You and your Friend's Friend's Friend
For those of us not actively toiling in a university, most modern writing in the social sciences can be placed into one of three categories. The first category, which is vast, consists of the arcane and the incremental - those studies so obscure, or which advance scholarship so infinitesimally, that they can be safely ignored by the general reader. (Not that this work isn't important; it keeps academic publishing in business, and significant knowledge accretes in tiny drips on the way to tenure.) The second category consists of statistical proof of the obvious. (Some actual study findings published recently: "the parent-child relationship . . . commonly includes feelings of irritation, tension and ambivalence"; women are more likely to engage in casual sex with "an exceptionally attractive man"; and driving while text-messaging leads to "a substantial increase in the risk of being involved in a safety-critical event such as a crash." Thank you, social science!) And in the third category, which is surely the smallest, are works of brilliant originality that stimulate and enlighten and can sometimes even change the way we understand the world.
The Surprising Power of Our Social Networks and How They Shape Our Lives
But here's a funny thing: Some research, defying logic, manages to straddle the boundary between the second category and the third, seeming alternately (or is it simultaneously?) obvious and brilliant. "Connected," by Nicholas A. Christakis and James H. Fowler, is full of this kind of research. "What a colossal waste of money it is for social scientists to prove the obvious," the authors themselves write, characterizing a typical response they got to an attention-grabbing study they published two years ago in The New England Journal of Medicine.
That study, about how obesity can be contagious, has already won the authors a modicum of fame. Earlier this year Time magazine named Christakis, a professor of medicine, sociology and health care policy at Harvard, one of the 100 most influential people in the world. Fowler, an associate professor of political science at the University of California, San Diego, gained additional notice by providing the first statistical evidence of the "Colbert bump" - the boost of political support elected officials enjoy after appearing on "The Colbert Report." In their writing, they are endearingly excitable, ranging enthusiastically across science and culture to find gee-whiz insights and unexpected results that support their arguments. For instance: "A study published in the scientific journal Nature revealed that a typical article in Wikipedia was almost as accurate as a typical article in the Encyclopedia Britannica." And: "Fungi can even ‘collaborate' to find the best path through mazes into which they have been placed by human experimenters." And my favorite: "In fact, the model that best predicted the network structure of U.S. senators was that of social licking among cows."
Do the arguments and ideas present in this book merit such far-reaching exploration, not to mention the attention they have brought the authors? Quite possibly so. For starters, that obesity study. Poring through the meticulous records of the Framingham Heart Study, conducted from 1948 to the present in a small Massachusetts city, Christakis and Fowler mapped out the relationship of 12,067 people with more than 50,000 ties (or connections between friends and relatives) among them. Analyzing the network, the authors noticed that obese people tend to be friends with other obese people, while thin people tend to be friends with other thin people. On one level, this is obvious and unsurprising; birds of a feather and all that. But based on their reading of the data (which some other researchers have questioned), the authors concluded that the relationship was causal: being associated with overweight people, even indirectly, is likely to make you overweight.
As Christakis and Fowler (along with other researchers) have found, obesity spreads by contagion. So if your friend's friend's friend - whom you've never met, and lives a thousand miles away - gains weight, you're likely to gain weight, too. And if your friend's friend's friend loses weight, you're likely to lose weight, too.
Christakis and Fowler explore network contagion in everything from back pain (higher incidence spread from West Germany to East Germany after the fall of the Berlin Wall) to suicide (well known to spread throughout communities on occasion) to sex practices (such as the growing prevalence of oral sex among teenagers) to politics (where the denser your network of connections, the more ideologically intense and intractable your beliefs are likely to be). And while it's hardly surprising that emotion can be transmitted from person to person, the authors report that getting a $10,000 raise is less likely to make you happy than having a happy friend is - in fact, the raise is less likely to make you happy than is having a friend who has a friend who has a friend who is happy. They even argue - and this is sure to generate controversy - that the obsessive drive to create "nut free" environments is not the result of any real increase in children's allergies but rather something akin to an epidemic of adult hysteria, spread via network transmission.
How does network contagion work? What is the mechanism whereby your friend's friend's obesity is likely to make you fatter? Partly, it's a kind of peer pressure, or norming, effect, in which certain behaviors, or the social acceptance of certain behaviors, get transmitted across a network of acquaintances. In one example the authors give, Heather stops exercising and gains weight, which influences her friend Maria's thinking about what normal weight is, so that when Maria's other friend Amy (who has never met Heather) also stops her exercise regime, Maria is less likely to urge Amy to resume it. So Heather's weight gain influences Amy's, even though the two women never meet. Of course, there's an awful lot of supposition here; the authors are more convincing in arguing that this sort of contagion happens than in explaining how it happens.
The Surprising Power of Our Social Networks and How They Shape Our Lives
They're also pretty interesting on why it happens. During the early stages of human evolution, selective advantage was probably conferred on those individuals who lived in social networks and could share information about food or predators. The primatologist Robin Dunbar has argued that the human brain evolved to its present size to keep track of a network of 150 people. Dunbar argued furthermore that grooming (picking the nits out of the fur of the other individuals in your group), which is the behavior used by other primates to maintain relationships, becomes inefficient in a group of 150. So we evolved the capacity for language, a "less yucky and more efficient way to get to know our peers, since we can talk to several friends at once but only groom one at a time," as Christakis and Fowler put it.
As among primates, those humans who are best able to manipulate social networks to their advantage thrive, and that ability may be genetically encoded. Using a clever study of young twins, the authors observed that genes accounted for "46 percent of the variation in how popular the kids were." (This is either brilliant or the reductio ad absurdum of genetic analysis or both.) And in a series of "cooperation game" studies, in which altruistic behavior among a social network was rewarded with money, the most popular girls ended up with four times as much money as the least popular, proving once again the propensity of the rich to get richer - not to mention the compounding cruelty of unpopularity.
Network science has implications for public policy. By learning more about the structure of various networks, we can identify where the hubs are - the most "influenceable" nodes that are likely to spread an idea (or a behavior or a germ) quickest - and intervene at those points to stop the spread of, say, an unhealthy behavior (like smoking or overeating or suicide), or to promote a positive one (like voting or becoming an organ donor), or to vaccinate more efficiently against disease.
But the more interesting implications are philosophical. A social network, while not quite sentient, acquires its own agency; it wants things, and it wants us, the nodes of which it consists, to do certain things, whether to gain weight or have oral sex at age 13. Mathematical models of flocks of birds, or colonies of ants, or schools of fish reveal that while there is no central controlling director telling the birds to fly one direction or another, a collective intelligence somehow emerges, so that all the birds fly in the same direction at the same time. Christakis and Fowler argue that through network science we are discovering the same principle at work in humans - as individuals, we are part of a superorganism, a hivelike network that shapes our decisions. "A smoker may have as much control over quitting as a bird has to stop a flock from flying in a particular direction," they write. The authors take a benign view of this and argue that as we become aware of the networks in which we're enmeshed, we'll all be better off.
As described by the authors, network science has potential to be used for good. But then again, if all the strutting and fretting that we believe to be the product of our individual free will is really only the antlike scurrying of a collection of nodes, can anyone really be said to "use" the network? Or is the network always using us?
- You and your Friend's Friend's Friend (Scott Stossel, New York Times, 1th October 2009)