Re: Mechanisms are Needed to Explain Cohen's Data

[RuthWeiner@AOL.COM]

Maybe I am mistaken, and I haven't looked at Benrie Cohen's papers for some time, but I seem to recall that the idea was more to show the absence of a positive (linear) correlation than to demonstrate a negative correlation.  Certainly, one can conclude that  there is no positive correlation, let alone a linear one.


The burden of proof is considerably greater and more difficult if one wishes to demonstrate a correlation, whether positive or negative.


Also, I don't understand how epidemiologic studies can demonstrate a mechanism of action.  They can lead to demonstration of causality, but a cause is not necessarily a mechanism, and may not elucidate a mechanism.  For example, the recent elucidation of a heliobacter as the cause of stomach ulcers.  Whatever epidemiology was done pointed to "stress".  Careful case-control studiesopinted to the cause -- the heliobacter.  Then the mechanism could be studied.


Ruth Weiner, Ph. D.

ruthweiner@aol.com