper null
hypothesis.
But does pure abstract statistical theory really give you all
that freedom? In view of our carefully worded letter to Vicki
Bier, it may be worth the effort to look in more detail at your
comment which is worded with equal care, and which is worth
repeating:
"In fact, when you use statistical tests of hypothesis to do
science (or anything else, for that matter), you have almost
complete freedom to select your null hypothesis, with the
following constraints. First, you want to choose as your
Null the thing you would like to disprove. Second, you
need to choose a null hypothesis that is capable of pre-
dicting the outcome of your experiment or the expected
properties of your data set. The predictions are used to
establish the "Accept H0" and "Reject H0" regions."
First, your claim of "almost complete freedom" is at quite
some odds with the two conditions that follow. Indeed,
from our point of view, you have an almost complete lack
of freedom because you have to fulfill these two conditions.
Here, we think that we have the root of our difference. We
know that scientists like to think in terms of questions, but we
also know that statisticians often do not like to think in such
terms.
Second, on closer inspection, however, your conditions
seem to be a quite good approximation of what we said that
‘doing good science' requires. The main difference seems to
be that you start from a theory that we call ‘abstract statistical
theory', i.e., a theory which quite properly is not tied to any
particular example, in order to arrive at your conditions,
whereas we apply the Scientific Method to arrive at ours.
Well, the Scientific Method is actually an abstract theory also
but it is clearly focused on the application of statistics to its
problems. This is why we like our approach better, it depends
explicitly on the scientific question asked and allows no logical
twists or machinations. Nevertheless, it should not be too
surprising that both arguments lead essentially to the same
requirements!
If we had not known that, we might have taken exception to
your characterizing Bernie Cohen's paper as one in which "he
chose as his Null the thing that he wanted to reject." We think
that this problem is really a question of approach and of attitude
toward the scientific question. But then, as the title "Test of the
linear-no threshold theory of radiation carcinogenesis for inhaled
Radon decay products" (Cohen, B.L.; Health Phys. 68:157-
174; 1995) makes quite clear, it is actually a test of the linear
paradigm and is thus in agreement with the best scientific tradition.
As it is, however, it is also an approach which fulfills the two
conditions which you propose!
Freedom or not, we find it rather amusing that we both want
the same thing, but we get there from completely different points
of view. The only real difference is the use of the term "almost
complete freedom"! :-)) :-)
As for the rest of your remarks, we think that we are all a bit
overly sensitive and should not over-interpret each other's posts.
The "Fastest Gun in the West" approach somehow does not cut
it in the case of scientific discussions. We should all try to think
a bit more carefully about what the other side is really saying.
They might actually agree with you!
Best regards,
Joe and Fritz
*************************
Fritz A. Seiler, Ph.D.
Principal
Sigma Five Associates
P.O. Box 14006
Albuquerque, NM 87191-4006
Tel. 505-323-7848
Fax. 505-293-3911
e-mail: faseiler@nmia.com
**************************
Joe Alvarez, Ph.D., CHP
Auxier & Associates, Inc.
10317 Technology Dr., Suite 1
Knoxville, TN 37932
Phone (423)675-3669
FAX: (423)675-3677
Email: jalvarez@auxier.com
**************************
************************************************************************
The RADSAFE Frequently Asked Questions list, archives and subscription
information can be accessed at http://www.ehs.uiuc.edu/~rad/radsafe.html