Is fat hatred contagious?

"There are three kinds of lies:
lies, damned lies, and statistics."

                                  Benjamin Disraeli

The researchers behind the recently published New England Journal of Medicine study claiming to document "the spread of obesity in a large social network over 32 years" have apparently never heard of Disraeli's comment on statistics (popularized by Mark Twain) -- or have conveniently forgotten it in their eagerness to use statistical analyses and a nifty Social Network Image Animator to purportedly track "the person-to-person spread of obesity."

And the media, in parroting the study's claims to have documented the "social contagion" of obesity, are fueling the war on fat people by uncritically reporting that "obesity" is contagious.

What can we expect next? Quarantine? Apartheid?

Nicholas A. Christakis, M.D., Ph.D., M.P.H. and James H. Fowler, Ph.D. were researching the "social spread of obesity" as they mined a portion of the almost 60 years' worth of data collected in the Framingham (Massachusetts) Heart Study. In almost all of the nine media reports of this research I found in a quick web search, there was no differentiation between physical/medical contagion (the spread of a disease) and social contagion (the spread of ideas and behavior).

The research is being reported in a biased manner, which I'll go into in a minute. But I suppose that's only to be expected, since the research was biased to begin with.

The researchers started -- and apparently ended -- with assumptions that fatness is always unhealthy and undesirable, weight loss is always healthy and desirable, and that permanent, safe weight loss is achievable. None of those assumptions is empirically supported, and each has been debunked in research documented in books ranging from Glenn Gaesser, Ph.D.'s Big Fat Lies: The Truth About Your Weight and Your Health to Fat Politics by J. Eric Oliver, Ph.D. and Rethinking Thin: The New Science of Weight Loss and the Myths and Realities of Dieting by New York Times science reporter Gina Kolata.

What the Christakis-Fowler research actually shows is the social contagion of fat hatred, especially in regard to the way it's being disseminated and reported.

For example, in my admittedly non-random sample of media reports, the research was consistently reported as indicating that the body mass index of "friends" (no gender indicated) influences each other. (Yes, they confused correlation with causality. That seems to be standard in "obesity" research.) Every photo or video or audio clip purportedly illustrating that "obesity" is socially contagious among friends featured women. (Women who joined Weight Watchers, in at least two instances -- making me wonder if the diet industry rather than the National Institute of Aging might be the real funder of the study. The exception being the headless-torso-photos so typical of fat-related journalism, in which gender can't be determined.) But the original study found that the so-called "influence" of friends in regard to BMI only existed in male same-sex friendships.

Among friends of the same sex, a man had a 100%...increase in the chance of becoming obese if his male friend became obese, whereas the female-to-female spread of obesity was not significant...

Yet not one of the media outlets reported this fact. Indeed, by illustrating their stories with female friends, they falsely implied that the friends-make-friends-obese "connection" was true of female-female friendships. Which I find particularly egregious given that women are more preoccupied by -- and penalized for -- fatness than are men, and have a higher incidence of eating disorders.

(You know, somebody ought to point out to Rush Limbaugh than this new study could be used to blame him and his body-mass cohort for the so-called "obesity epidemic." That might make an -- ahem -- interesting show topic.)

But in the media's haste to keep the "war on obesity" flaming -- fueling it in part by their own prejudices and assumptions about fat -- they're missing a far more important story, in my opinion. No, not "why did the New England Journal of Medicine publish this p.o.s. in the first place?" (although that's a good question), but "what effect is this research having on fat people?"

Because the fat people (and their allies) who know the truth about the "myths and realities of dieting" are horrified. And many are scared.

They (we) see this study being used to justify more intrusive public health programs and actions. More fat children being removed from their homes to "save" them from their allegedly fattening parents. (Well, maybe not "allegedly" fattening, since "obesity" has been shown to be highly heritable, a fact the social-contagion-study researchers discount.) More shunning of fat kids and adults, more job and academic and social discrimination.

