3 ms·

TL;DR: Among light and moderate drinkers as a group, 27 of the 30 factors were correlated with drinking levels, and in the direction that could create a false a

TL;DR: Among light and moderate drinkers as a group, 27 of the 30 factors were correlated with drinking levels, and in the direction that could create a false appearance of moderate drinking leading to better health. For example, as people enter old age, they drink less and become more likely to die. ... In light of these facts, Occam’s Razor argues for a simple theory: the net marginal impact of drinking on health is negative at all levels; and moderate drinking is a marker for relative youth, affluence, and healthy habits rather than a cause of good health. I guess he's suggesting that the apparent positive correlation between moderate drinking and health is an example of Simpson's paradox [0]. [0] http://en.wikipedia.org/wiki/Simpson%27s_paradox http://en.wikipedia.org/wiki/Simpson%27s_paradox

mulligan 12y ago
From the article: argues for a simple theory: the net marginal impact of drinking on health is negative at all levels; and moderate drinking is a marker for relative youth, affluence, and healthy habits rather than a cause of good health
coldtea 12y ago
>In light of these facts, Occam’s Razor argues for a simple theory: the net marginal impact of drinking on health is negative at all levels; and moderate drinking is a marker for relative youth, affluence, and healthy habits rather than a cause of good health. That would be true only if all studies failed to control for such parameters, which sounds BS. According to Occam’s Razor it's more probably that the author is yielding Occam's Razor wrongly, than that all those researchers are wrong and fail at such basic research methodology as controlling for age and those other parameters...