A new study shows how much social capital matters
- Read original article here
For a look behind the scenes of our data journalism, sign up to Off the Charts, our weekly newsletter
P 2000 “Bowling Alone” transfixed Americanists in a way that few books ever do. The landmark study by Robert Putnam, a political scientist at Harvard, offered a sweeping survey of what ailed American society by detailing a steady erosion in the amount of social interaction among people who are not family, close friends or work colleagues—a concept called “social capital”. The withdrawal of Americans from bowling leagues, religious life and volunteering was a symptom of anomie, Mr Putnam argued. It would adversely affect not just the lives of individuals, but also American democracy itself. Social scientists have argued about the supposed importance of social capital in the intervening decades. A comprehensive set of studies by a team of economists, just published in Nature, may help settle things. It strikes a blow in favour of those who have championed the importance of social capital, in and of itself.
Fittingly for the new age of social science, the data come from Facebook—which allows for the kind of sophisticated network analysis that previous generations of scholars could have only dreamed of. The large research team—which included Raj Chetty of Harvard University, Matthew Jackson of Stanford University, and Theresa Kuchler and Johannes Stroebel of New York University—worked with the social-media giant’s proprietary data on 72.2m young American adults and the 21bn friendships they had formed on the site. The shape of every person’s network of friends could be converted into a few simple, empirical scores of social capital. Data on the hometowns, workplaces and educational histories of Facebook users allowed the researchers to estimate how social capital varied between places in a much finer-grained way than previously possible. And, most important, they were able to correlate these neighbourhood-level measures of social capital with important outcomes like graduating successfully from high school and having upward social mobility.
One measure seemed especially important. What the researchers called “economic connectedness”—the extent of friendships across social classes—was strongly associated with increased rates of high-school completion, reduced rates of teenage pregnancy and increased income for those born poorer. Moving from a place where friendships across social strata are relatively uncommon (a one-in-four chance of a friendship between someone in the bottom half of the class distribution and someone in the top) to one where it was relatively standard (a one-in-two chance) translated into an 8.2% increase in future earnings, the researchers found.
This corroborates findings in other recent large-scale examinations: places in America that are more segregated by race or income seem to give poor kids less of a chance. This can be difficult for policymakers to counteract. The most important institutions for encouraging these friendships seem to be colleges and universities. Only there are poorer students surrounded predominantly by their wealthier peers. Past research has also found long-term benefits for poor kids who attend integrated schools or who do not live in neighbourhoods where poverty is extremely concentrated. Divine intervention helps, too. The researchers also found that intra-class friendships were most easily struck up in religious settings.
Other classical definitions of social capital like those Mr Putnam employed—such as the extent of civic engagement and the cohesiveness of bonds within social classes—did not appear to be correlated with improved life outcomes. That may prompt a refinement of the theory of social capital. Verifying it has always been difficult. And even though this study is studded with quantitative measures derived from big data, its findings remain largely based on associations and correlations. It is hard to think of a way to experimentally vary a person’s degree of social capital—even leaving aside the ethical problems with doing it. Whatever uncertainty remains, the imperative for integration seems to have become stronger still. ■