Indeed, I see the justification and agenda for this study in the reseacrhers' conclusion that

The spread of obesity in social networks appears to be a factor in the obesity epidemic. Yet the relevance of social influence also suggests that it may be possible to harness this same force to slow the spread of obesity....medical and public health interventions might be more cost-effective than initially supposed, since health improvements in one person might spread to others. [Assumes thinness and/or weight loss equals good health.]....This highlights the necessity of approaching obesity not only as a clinical problem but also as a public health problem. (Emphases mine)

More money for public health initiatives! (And those whose salaries depend on them.) More money for weight loss treatments, especially group-focused interventions! (Never mind that Weight Watchers, for instance, has been offering dieting-based groups for more than 40 years, with no demonstrable effect on the nation's waistline except, possibly, to expand it through the yo-yo-dieting weight-loss-and-regain effect.)

The first reports of this study indicating "obesity is contagious" (reports assuming, of course, that "obesity" is a terrible thing) brought chilling visions of quarantine, apartheid, and even lynching to some minds. Including mine.

When my husband came across a report of the study online, he emailed the link to me. He thought it was funny and couldn't imagine anyone taking it seriously. "Yeah, you laugh until they put me in a boxcar and send me to Fat Land," I told him. (That was before I learned the strongest "effects" of this so-called contagion were male to male.) My husband's reaction? "Well, you know where the gun is." That's a Southern male response for you.

Here's a sampling of the way this research was reported. As you read, keep in mind that the "friends" effect was only for male same-sex friendships, and the average amount of weight a friend gained when his male friend was "obese" was only 5 pounds.

Time magazine (Laura Blue):
Obesity Is Contagious, Study Finds
Wondering why your waistline is expanding?  Have a look at those of your friends. Your close friends can influence your weight even more than genes or your family members, according to new research appearing in the July 26 issue of The New England Journal of Medicine.

NPR Morning Edition (Allison Aubrey):
Are Your Friends Making You Fat?
A new study suggests that your best friend's weight may be very influential in determining whether you'll gain or lose weight over the years.

NewScientist.com
Is obesity contagious? (Roxanne Khamsi):
They say it's not what you know but who you know. A new study suggests that this aphorism might hold true when it comes to your body weight.

MSNBC.com (Reuters):
Your best friend could be making you fat
If your BFF gets obese, study says you're likely to follow suit
Could your best friend make you fat?

Researchers who have studied "networks" of obesity think so; they found that if someone's friend becomes obese, that person's chances of becoming obese increase by more than half.

BBC News (no attribution):
Obesity 'contagious', experts say
Having a friend, sibling or spouse who is overweight raises a person's risk of being obese too, US researchers say

The New York Times (Gina Kolata*):
Find Yourself Packing It On? Blame Friends
Obesity can spread from person to person, much like a virus, researchers are reporting today. When one person gains weight, close friends tend to gain weight, too.

*Yes, that Gina Kolata, the author of the recently published book Rethinking Thin, who has apparently dissociated the contents of her book, including her ending plea that we "lay off the beleaguered fat people."

CBSNews.com Couric & Co (Dr. Jonathan LaPook):
Is Obesity Contagious?
...If one of a pair of mutual friends BECOMES obese (defined as Body Mass Index, or BMI >= 30) then the risk of the other becoming obese increases by 171%!

My Way News (Alicia Chang):
Study: Obesity is 'Socially Contagious'
If your friends and family get fat, chances are you will, too, researchers report in a startling new study that suggests obesity is "socially contagious" and can spread easily from person to person.

Each of these stories, of course, leads with the "friends can make you fat" claim, but fails to clarify that the friendship "contagion" was only significant in male same-sex relationships. Of the nine articles I read, only one, by WebMD, clearly stated in the headline or lead paragraph that the contagion wasn't "about viruses or germs." But WebMD reporter Miranda Hitti still reported the friendship-obesity claims without clarifying that they applied to male same-sex friendships only.

This might be an example of the assumption of male privilege by the male researchers -- findings applying only to men being reported as if they were true of men and women both. I have a hard time imagining that findings applying only to women wouldn't be clearly indicated as such. I suspect the consistent failure of reporters to clarify the friendship gender issue stems from their not having read the original research and relying instead on a press release that spins the research in the direction most likely to fuel the war on "obesity."

