Almost 18 million people died in the U.S. from cancer between 1950 and 1997. 15,000 deaths in this number is a miniscule fraction--what would be considered noise in any other study. What statistical leaps of faith are required to connect these deaths to a particular nucleotide from a particular source? I have some knowledge of how cancer deaths by radiation are calculated, and I can tell you it is more guesswork than anything else. One assumption often made is that deaths vs. dosage are a straight line curve, which ignores the hormesis effect.

Even the report writers say that the study could have a "considerable margin of error," and that "the progress report makes clear that the study is not meant to be a conclusive finding but a preliminary look at the feasibility of in-depth research." Could anything be more vague? What's the point of this, other than fishing for research money? Read the rest of the article--most of the other people quoted are already talking as though this 15,000 figure, pulled out of someone's ass, is carved in stone.

Should this be studied further? Maybe. Is there really any feasible way in which causality can be established between atmospheric radiation barely over background, and specific cancer cases? Hardly a day goes by that some environmental factor isn't "connected" to cancer deaths--some with more credibility than others.

The original articles are here, and I will be looking them over, but I bet they don't show the work.