RE: Cohen's Ecologic Studies

[epirad@mchsi.com]

Ted,



A specific, credible example is cross-level bias and 

confounding at the group level.  It is up to Dr. Cohen 

to demonstrate this is not the cause.  I also provided 

the paper by Sheppard et al. which presents methods he 

can possibly use to prove it is not the cause.



>From the perspective of a non-epidemiologist, I can 

understand why you think a specific, credible example is 

needed to refute Cohen's findings.  However, if you read 

our papers in Health Physics, you may understand why it 

is both not needed and often not possible to provide.  



The majority of this following text is taken from the 

rejoinder from years ago, but it is very applicable to 

your comment today.  The ecologic study design and 

existing data precludes us from providing an explanation 

in definitive analytical terms.  Greenland and Robins 

(1994) and Lubin (1998, 2002) have already presented 

plausible theoretical examples of how Cohen’s data can 

produce incorrect and even contradictory risk estimates. 

Greenland and Morgenstern (1989) previously pointed out 

that it is not possible to identify empirical sources of 

ecologic bias from aggregate data alone. Rather, 

researchers must rely on prior knowledge of intergroup 

variation in the distribution of other risk factors and 

effect modifiers.  Morgenstern (1982) has shown that 

factors responsible for ecologic bias may not be 

confounders or effect modifiers at the individual level 

and that identifying the bias is even more difficult 

because factors may not even “appear” to be confounders 

or modifiers at the ecologic level.  These ecologic 

biases can even reverse the direction of an observed 

association, especially when the range of average 

exposure levels across groups is small or the exposure 

under study is not a strong risk factor (Greenland and 

Morgenstern 1989; Richardson et al. 1987).   



In addition, if his findings are valid, he should not be 

finding an inverse association with other smoking 

related cancers.



I agree with others who feel the burden of proof is on 

Cohen to show that he is right and everyone else is 

wrong. 



http://www.vanderbilt.edu/radsafe/0003/msg00651.html

  

Regards, Bill Field

> I don't know why you guys refuse to face the issue.  You argue that

> ecological studies have certain limitations.  No one disputes that.  You

> keep describing what these potential problems are.  No argument.  But no one

> has shown how, with a specific, credible example, such a problem could

> dispute the actual data Dr. Cohen has presented.

> 

> This is like arguing that a clock might not give the right time.  It might

> not have been wound.  Or reset after winding.  Or be in the wrong time zone.

> Or be a lousy clock.  All true.  But when a guy says "My clock says the time

> is now 8:15,' do you sit around and present the kind of arguments I

> described, or do you check out the specifics?  What Cohen keeps asking for

> is:  Give me an example of a credible situation that could reverse his

> negative correlation.  He agrees to calculate the results, so you don't even

> have to do any work.

> 

> Seems like a reasonable request.

> 

> Ted Rockwell

> 

> 

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

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/