Here are some more gems from the media reports:

The study's authors suggest that obesity isn't just spreading; rather, it may be contagious between people, like a common cold. (Time; emphasis mine)

Confuses social contagion and physical contagion.

"You are what you eat isn't the end of the story," says study co-author James Fowler, a political scientist at UC San Diego. "You are what you and your friends eat." (Time)

Erroneous assumption that fatness is simply and directly related to the amount and type of food ingested. Fat people can eat healthily and still be fat, just as some thin people can eat unhealthily and still be thin.

Having fat friends makes being fat seem more acceptable. (Time)

Assumes fat bodies are "unacceptable."

Working on your weight? Enlist your social network.

"If you're going on a diet, then you want to convince them to go on a diet. If you want to start to run or to change your exercise behavior, you also want to encourage your friends to engage in those behaviors," says Fowler. (WebMD)

Like no one has ever thought of this before. Well, maybe men aren't as likely to make dieting a social or group activity as women are. Or maybe Fowler, being a political scientist rather than a clinician, doesn't know beans about the history of dieting. Or of Weight Watchers and TOPS (Take Off Pounds Sensibly) and Overeaters Anonymous and other weight-oriented groups.

If the new research is correct [well, thanks for that "if"], it may say that something in the environment seeded what some call an obesity epidemic, making a few people gain weight. Then social networks let the obesity spread rapidly. (New York Times; emphasis and sarcasm mine)

Maybe to stop this rapid spread of "obesity" they should train people to yell "Fat! Fat!" when they spot an "obese" person, and "drop and roll" if they get too close. Or they could create fatbreaks -- you know, places that are inaccessible to fat people and thus could stop the spread. Oh, wait -- there are plenty of those places already, and they apparently haven't had an effect upon the conflagration.

It may also mean that the way to avoid becoming fat is to avoid having fat friends. (New York Times)

That's the take-home message being blared by the media, even though the researchers claim it's not their intention. (Just what we need, more isolation, discrimination, and shaming. After all, discrimination and stress have great effects on health -- oh, wait. No, they don't.) Of course this statement, just as every other media report as well as the original journal article, confuses correlation with causality.

Almost all the media reported the study uncritically. Only the New York Times, My Way and BBC News articles (among those I reviewed) included comments from other scientists that could be considered mildly cautious about the study, but those remarks were buried at the end of the articles and were nowhere near as critical as the research warranted.

Other problems with the study:

• The data was reported as if it applied to all Americans and all social classes, even though it was a non-random sampling of a white middle-class community. (The NewScientist.com article actually reports that Christakis and Fowler say the study "might mean that the problem of obesity [a phrase reminiscient of "The ________ Problem"....fill in the blank with the population of your choice] could spread more quickly through small towns -- where social networks tend to overlap more than in big cities." [Emphasis mine.] Maybe we should stamp out small towns. Or rigorously police them to make sure the fatties stay away from the thinnies.
  

• The "friends" and family members included in the study were only those whose names the Framingham Heart Study (FHS) participants had given to the FHS researchers when asked for emergency contacts -- and of those people, only ones who were also participating in the heart study were included in the "social contagion" analyses. Christakis and Fowler apparently excluded any "friends" or family members who were not also in the FHS, because weight data weren't available for those people.

• Using BMI as an indicator of "obesity" is inherently problematic especially for men (for whom the only "friendship" effects were found, remember), because the BMI formula doesn't discriminate between muscle and adipose tissue. Muscular men could be identified as "obese" even though their body fat percentage might be within "ideal" range.

It appears that Christakis and Fowler pulled social data from a study not intended for that purpose, tossed it into a statistical program, and looked for patterns that fit their preconceived notions -- err, hypotheses.

That reminds me of the scene from the film A Beautiful Mind in which Russell Crowe shows Jennifer Connelly how he could find any image she named in the stars. Given enough data (or a starry night), you can find find all sorts of patterns and assign the meaning to them you wish.

