RE: Risks of low level radiation - New Scientist Article

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

On Fri, 7 Dec 2001, Jim Nelson wrote:



> Dr. Cohen,

>

> I understand the meaning of r-sqaured very well.  I do not see in the

> reference provided that you addresses this issue at all?  Jim



	--I will give you an extreme example: Suppose there is very little

spread in smoking rates in various counties, so this would cause very

little difference in lung cancer rates. But there are always

fluctuations up and down due to ethnic variations, medical services,

reporting variations, chemicals in the environment and in the food,

 respiratory illness, etc, etc, and just plain statistical variations;

these would cause more differences in the lung cancer rates than the very

small differences caused by the very small differences in smoking. Then

R-squared would be close to zero. The true test would be the number of

standard deviations by which the regression of lung cancer on smoking

differs from zero. On that basis, my smoking data test out very well.



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

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.