It’s too good an idea to resist: Happiness is contagious. A new study published online Dec. 4 in the British Medical Journal shows it.
Maybe you’d be more inclined to resist these ideas, though: Headaches are contagious. Acne is contagious. Height is contagious.
According to another study published the same day in the same journal, by Ethan Cohen-Cole of the Federal Reserve Bank of Boston and Jason Fletcher of YaleUniversity, these latter claims are nearly as likely as the first. The researchers “proved” them using the same methodology.
Sure, say Cohen-Cole and Fletcher, happy people tend to have happy friends. But that may not mean that happy friends make you happier, any more than tall friends make you taller, your friends’ headaches hurt your head, or the pimples on your friend’s face infect you. Instead, happy people might choose happy friends in the first place. Or an outside event, say a shooting in the neighborhood, could make a whole group of friends unhappy at once.
The emergence of social network theory has allowed scientists to begin studying much more deeply how our friends and colleagues affect our health, and it is becoming clear that social effects are big and important. But the field is so new that researchers are still disputing which methods are necessary to preclude detecting influences that aren’t really there.
Cohen-Cole and Fletcher say that the happiness researchers, James Fowler of the University of California, San Diego and Nicholas Christakis of HarvardUniversity, didn’t do enough to control for factors like similarity of friends or common environmental influences. As evidence, they cite their own study, in which the apparent social impacts of headaches, acne and height went away once they controlled for environmental factors.
For their part, Fowler and Christakis acknowledge in their paper that there are plenty of reasons why happy people might tend to have happy friends. But they defend their methods, saying the only explanation that explains their data is that happiness is contagious.
Fowler and Christakis made clever use of data that were likely never intended for use studying social networks: the Framingham Heart Study. A project of the National Heart, Lung and Blood Institute and BostonUniversity, the study collected health data on several thousand people from a small city in Massachusetts over decades.
The study also asked people how often in the last week they experienced feelings such as, “I enjoyed life” or “I felt that I was just as good as other people.” Participants gave the name and contact information for one friend, just to make it easier to track them down for the next survey a few years later. Christakis and Fowler constructed a network of people’s social interactions using this data together with information about where people lived and worked and who their spouses were.
The researchers wanted to understand not just how one person’s mood affects another person’s, but how an entire web of interactions contribute to or detract from happiness. A quick glance at a drawing of the network revealed what the team was looking for: Clumps of happy people. People at the centers of clumps were particularly likely to be happy. And it wasn’t just people’s friends that made them happy — it was also their friends’ friends’ friends.
But Fowler and Christakis realized those effects might only indicate that people like to hang out with folks who are about as happy as they are. So they also looked at changes in happiness levels, computing how likely it is that when your friends get happier, you do too.
A single friend’s increased happiness, they found, increased by 9 percent the chance that you would get happier, a statistically significant amount, Fowler and Christakis report. A happier spouse, nearby siblings or neighbor are even more likely put a smile on your face. Best of all, oddly enough, is a happier next-door neighbor. Happier coworkers, on the other hand, probably wouldn’t do much for you.
That still left the questions of whether common environmental effects could be causing the simultaneous changes in happiness levels. A new neighborhood park, for example, could make lots of people happy at the same time, even if happiness isn’t contagious.
To rule that out, the researchers looked at pairs of friends who both participated in the study, say, for example, Charles and John. They found that if Charles listed John as his friend but John didn’t name Charles, then Charles tended to be more influenced by John’s happiness than the other way around. That difference, they argue, couldn’t possibly be explained by environment, since Charles spends the same amount of time with John as John does with Charles.
But those checks aren’t enough to reassure everyone. Cohen-Cole and Fletcher say that it’s remarkably easy to see social influences where they don’t exist, as evidenced by their own study. “We took their same method and applied it to silly things that can’t possibly be contagious,” Cohen-Cole says, “and we found similar effects.” So, Cohen-Cole and Fletcher concluded, the methodology must be flawed.
Cohen-Cole and Fletcher used the data from the National Longitudinal Study of Adolescent Health, which followed several thousand adolescents over a decade. They found that if a person’s acne got worse, their friend had a 46 percent greater chance of worsened acne, and if a person’s headaches got worse, their friend had a 35 percent greater chance of worsened headaches. (Large as both those effects were, Cohen-Cole and Fletcher determined that they were not quite statistically significant.) And for every inch a person grew, their friend grew an extra 0.2 inch (which was statistically significant).
Then they applied a stronger method of controlling for environmental influences than Fowler and Christakis had used, and found that the apparent effect went away. They argue that these more conservative methods need to be used all the time.
Fowler and Christakis say that the idea that acne, headaches and height could be transmissible might not be so absurd. Friends might learn remedies from one another, or share dietary habits that would affect their acne or headaches. They also point out that all these properties were self-reported, and a short person with tall friends might be more inclined to exaggerate their height.
Despite these lingering questions, Ana Diez-Roux of the University of Michigan School of Health says that Fowler and Christakis’ work is groundbreaking. “I’m not ready to say that it’s totally convincing that the effects they’re seeing are pure contagion,” she says, partly because of the concerns that Cohen-Cole and Fletcher have raised. “But they’ve done lots of interesting things to try to support their case.” With time, she says, the community will develop better techniques to resolve the uncertainties.