In the BBC News article, Christakis says "It's not that obese or non-obese people simply find other similar people to hang out with." [Actually, the researchers claim this isn't the case, but I suspect it's just one of the constellations they wanted to see in the data. It's not at all clear in the journal article that their data disprove this.]

"Rather, there is a direct, causal relationship.[Correlation does not equal causality. Twinkle, twinkle, little star...] What appears to be happening is that a person becoming obese most likely causes a change of norms about what counts as an appropriate body size.

"People come to think that it is OK to be bigger since those around them are bigger, and this sensibility spreads." [Horrors.]

So there's another take-home message from the researchers: size acceptance -- or, heaven forbid, fat acceptance -- is dangerous because it can seduce others into fatness.

Of course,it is true that those who recognize natural diversity in body size and champion size -- even FAT -- acceptance might, yes, influence others to stop food restriction and other unhealthy weight loss practices. And that might result in those individuals' regaining any lost weight -- although they'd eventually regain anyway, since there is no diet or drug or exercise or medical "treatment" proven to result in permanent, safe weight loss. Size/fat acceptance just helps them understand that the failure lies in the diet (or other intervention), rather than the individual.

But is that dangerous? I think it's most threatening to those whose livelihoods are dependent on the notion that fatness is a killer disease that must be eradicated.

Earlier today I shipped a box of books published by my publishing company, Pearlsong Press, to the young woman who will be our first writing intern this fall. I wanted her to become acquainted not only with the books and authors, but with the company's mission and vision of publishing only books that promote (or at least do not contradict) size acceptance and Health At Every Size, the weight-neutral approach to well-being.

The first books I published, in 2004, were Pat Ballard's short story collection and romance novels featuring plus-sized heroines. (Or, as I like to put it, books in which the fat chick gets the guy.) Romance novels might seem like lightweight armor in the war against fat people, but I think it's important that people be able to see positive images of themselves in the world around them. I'm old enough to remember when there were no African American people on television in any role except servants. It was a big deal when Diahann Carrol became the first black woman to star as a professional woman on television, in the show Julia. It's almost as rare today to see fat women portrayed positively in television, film, and literature. (I'm talking fat, here, not Bridget-Jones-chubby .... although even that, too, is rare.)

Anyway, as the postal clerk weighed the book package this afternoon she went through her usual litany, asking me if the box contained any liquid, flammable, or hazardous material.

"Just dangerous ideas," I said.

UPDATE 7/30/2007: Sandy Swzarc has an excellent post on the study and its "virtual reality" at her Junkfood Science blog.

She points out that although the NEJM reported the authors had no conflicts of interest, in fact Christakis is on the Robert Wood Johnson Scholars in Health Policy Program at Harvard, which is a program of the Robert Wood Johnson Foundation. The RWJF is the largest organization in the world funding policies against "obesity." (See my April 4, 2007 post about the foundation funding the war on fat kids here.) And they call that "no conflict of interest"...?

Comments

[Sandy Schaffer]

Great article Peggy. Beautifully written and so well researched. I would love to post this article on my NY NAAFA site.

[Kate Harding]

Peggy, I adore you for this entire post, but especially for rounding up outrageous quotes from the mainstream articles. I wanted to do that but don't have time today!

[Cynthia]

I've taken to calling this contagious obesity theory the cooties philosphy of fatness.

[Susan Stinson]

Hi Peggy. That's a great analysis. I linked to it here, along with some information about a call for people's stories in support of the proposed Massachusetts law against discrimination based on size. If you'd like to help spread the word -- or have stories of your own -- that would be great.

And how very, very cool that you've recently read Venus of Chalk!

[stef]

I blogged this post here. Excerpt:
The main thing that's different about this particular story is that the hatred and fear are being openly voiced, by mainstream supposedly objective news media, as advice. The hatred is usually more couched in pseudo-concern or "too bad, but this is the way it is" language, but here fat people are being portrayed as actual enemies of thinness to be actively defended against, rather than as victims or losers.

[Peggy Elam, Ph.D.]

My apologies for the delay in publishing these comments -- I left town to attend a family wedding shortly after writing the blog post, and just returned home this afternoon. Didn't have email/online access in the interim.

