In a heterogeneous society, practice tends to be normative. That is why homosexual activists greatly exaggerate the prevalence of homosexuality--asserting, on the basis of a misreading of Kinsey's famous studies, that 10 percent of the population is homosexual, whereas the true figure is probably at most 2 percent. The more homosexuals there are, the stronger their claim to be normal, a claim that would fail in a society that had a strict moral code condemning homosexuality.
Similarly, the more fat people there are, the more being fat is seen as normal. A half century ago, when obesity and overweight were relatively rare in this country, fat people were regularly ridiculed by entertainers, and this ridicule helped to keep people thin. As more and more people become fat, fatness becomes more normal-seeming, and the ridicule ceases (though another factor is the march of "political correctness," which discourages criticism of people's weaknesses).
It makes sense, as the recent article in the New England Journal of Medicine finds, that friends' fatness would have an influence distinct from that of the culture as a whole. We each inhabit a subcommunity or subcommunities within the national (and world) community as a whole, and we are more likely to take our clues from these subgroups than from the broader community. In my own ingroup of 16 judges (11 active members of my court, 4 senior members, and 1 nominee, who will replace an active member who will be taking senior status), only 2 are overweight (12.5 percent), compared to a nationwide average of 66 percent. Among my other friends, judicial and otherwise, the percentage who are overweight is probably no greater than 12.5 percent.
But separating out common causes from social influence is difficult. My social network consists almost entirely of affluent, educated people who are knowledgeable about calories and about the health effects of overweight, who can easily afford both expensive alternatives to junk food and membership in health clubs or ownership of exercise equipment, and who (this of course is related to their affluence and education) have a low discount rate and thus do not neglect long-term consequences of current behavior. One expects these people to be thin even if they are uninfluenced by the weight of the other people in their network. And likewise at the other extreme, with networks composed of people who are poor and badly educated and have high discount rates, all of which are correlated with obesity.
Still, it is plausible that there would be some social influence within these networks. The reason is that there is no clear notion of an optimal weight. Nobody bothers to compute his body mass, which requires translating one's weight from pounds to kilograms and then dividing by the square of one's height in meters (the normal range for the body mass index so computed is 18 to 24), simple as this computation is. A 2006 survey by the Pew Research Center finds that while people are acutely aware of the weight problem, they tend to regard themselves as of normal weight even when they are overweight, and this tendency to self-deception can be expected to be greater the heavier the people they associate with are. If you weigh 180 pounds, though you should weigh only 150, but your friends weigh 200 pounds, you will tend to think of yourself as thin. As Becker explains through the concept of a social multiplier, when you weighed 150 pounds your thinness may have constrained your friends, but when you move up to 180 you exercise a lesser constraint.
The social multiplier effect can of course operate in either direction. Among young professional women, there is a cult of thinness that seems to illustrate the effect. But it would be difficult to initiate such a downward cycle among the currently overweight.
Overweight may also be related to a decline in social conformity. There is more variety in people's dress, hair length, etc., than there was fifty years ago. Maybe it has become easier to make judgments about people without relying on crude signals, such as physical appearance. Then the costs of a nonnormal appearance would fall, and this would help to explain the increase in overweight.
A further point has to do with sedentary life style. That is rightly regarded as a risk factor for obesity. But in addition, being sedentary reduces the cost of being obese, since the less active one is, the less one is impeded by being obese. Fatness tends to creep up on one, and the less costly it is, the less incentive one has to lose weight, which is difficult.