Re: Recommendation to resolve Cohen's ecologic study questions

["Field, R. William" <bill-field@UIOWA.EDU>]

Dr. Cohen,



The explanation is simply that you have group level confounding that you 

have not accounted for.



Smoking (among other factors) confounds your radon covariate at the 

regional level, but most of the smoking variability occurs within counties 

rather than between counties.  Therefore, as suggested in Guthrie's paper, 

you need county level population incidence data and covariate data from a 

sample survey.  As we pointed out before, in order to explain your findings 

you will likely need more data than is available from existing published 

sources.



Would you like me to fax you a copy of Guthrie's paper?



Bill Field



At 01:08 PM 01/23/2002 -0500, you wrote:



>On Tue, 22 Jan 2002, Field, R. William wrote:

> >

> > Here is my specific recommendation.  If you are really interested in using

> > your ecologic data and getting at the potential sources of your bias, you

> > need to identify what inter county adjustments need to be made by using

> > more innovative methods.  Possibly those suggested in: Guthrie, KA and

> > Sheppard, L, Overcoming biases and misconceptions in ecological studies,

> > J.R. Statist. Soc., volume 164, pp 141-154, 2001.

>

>         --Where is your specific suggestion for an explanation for my

>discrepancy with LNT? I ask for a specific suggested cause, not for a

>suggested procedure. I am surely entitled to any procedure I like as long

>as it justifies my conclusion.



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

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/