So a team of American researchers turned to Facebook, the world’s largest social database with nearly 3 billion users, to study the problem with unprecedented scale and precision. They analyzed the privacy-preserving data of 72 million U.S. Facebook users between the ages of 25 and 44. Friendships on Facebook are used to represent real-world friendships.
The researchers used an algorithm to rank users by factors such as socioeconomic status, age and region. They then measured the degree of interaction between richer and poorer people and created the term “economic relatedness” to represent the share of a person’s friends above or below the average socioeconomic level. They then compared this measure with previous studies of children’s ability to lift themselves out of poverty in every zip code in the United States.
The results are strikingly similar, said Harvard economist Raj Chetty, lead author of the papers on both studies.
He pointed to the first paper showing that “economic relevance” is one of the strongest predictors of economic mobility that anyone has found to date.
The second paper sought to find out why children from wealthier or poorer backgrounds were in some ways more likely to make friends than others.
The researchers found two main factors: One was the degree to which the two groups were in contact with each other — such as whether they attended different high schools or lived in different neighborhoods.
Yet even if wealthy and non-wealthy students do attend the same schools, they may still not socialize with each other — a factor the researchers call friendship bias.
The study found that half of the social disengagement between the rich and the poor is due to a lack of opportunities to engage with each other.
“But the remaining half is explained by dating bias,” Chetty said.
Chetty went on to add that the findings suggest that U.S. policies aimed at reducing economic segregation between schools and districts are important, but not sufficient.
The study found that where richer and poorer children met had a big impact on whether they became friends – meaning institutions played a big role. Chetty points out that friendships in religious institutions such as churches are more likely to cross class boundaries.
Researchers released data on contact and dating bias on socialcapital.org on Monday, which they hope will prompt action by authorities across the United States.
Similar results could be found in other countries, Chetty predicted, urging researchers and governments around the world to access their own Facebook data.
Noam Angrist of the University of Oxford and Bruce Sacerdote of Dartmouth College in New Hampshire said the research represents “an important contribution that will lead to a greater understanding of social capital.”