Thanks for the positive comments. (Susan, I loved VENUS OF CHALK!)

D.B.A., I hurriedly published all comments before reading them, and only afterward read yours and felt it to be inappropriately supportive of the NEJM study and weight loss as well as condescending to fat people. (If you were being sarcastic, I apologize -- that wasn't apparent in my quick read.) I didn't see a way to edit your post after I'd published it, so I deleted it. There is enough of that already on the web. I won't have it here.

[James Fowler]

Hi Dr. Elam,

Thanks very much for your work in fighting discrimination. I read your article with great interest and I very much appreciate the effort to counteract media coverage that plays on stereotypes. However, I feel compelled to respond to some of the criticisms that have been raised.

1) For the record, I am a big fan of Mark Twain and I know the quote about statistics. It is certainly possible to use statistics to cherry-pick conclusions. That's why I'm extraordinarily careful whenever I study a problem to conduct several different kinds of analyses on the same data. I encourage your readers not only to read the paper but to examine the supplementary information. We literally sliced this data 1000 ways and got the same conclusions every time, which we report in the paper. And this data will soon be available to other scholars who wish to study the social network in the Framingham Heart Study.

2) You say that we assume that weight loss is always healthy, but this is not true. In fact, in the press conference we held on monday before the article was published, there was a discussion specifically about Bulimia in which I note that weight gain may actually be the healthy behavior and weight loss the unhealthy behavior--so it depends on context. Given the strong evidence for both the increase of obesity and its linkage to disability, especially in older Americans, we thought readers of NEJM (especially doctors) would most likely be concerned with weight gain, so we mainly described our results in those terms. However, we also noted that our results apply to the spread of weight loss, as well.

3) On the issue of correlation vs. causation, I think you are missing an important and unique element in the study. We showed that these correlations are directional -- if Joe names Ahmad as a friend and Ahmad gains weight, then Joe will also gain weight. But the reverse is not true -- Ahmad is unaffected by Joe. The same holds true for losing weight. If it were just some sort of association through shared experience, the effect would be equally strong in both directions. But it's not. Thus, unlike the typical observational study, we have evidence that influence is flowing causally from one person to another.

4) On the issue of men vs. women in same sex relationships, if you'll notice in the statistical models, the difference between men and women is not statistically meaningful. In other words, we can't be sure that males have a greater impact on each other than females have on each other. If it were an election, we'd say "it's too close to call." As a result, the standard procedure is to rely on the pooled results for both men and women. We presented these results separately to demonstrate that pooling was appropriate, and that it made sense to talk generically about same sex friendships of either gender.

5) You are certainly welcome to your opinions, and I emphasize that I appreciate the effort to reduce discrimination, but I'm not sure what purpose personal attacks serve (unless I misunderstand the meaning of p.o.s.).

6) On the issue of contagion, we actually point out in the article that the fact that neighbors have no effect and the fact that the effect is not reduced by distance between friends means that a recent theory that obesity is spread via a pathogen is unsupported. What we are talking about is the spread of ideas about behaviors between one's closest friends. The media way over-generalized the result in many cases and gave the impression that physical proximity to a person was sufficient to generate the effect, which was false.

7) I spent the whole week trying to combat the "dump your fat friends argument." For instance, I noted over and over again that the social support literature tells us every friend makes you healthier, regardless of whether or not they are obese. I also emphasized that our study said nothing about what happens when we change friends -- only what happens when our current friends change weight.

8) On the issue of using data "not intended for that purpose", I urge you to search for "framingham obesity" in Google Scholar. There have been dozens of studies of obesity using the Framingham Heart Study data, from the 1960s to today. We created the social network data for our study from previously unused information, but Lisa Berkman has also used surveys on social support in the Framingham data to study the effect of friends on health (Berkman's seminal work contributed a lot to our knowledge that additional friends make us healthier).

