Re: Cohen's ecological data: a test of LNT?

[Jim Muckerheide <jmuckerheide@delphi.com>]

Ken, Group,

I'm sorry to have been unable to respond to a few key msgs, as follows:

Ken Mossman wrote:
> 
> This my last shot!
> 
> Cohen and I agree that LNT is a causal relationship. Howevever, I am now
> totally confused about whether Cohen's data tests for a causal relationship.
> In his April 1997 paper in Health Physics (page 625), Cohen indicates that
> his data do not test for a causal relationship but has a much more limited
> objective. Now Cohen says that his data tests for a causal relationship.
> Which is  it? Either the data test for a causal relationship or thery don't!
> 
> If the data test for a causal relationship, as Cohen now states, then the
> ecological fallacy becomes operative. If the data do not test for a causal
> relationbship, then the data are not a test of LNT because LNT is a causal
> relationship (as Cohen and I now agree).
> 
> Kenneth L. Mossman

These semantics seem rather disingenuous. The data don't care about semantics. 

Bernie disproved the LNT as a causal relationship. He further showed a
definitive negative correlation. He chooses to only address the patently
obvious and definitive fact of the failure of the LNT. But the extent of the
analysis by now has also, especially by stratification of low-dose and
high-dose counties, essentially negates the possible 'ecological fallacy'
effect, despite the lack of individual dose-responses at the county data
level. 

BUT: The 'eco fallacy' is ever only AT BEST a remote possible chance of an
effect, which is now essentially eliminated by the extent of the data (i.e.,
the whole population against which you can not make 'corrections' due to some
possible, and massive, unknown conditions affecting a subset). This is further
confirmed by the definitive and consistent testing and results of the data
analyzed in hundreds of substantial subsets, with stratification! 

Now, it would seem certain IF any ONE subset did not have the inverse
relationship, the ONLY conclusion can be that IT would be an demonstrable
effect of the statistical equivalent of the 'ecological fallacy!' IT would be
the subset that DOES NOT agree with the data for the whole population! 

IT would then provide ONE example of an actual effect (of the remote
possibility) of an inverse relationship actually manifesting in a real data
set (though the lack of individual dose/effects defines the 'eco fallacy').
The lack of one such subset in hundreds indicates the remoteness of finding
such inverse relationships in substantial samples - consistent with the
applicable premises of statistics. But which the anti-nukes take to be
'definitive,' i.e., the ecological radon dose-response data by Cohen and
others is CERTAIN to be in error (not just a remote possibility), in every
case, despite statistical rules for defining error and variance. 

But surely at some small sample size (if they run through enough subsets -
Cohen's data is available - they will find a subset of the data that is
inverse to the actual relationship of the whole dataset!? :-)  I'm sure they
would then say this 'proves' their case - at least for the innumerate
politicians!

LNT-advocacy, the LNT-defenders, especially since the BEIR VII Committee
disaster, seem to be getting ever more tortured and desperate! 

Regards, Jim
============
************************************************************************
The RADSAFE Frequently Asked Questions list, archives and subscription
information can be accessed at http://www.ehs.uiuc.edu/~rad/radsafe.html