RE: Re: Cohen's Fallacy

[BERNARD L COHEN <blc+@PITT.EDU>]

On Mon, 28 Jan 2002, Field, R. William wrote:



> Dr. Cohen,

>

> I would be very willing to write a letter to the Journal of Radiologic

> Protection suggesting that the specific cause of your inverse findings is

> inter county (group-level) confounding.  We and others have pointed out

> that this error is potentially unbounded.



	--I addressed this point in my papers in Health Physics

76:437-439;1999, and J. Radiol. Protection 21:64-65;2001, even giving

some specific examples of how this can explain the

discrepancy between my observational results and LNT theory. But the

examples I gave were completely implausible. What is needed is a specific

example that I cannot show to be completely implausible.

	A general treatment is not adequate because it does not include

the requirement of plausibility. As far as I can understand, plausibility

cannot be tested unless there is a specific example. That is why I need a

specific example. If you can suggest a way to test plausibility without a

specific example, please let me know.

	The Scientific method requires that any conflict between theory

and experiment must be plausibly explainable. You may say that cross-level

bias is the explanation, but if that is so, it would be easy to conjure up

a plausible example. I have 25 years of research experience in theoretical

(and experimental) basic physics, and have dealt with numerous general

proofs and am aware of others dealing with many more, but it is always

very easy to come up with a specific example. No physicist would accept a

general proof without being offered such examplles. If you find this hard

to understand, maybe you could consult a theoretical physicist.



  I have already been told by the

> past editor of Health Physics that further letters on this topic are not

> welcome, although I would try them first again if you like.



	--The agreement I made with Ken Miller, the former Editor, was

that he would give me one more response to our exchange and then he would

give you the last word. That applied only to our back and forth exchange

at that time. It did not apply to future papers on my work, and they

published my paper in Health Physics 78::522=527;2000 after that.



 I would also

> suggest that since this is your study, you examine (rather than making

> assumptions) directly the correlation between smoking and lung cancer on

> the county level by using the technique suggested in Guthrie's paper, which

> I already faxed you.



	--I will study the Guthrie paper, but that is irrelevant here. If

you think I am making unjustified assumptions, you should point them out.



  I am making the assumption that you are sincere about

> your desire to determine the cause of your inverse relationship.  If you

> say, my suggestion is not specific enough, I think people should really

> question your sincerity about determining the cause of the inverse

> relationship.



	--In my 50+ years as a scientist, no one (to my knowledge) has

ever before questioned my scientific sincerity or integrity. I value my

scientific reputation above my life itself. If you don't understand what I

have said above about why a specific suggestion is needed, please consult

a theoretical physicist. If you again question my sincerity or scientific

integrity, I will cut off all further communication with you.



************************************************************************

You are currently subscribed to the Radsafe mailing list. To unsubscribe,

send an e-mail to Majordomo@list.vanderbilt.edu  Put the text "unsubscribe

radsafe" (no quote marks) in the body of the e-mail, with no subject line. You can view the Radsafe archives at http://www.vanderbilt.edu/radsafe/