If you’re happy and you know it, maybe it’s catching

“Happiness is not merely the province of isolated individuals.” This is the conclusion of a new analysis of the results of a long-term cross-generational American community study: how happy you are depends on the happiness of the people around you.

Published in the British Medical Journal the study by researchers from Harvard and the University of California found that happiness is infectious; happy people tend to be found in social clusters.

From examining the social network inside the New England town of Framingham (population 60,000), the researchers were able to estimate that a person is 15% more likely to be happy if he or she has a social connection with someone else who is happy.

This relationship holds good, albeit less intensely, even if the connection one to the other is indirect. They two might never have met. For example, one might be the friend or colleague of the sister of the other.

Authors James Fowler, a political scientist from University of California, and Nicholas Christakis, a medical sociologist from Harvard, found a rich dataset in the shape of The Framingham Heart Study [1], a longitudinal study of community health initiated in 1948 and by now crossing three generations.

The heart study routinely collected information on depression, which they were able to recycle as an indicator of happiness. And, crucially, participants volunteered information about families, friends and colleagues, whom Fowler and Christakis were able to track down as a basis for mapping social networks.

They called their 5,000 strong core sample the “egos,” each connected to other individuals via family, friendship, neighborhood and work. People with relationships with egos Fowler and Christakis termed “alters” (many of whom in a town the size of Framingham were also “egos”). The Framingham Heart Study social network, included 12,000 individuals who were “connected” at some point between 1971 and 2003.

To make sense of this vast quantity of data the researchers applied social network analysis [2], a technique based on two simple building blocks: nodes and ties. Once nodes and ties are identified, it is possible to draw a diagram of the network and to work out the position of each person inside it. The shortest path from one person to another is known as the “degree of separation”.

The same marking and linking procedure also makes it possible to estimate an individual’s centrality within a network. It turned out that those closest to the centre of a network were more likely to be happy. The greater the number of direct ties – direct or indirect – the greater the likelihood of happiness.

How close relationships were - in terms of physical distance - also mattered. If a happy “alter” lived less than a mile away, it increased by one quarter the likelihood that an “ego” would be happy. Changes in happiness of friends who lived farther away than that did not affect an ego’s happiness so significantly.

Given the increasingly fractured condition of social relationships and especially the popularity of mobile phones and social networking, the importance of physical proximity intrigued Fowler and Christakis.

Based on this analysis they hypothesize (as must have numberless propagandists and Armed Forces Entertainment managers before them) that “the spread of happiness can ripple through social networks, giving rise to clusters of happy and unhappy individuals”.

They go on to argue that happiness is nevertheless more often regarded by social scientists as a self-contained personal quality. Prevention and early intervention programs that aim to improve well-being nearly always treat it as such, targeting individuals and estimating success in terms of program impact.

The Framingham evidence suggests that helping an individual may unintentionally make a difference on a collective level, but that, better understood, the infectiousness of happiness might become part of a more coherent strategy, and in the process enhance efficacy and cost effectiveness.

The ripples of laughter do not spread out for ever. The study also found evidence that three degrees of separation was the limit, confirming findings elsewhere arising from obesity and smoking behavior research. So there is a basis for supposing that the three degrees rule may apply to the dissemination of other phenomena across social networks.

On the other hand prosperous, middle-class Framingham [3] may be a case apart, and generalizing from such an isolated study perilous.

• See also extended coverage of the Framingham study in the New York Times [4]

See: Fowler J H and Christakis NA (2009), “Dynamic spread of happiness in a large social network: longitudinal analysis of the Framingham Heart Study social network,” British Medical Journal, a2338, 337