9) Finally, on a personal note, six of my closest family members are obese (or were obese before they passed away). One of them has just been diagnosed with type II diabetes at 33 years of age. So, like you, I have a very personal interest (I think you might call it a conflict of interest) in obesity. For the sake of my family, it would be nice if we understood the relationship between health and obesity better, and I very much hope that your work and my work brings us closer to that understanding.

All my best,
james fowler
ucsd

[fatfu]

"3) On the issue of correlation vs. causation, I think you are missing an important and unique element in the study. We showed that these correlations are directional -- if Joe names Ahmad as a friend and Ahmad gains weight, then Joe will also gain weight. But the reverse is not true -- Ahmad is unaffected by Joe. The same holds true for losing weight. If it were just some sort of association through shared experience, the effect would be equally strong in both directions. But it's not. Thus, unlike the typical observational study, we have evidence that influence is flowing causally from one person to another.

Dr. Fowler, with all due respect you did *not* show that it's directional. The difference between the two sets of friends was not statistically significant. It bothers me that you keep claiming this - since the idea of directionality underlies your entire hypothesis.

[Sandy]

Fatfu is right. Not only did this study use problematic data never intended to be used to study social relationships, an extremely problematic computer model that made a multitude of flawed clinical assumptions, but its weak hazard ratios would not be considered tenable for these types of studies (and just as likely due to chance, random error or a thousand other things). Sadly, the null findings were not reported as such. I looked at the most significant logical fallacies (correlation and directionality) in the post above, "Cats and email." Hope it helps.

[Peggy Elam, Ph.D.]

Dr. Fowler, I appreciate your efforts to counteract the media misrepresentation and hyping of your study. My bachelor's degree is in journalism and English, and I was a newspaper journalist for several years before going to grad school in clinical psychology, so I am particularly concerned by the way this and other research is reported in the media.

I did not intend the "p.o.s." comment to personally attack you or your co-author, but rather the study itself. However, I can see how you would take it personally, just as many fat people across the country have taken the study and its reporting personally, and for that I apologize. I probably would have chosen another term had I not been writing in the heat of my and others' emotional response to the study and its reports. As it was, I still have enough of my Southern Baptist upbringing that I wouldn't spell the term out. :-)

I am very sorry to hear about the health problems and deaths of people close to you who are or were fat. I can understand how that could fuel your concern about weight. However, I wonder if you have considered that family and social concern (even judgement and disapproval) about weight/fatness can itself contribute to unhealthy shame and isolation?

I still think focus on weight itself, rather than on healthy behaviors, is misguided -- in large part because of the negative effect fat stigma, prejudice and discrimination (which are reinforced/encouraged by "obesity prevention" campaigns and the like) can have on emotional and physical health. And also, of course, because weight-loss dieting (even disguised as "lifestyle changes") not only doesn't work in the long run, but has actually been shown to cause harm. (I refer you to Linda Bacon and colleagues' study comparing a Health At Every Size approach to a traditional weight-loss approach for "obese" middle-aged women, published in the Journal of the American Dietetic Association a couple of years ago.) Glenn Gaesser, J. Eric Oliver and Gina Kolata also provide a good review of decades of research on dieting & weight in their books "Big Fat Lies: The Truth About Your Weight and Your Health," "Fat Politics," and "Rethinking Thin: The New Science of Weight Loss and the Myths and Realities of Dieting," respectively.

I've found that many fat people with Type II diabetes, especially women, react to weight-loss-related food restriction with dieting-related bingeing, body shame and despair that does them no good. Whereas simply focusing on regular exercise and good nutrition (and stress reduction and other factors) can get their diabetes under control without much -- or any -- weight loss, and in some cases even without medication.

[S Sanders, MA]

Great article!!!! I have also done a deconstruction of this study which I put on Daily Kos. I will just put the link here:

http://www.dailykos.com/story/2007/7/28/84516/3178

I have been posted this everywhere to combat the media hype and have met with many favorable responses. Anyone who wishes may reproduce it in any form.

UNBELIEVABLE that Fowler thinks he proved causation. Dr Christakis e-mailed me back and also provided the same "we have fat family" song and dance. If they really cared about this issue, they would admit their study has limitations! And if they think they proved causation, they need to go back to research 101!