All products featured on WIRED are independently selected by our editors. However, we may receive compensation from retailers and/or from purchases of products through these links.
LONG BEACH, Calif -- Online social networks may be new, but social networks themselves, of course, are not. Nicholas Christakis, a professor of medicine and sociology at Harvard University and co-author with James Fowler of Connected, says that social networks -- online or offline -- follow some basic rules and influence the people in them. Their structure also often determines what the network accomplishes.
Wired.com spoke with Christakis prior to his Thursday talk at the Technology, Entertainment and Design conference about how social networks and friends influence us and how friends of friends can sometimes influence us more than our friends themselves.
Wired: What purpose do social networks serve? Are they just a holdover from the tribal life of our ancestors to aid survival?
Nicholas Christakis: The reason we form networks is because the benefits of a connected life outweigh the costs. It's to our advantage as individuals and a species to assemble ourselves in this fashion. There are very fundamental reasons we live our lives in social networks and if we really understood the role they're playing in our society we would take better care of social networks and find ways to take advantage of their power to improve our society.
Wired: What rules do social networks follow and do they all follow the same rules?
Christakis: The first rule is that we shape our network. Humans deliberately make and remake their social networks all the time. We form new friends according to certain rules, and we ditch old friends, and we choose our spouses and co-workers and so forth. This means that we literally create the network around us.
The second rule is that the network shapes us and where you are located in the network has significant implications for the experience you have in life. The simplest example I can give you is that you can take two different adolescent girls. Both of them have two friends. ... It turns out that if a girl's friends do not get along, she's more likely to think of killing herself than the girl whose friends do get along. ...
Another rule is that our friends affect us. So you could have two happy friends or two unhappy friends, and it matters. We are affected by what's going on around us.
The fourth rule is that the effect doesn't just start there. Our friends' friends' friends affect us – meaning that there's a kind of social domino effect or a social contagion. Things ripple through the network and we can come to be affected, not just by what the people around us are doing, but by what people further away, that we don't even know, are doing. The best example of this is a children's game of telephone. You're the fifth in line, and the person whispers something in your ear that is erroneous. But it doesn't just include the errors that that person introduced. It includes all the accumulated errors of everyone else. So that's how we come to be affected by people downstream.
The last rule is that the network has a life of its own. The network is a kind of super organism and in a way has its own existence and its own desires and properties, if you will. A very simple example of this is the kind of accuracy that's achieved by Wikipedia, [in which] every little person contributing something creates a whole that's more than the sum of its parts.
Wired: How deeply do the online social networks run or are they just the McDonald's version of networks, and how much do they influence us?
Christakis: We are influenced by our friends, but we are not influenced by our acquaintances, at least not to the same extent. People have just assumed that . . . if we call our Facebook acquaintances our friends, we must be influenced by them, too. But we're not. We and others have done a bunch of work to show that if your real friends online say or do something, it affects you. But if your acquaintances online say or do something, it does not. People on average have about 106 Facebook friends, but only 5 or 6 real friends.
Wired: You've been monitoring 1,700 Facebook profiles belonging to students at an unnamed East Coast college to see how their values, preferences and habits spread through the network. What have you found?
Christakis: We looked at the influence with respect to movies, music, and books in the network. We found there was no evidence for influence for those three kinds of products when you looked at all Facebook friends. If one of your 500 Facebook friends made a particular movie his or her favorite movie, it didn't affect the probability that you would. But if one of your 5 or 6 close friends -- people, for example, who you co-appeared with in a photograph and tagged it on your Facebook page -- liked a movie, you subsequently were more likely to like that movie, too. There's a very stark contrast between the effect of any one of your friends doing something vs one of your real friends doing something. But this is just one data set; we don't want to make a mountain out of it.
Wired: You liken social networks to capital that helps you do things you normally wouldn’t be able to do. How does this work?
Christakis: New properties emerge as a result of the connection that don't exist within the individuals. The example I give . . . is the graphite and diamond. . . . The properties of graphite are completely different than the properties of diamond, and those properties do not reside in the carbon. They arise as a result of the patterns of interconnections between the carbon atoms. Therefore a group of carbon atoms can have different properties that have nothing to do with the carbon per se and have everything to do with the ties between the carbon. And that's what we're seeing about social networks. The same people assembled in different ways can give rise to different properties.
My favorite example of this . . . was work done by Brian Uzzi on Broadway musical production companies and transitivity. I know Tom and Dick, and when Tom and Dick in turn know each other, you have a transitive relationship and you get a closed triangle. Brian Uzzi found that if you took the group of people that were producing the musical, how successful the musical was deeply related to how transitive the production company was. So if nobody had ever worked with each other before, the transitivity was very low, the show was a flop. If everybody had worked together before, the show was a flop. . . . However if there was intermediate transitivity and some of the people knew each other and some did not, the show was a runaway success. . . .
There's an optimal amount of connection to maximize success. . . . What you want is the right balance between some people who have worked together before and trust each other plus some new blood. . . . The point is that how I take a group and interconnect it affects the productivity of the group. Now I should say it's not always that way. Sometimes more transitivity is just better. If you're trying to assemble a terrorist organization, you want to minimize transitivity – less and less is always better, because if the cops catch one of the bad guys, if he doesn't know anyone else in the organization, he can't bring it down.
Wired: You've studied the spread of obesity in social networks, examining 12,000 people over 32 years and found that obesity spreads from person to person, like a virus. That seems obvious. But you also found that it wasn't just your closest connections that influenced you.
Christakis: People were most likely to become obese when a friend became obese. That increased a person's chances of becoming obese by 57 percent. There was no effect when a neighbor gained or lost weight, however, and family members had less influence than friends. It did not even matter if the friend was hundreds of miles away, the influence remained. But the greatest influence of all was between close, mutual friends. There, if one became obese, the other had a 171 percent increased chance of becoming obese, too.
What we were able to do was to show not just that people were influenced by others, because this is common sense. But we were able to show that people were influenced by others for a class of phenomena that many people think of as deeply individualistic, like what your body size is. And we were able to show that the effects . . . don't just spread from person to person but also from person to person to person. There is a kind of social contagion, social domino effect.
Suzy makes Betty eat poorly. And then Betty makes Jane eat poorly. And Jane makes Ann eat poorly. Suzy does not know Jane or Ann, but Suzy's behavior and actions are influencing the interaction between Jane and Ann.
People often asks us, How can you be saying that obesity is contagious? The supermodels are as thin as they ever were. Part of our argument about the way obesity is contagious is that norms regarding acceptable body size are changing, and as people around you gain weight it resets your ideas about what an acceptable body size is and it creates a more permissible environment for you to gain weight. The real influence on people's lives is not these abstract ideals – supermodels – but rather what the people around them are doing. . . . Both weight gain and weight loss spread equally. If your friends lose weight it affects you, and if your friends gain weight it affects you.
Photo: Rick Friedman
