With a tanking economy and global violence on the rise, there's at least one thing to smile about: A pair of scientists is reporting that happiness can spread through social networks, meaning that friends of the cheery contract the happiness bug themselves.
The data behind the new findings come from, of all things, a massive study of cardiovascular disease. In 1948, researchers began collecting health and other information on U.S. adults as part of the Framingham Heart Study. Today, the project has data on more than 14,000 people, and it has helped researchers identify many of the major risk factors behind heart disease and stroke. Because the Framingham leaders, trying to track volunteers over many years, worried about losing contact with them, they asked all subjects to provide the name of a friend who would know how to find them if necessary. Often, those friends were also part of the study.
Nicholas Christakis, formerly a hospice physician at Harvard, and James Fowler, a political scientist at the University of California, San Diego, used these data to create a social network of nearly 5000 people. The duo then matched the information with various health data. Last year, they reported that weight loss and weight gain could "spread" through the network, meaning that a guy whose friends were overweight was more likely to pack on the pounds himself (ScienceNOW, 25 July 2007). Christakis and Fowler published a similar finding on smoking earlier this year.
Now, the two have turned to something more ethereal: mood. To measure this, they relied on a questionnaire that was part of a depression assessment in the Framingham study. The volunteers were asked to rate how strongly they agreed with four statements that described their mood in the previous week, including "I enjoyed life" or "I felt hopeful about the future." The 60% who gave the highest score on every question were deemed happy. The rest were considered unhappy, the two report online this week in the British Medical Journal. The happiness assessments came from three examinations that occurred between 1983 and 2001.
Christakis and Fowler found that when one person was directly connected to another--for example, a close friend or immediate family member--who was happy, the original individual was 15.3% more likely to be happy themselves. If the person was once removed--the friend of a next-door neighbor, perhaps--the increase was 9.5%. And an individual two times removed from the original person boosted their chance of happiness by 5.6%. Unhappiness could spread as well, though the chance of becoming unhappy if one's friend became unhappy was slightly lower than for happiness. "It's person-to-person-to-person transmission," says Christakis.
The effect only held for friends in very close proximity, however, a mile or less apart. The researchers also found that mutual friends had a stronger effect than when only one considered the other a buddy. As in their obesity work, Christakis and Fowler also found that same-sex friendships and neighbor relationships were more potent than those of the opposite sex.
The biggest difficulty in social-network studies like these is showing that effect is actually spreading. It's equally possible that happy people simply like spending time with other happy people, or that they share an environment of safe neighborhoods and good jobs that make them happy. Responding to the former argument, Christakis and Fowler say they found that happy and unhappy people still become friends, but that--over time--the mood of one evolves to match the other. As for environmental effects explaining the clustering, the researchers found that although next-door neighbors had an influence, those living farther down the block--where the environment was presumably no different--did not.
Some experts remain dubious. In an accompanying paper, for example, two economists argue that applying the same methods to qualities such as height and acne in teenagers indicates that these traits too are contagious, even though no one believes that growth spurts are caused by friendships with tall people. Christakis and Fowler also didn't control for certain things that might affect happiness in a community, such as crime, weather, or unemployment, says one of the economists, Jason Fletcher of Yale University.
"Their study would really not permit differentiation" among the different explanations for why happy people, or unhappy people, cluster together, Richard Rothenberg, an infectious disease specialist at Emory University in Atlanta who has studied how HIV spreads through social networks, notes in an e-mail message. That said, the work is enticing and "brings a smile to your lips, because it's a funny thing, that happiness should spread," says Rothenberg, but the proof is not quite there.
Christakis and Fowler are continuing to plumb their data, looking also at loneliness, depression, and alcoholism.