[supercat]

Interesting example of data double-speak...

Statistical homeopathy can be a wonderful magical thing. Under traditional statistical methods, if someone gets cancer after having been exposed to what would generally be regarded as a trivial amount of smoke, that would be evidence reducing the apparent correlation between smoking and cancer (I'm not trying to suggest such a correlation does not exist, but accurate judgement of its magnitude requires recognizing the existence of non-smoking-related cancer). By classifying such cancers as smoking related, however, the statistical homeopath can enormously magnify the apparent harm of smoking. Indeed, cigarette smoke is so powerful that it could kill a person who's never been within 100' of a cigarette, if such a person existed.

[babygene]

"(I'm not trying to suggest such a correlation does not exist, but accurate judgement of its magnitude requires recognizing the existence of non-smoking-related cancer)."

Here's what I don't understand...

Clearly, the active ingredient and addictive ingredient of tobacco smoke (nicotine) does not cause cancer. Nicotine chewing gum that is available without a prescription carries NO cancer warning.

Cigarettes have hundreds of other ingredients, and the ingredients vary from brand to brand. Why has no one tried to identify the ingredient(s) that cause the "cancer" and remove them? Or identify the brands that don't contain those substances.