Monday, May 10, 2004

Weighty Considerations

The comments on my previous post on obesity and overweightness got me thinking more about the issue. Specifically, Tim noted:
Obesity and anorexia are both bad things. People should have reasonable expectations for their body weight, and for an average-height woman, 100 pounds and 200 pounds are both unhealthy. The people prone to obesity are typically different from the people prone to unhealthy dieting. They have different problems, and therefore need to hear different messages.
Fair enough, although I wonder if the messages are reaching the intended targets. Maybe the anorexics are hearing the anti-fat messages, while the obese people are hearing the anti-skinny messages.

Tim’s point also made me wonder about the prevalence of the different problems. In the previous post, I said the average American woman is 5’ 4” and 140 pounds. I assumed, given the recent reports that 60%+ of Americans are overweight or obese, that 140 pounds was too much for a 5’ 4” woman according to the new government standards. But I just plugged the numbers into a Body Mass Index (BMI) calculator, and it turns out that 5’ 4” and 140 pounds is in the government-approved range.

So what’s going on here? How can over half the public be fat, when the public still has an average weight that’s just fine? I can see a few possible explanations.

• First, the mean and median are not the same. It could be that, once the anorexics in the bottom 40% of the population are taken into account, their extreme underweightness balances out the top 60%’s overweightness to create a “healthy” mean, even though the median person is still overweight. But this strikes me as an unlikely explanation. It’s easier to be 200 pounds overweight than 200 pounds underweight; therefore, I would guess the weight distribution is positively skewed (much like the income distribution), leading to a mean that’s higher than the median, not lower.
• Second, a gender difference could resolve the paradox. Perhaps men are overrepresented in the overweight/obese category, meaning that a smaller percentage of women are overweight/obese. However, this fact sheet says that 61.9% of women over 20 years old are overweight/obese, so the paradox remains. (The percentage of men who are overweight/obese is indeed higher, though.)
• The comparison between the average woman figures and the percent overweight/obese figures is confounded by the latter being calculated based on BMI, which considers height as well as weight. Maybe there are a lot of short and fat women who still have relatively low weights compared to women of average or greater height, and their weights pull down the average even though they are overweight. I haven’t fully worked through the logic on this one, but my instinct is that it means short women must be more likely than tall women to be overweight.
• The “average woman” is not really an average woman, but an ideal woman according to the government’s standards.

Anyone who has the answer is invited to email it to me or put it in the comments box. Incidentally, nothing I’ve said here should be taken to mean I agree with the government’s standards for healthy BMI. Read this article to find out how the government actually changed the criteria, yielding a sudden jump in the number of people classified as overweight.

No comments: