residual confounding

["John Williams" <JohnWi@law.com>]

Dr. Cohen,



I think it is important to determine the the source of the strong 

inverse association between radon and smoking.  Many feel that this 

inverse relationship is what is driving your findings.  There is 

residual confounding of your smoking data that can not be corrected 

because of the cross-level bias inherent in the aggregate form of the 

data.  The residual confounding goes along with the other inverse 

relationships seen if you use your data for other cancers.  But, the 

inverse relationship is less for other cancers because the smoking 

effect is a smaller risk factor in those cancers.  







That is not for me to say. Correlations of various kinds appear in

all data. I have considered over 500 potential confounding factors and

essentially all of them have some correlation with radon. The 

important

question is how strong these correlations are. Studying them serves 

as the

basis for my method of using "plausibility of correlation" to evaluate

confounding effects. For example, if all 500 have correlations less 

than

0.5, it is highly implausible for other untested factors to have

correlations as high as 0.9. You can read about this in my paper on

"Treatment of confounding factors in an Ecological Study" which is 

posted

on my web site. The correlation between smoking and radon is not 

outside

of the distribution I find for other potential confounding factors, 

and is

thus not implausible.









Sent by Law  Mail

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

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/