RE: Risks of low level radiation - New Scientist Article

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

On Thu, 6 Dec 2001, Jim Nelson wrote:



> Are you a statistician?  Larger populations are good if you have good

> underlying data to describe the population.



	--I have treated over 500 potential socioeconomic factors on which

there is good information from the U.S. Census Bureau.



  Unfortuantely, it is very hard

> to derive good data from surrogate ecologic measurmements such as sales tax

> on cigarettes (used to infer smoking).



	--Sales tax on cigarettes is only one of three methods I used to

infer smoking prevalence. My preferred approach was to use a survey by the

U.S. Bureau of Census. But most importantly, I showed that even if my

smoking prevalences were wrong and there is a perfect negative

correlation between smoking prevalence and radon exposures, there would

still be a very large discrepancy between my data and the prediction of

Linear-no threshold theory



> The other study I provided a link to did use data from individuals as Dr.

> King suggested.



	-- The problem is that these studies do not provide a

statistically meaningful result in the low level region of interest



> As I told Dr. Cohen a few weeks ago, I agree with the papers by Smith et al.

> that describe the limitations of Dr. Cohen's work.  The smoking data he uses

> is so bad, it can only predict a little over 30% of the the lung cancers in

> the counties. If there was no confounding, it should be able to predict 85%

> or so.  I do not call that good control of confounding.



	--As I responded to you a few weeks ago, you apparently do not

understand the meaning of R-squared. This is clarified in my paper in

Health Phys. 72:489-490;1997. Note also the point above that even if my

smoking prevalences were all wrong and there were a perfect negative

correlation between smoking prevalence and radon exposure, my conclusions

would not be changed.



	--If you have reservations about my work, why not write a paper or

letter to Health Physics about them? Isn't that the normal procedure? I

responded to the criticisms of Field et al (that you support) as far

as I was allowed; the Editor decided they should have the last word. But

if you raise an issue again, I will get a chance to respond.







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

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.