Re: Re: Cohen's Fallacy

[Kai Kaletsch <info@eic.nu>]

Philippe,



Understanding the mechanism helps you fill in gaps in the data. Suppose that

the relationship: "smokers have 0.9 times the radon concentration of

non-smokers" is based mainly on screening measurements performed in

basements. If you don't care about the "why", you might be tempted to

conclude that this can be translated directly into an exposure relationship.



However, if you postulate that this relationship is because smokers open

more (main floor) windows, then you would have to conclude that the

difference in exposure is much larger than 0.9. Opening main floor windows

would have much more of an effect on main floor radon levels than on

basement levels.



At this point, I don't know how many gaps need to be filled in. Missouri,

Iowa. New Jersey... (the places with case-control studies) are located in

the US. Cohen's data also deals with the US (overlapping data sets). The

case control studies write down radon level on the main floor, smoking, age

and whatever else they think could have a bearing on lung cancer on an

INDIVIDUAL basis. So, if there are any systematic relationships among these

factors that could effect Cohen's county level analysis, why don't they

share these findings with us? Why are we still talking in abstract terms and

keep using big words?



If they came out and said: "In low radon counties, all the old people live

in high radon homes and all the young people live in apartments. That's why

low radon counties have a high lung cancer mortality."  I would be

satisfied.



We should note that the cross level bias needs TWO systematic associations

to make any difference: 1) "the old people live in high radon homes and the

young people live in apartments" and 2) "This process is inversely

proportional to average county radon level". That is why I think the cross

level bias theory is fairly unlikely.



Kai





----- Original Message -----

From: "Philippe Duport" <pduport@uottawa.ca>

To: "Rad health" <healthrad@HOTMAIL.COM>

Cc: <radsafe@list.vanderbilt.edu>

Sent: Monday, January 28, 2002 2:53 PM

Subject: Re: Re: Cohen's Fallacy





> The "why" is nice to know but not necessary to acknowledge an observation.

> Steam power was known a few hundred years BC; steam engines were used in

> industries in the mid 17th century, and nobody knew "why" they worked

until

> Sadi Carnot, a quarter of a century later, explained "why".

>

>

> Philippe Duport

>

>

> ----- Original Message -----

> From: "BERNARD L COHEN" <blc+@PITT.EDU>

> To: "Rad health" <healthrad@HOTMAIL.COM>

> Cc: <Jack_Earley@RL.GOV>; <radsafe@list.vanderbilt.edu>;

> <motnivas@YAHOO.COM>; <tjsav@LYCOS.COM>

> Sent: Monday, January 28, 2002 12:54 PM

> Subject: RE: Re: Cohen's Fallacy

>

>

> >

> > On Mon, 28 Jan 2002, Rad health wrote:

> >

> > >

> > > That is the question I have asked Dr. Cohen to explain numerous times.

> > > Please provide a definitive explanation as to why the radon

> concnetrations

> > > show a rather large inverse correlation with smoking?

> >

> > --It is not my responsibility to explain this, although I could

> > speculate on reasons if that were important. It is not important for my

> > studies. My job is to find out what the correlations are and take them

> > into account in my analyses.

> >

> > ************************************************************************

> > 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/

> >

>

> ************************************************************************

> 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/

>



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

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/