To see a proper critique of dubious statistics, look at The New Republic’s debunking of the Body Mass Index. This rough and ready statistic (weight in kilos divided by the square of height in metres) has become the standard number in claims about obesity, with values over 25 being taken as overweight, and values over 30 as obese.There are two big problems with this, according to TNR
(a) Since it doesn’t distinguish fat from muscle, the BMI is a lousy measure of fatness. Estimates of body fat percentage can be obtained reasonably reliably using calipers for skinfold measurements
(b) There’s little convincing evidence that fatness per se, as opposed to inactivity and lousy diet, poses a health risk

Point (a) is well-known, but the persistent use of the BMI reflects the perceived need to simplify health messages, even at the risk of distorting them. The most recent issue of the Scientific American has an article on rebuilding the food pyramid that makes the same point regarding the famous food pyramid. A revised and more complex version (refined starch including potatoes are bad, and unsaturated fats and oils are good) has now been issued.

On point (b) it’s difficult from the evidence in the paper to disentangle the problems associated with the BMI with the evidence regarding the effects of fatness. I suspect it’s somewhat moot, in that with regular exercise and a good diet, you’re unlikely to have excessive body fat, whatever your BMI says.

(Disclosure: As measured most recently, my BMI is 25.2)

Update As if to prove TNR wrong The Age has a story showing a big reduction in life expectancy for those who are obese (BMI>30) at age 40. The objection that obesity per se is not disentangled from phsyical inactivity still appears to apply.