NEWS

Facebook gives social scientists unprecedented access to its user data

Projects from around the world will delve into questions such as how misinformation spreads on social-media platforms and who distributes it.

Search for this author in:

Supporters of Presidential Candidate Sebastian Piñera wave flags in Santiago, Chile

Supporters of presidential candidate Sebastian Piñera wave flags at a gathering during Chile's 2017 presidential election.Credit: Sebastián Vivallo Oñate/Agencia Makro/Getty

Facebook is giving social scientists unprecedented access to its data so that they can investigate how social-media platforms can influence elections and alter democracies.

The first group of projects selected for funding involves more than 60 researchers split into 12 teams. They will tackle questions such as how fake news spreads, who distributes it and how to identify it. Their projects, announced on 28 April, will focus on countries including Germany, Chile, Italy and the United States.

The scientists will have access to reams of Facebook data, such as the URLs that users have shared and demographic information including gender and approximate age. The company — which has been accused of privacy violations in the past — is developing new protections aimed at shielding the identities of its users.

The research teams were chosen by the non-profit group Social Science Research Council in Brooklyn, New York, and Social Science One, an academic–industry partnership with ties to Harvard University in Cambridge, Massachusetts. A coalition of eight charitable organizations will fund the work. Facebook had no say in selecting the projects.

The programme could set a precedent for how social scientists work with companies to access information about social media, a growing force in shaping public discourse, says Simon Hegelich, a political data scientist at the Technical University of Munich in Germany. “It’s still a problem that social media is so important — especially for social science — but we don’t have access to the data,” he says. “Sometimes you spend a lot of time trying to find things out that were already known in companies.”

Worldwide webs

Hegelich leads a team that will study the spread of false information during Germany’s 2017 general election, using data from Facebook and Twitter. The researchers already have a list of Twitter accounts implicated in propagating misinformation during the 2016 presidential election in the United States. They found that some of those accounts were also active in Germany, and linked them to Facebook accounts with similar user names and content.

The team will have access to Facebook’s data on URLs that were shared more than 100 times on the site to determine the scope of the false-information campaign in Germany. Hegelich and his colleagues will try to determine how many people shared links promoted by these accounts — as well as the users’ genders and approximate ages. The researchers hope to use their findings to identify other such misinformation efforts, if they exist.

The opportunity to combine data across social-media services is particularly welcome, says Hegelich. Social scientists tend to analyse information from one platform at a time, he says, but that doesn’t reflect how data flows in the real world. “You have information taken out of Twitter and then shared again on Facebook,” Hegelich says.

The depth of access the project teams will have to Facebook data goes beyond many previous studies, in which researchers were restricted to more-limited data sets, says Michael Veale, who studies technology policy at the Alan Turing Institute in London. And that raises questions about the conclusions in those studies, he says. “It’s really good that Facebook — and hopefully some other companies — are saying that we are aware that we need to give researchers access to proper data sets within secure environments.”

Zeroing in

Facebook won’t have a say in approving or blocking the publication of the projects’ work — so even results that might cast the company in a negative light will be shared, says Sebastián Valenzuela, who studies communication at the Pontifical Catholic University of Chile in Santiago.

He is co-leading a project on the spread of fake news on Facebook during Chile’s 2017 general election. The team hopes to quantify which demographic groups of Chileans were most likely to be exposed to fake news.

The project will be the first of its kind to concentrate on Chile, says Valenzuela. “The evidence we have is mostly focused on the United States,” he says. “I’m not so sure it’s going to be applicable to countries in Latin America.”

Veale is happy to see projects from around the world — but he is disappointed that none of the studies focuses on India, the country with the largest Facebook user base.

Geographic specificity might be particularly important as policymakers grapple with how to handle disinformation campaigns, Valenzuela adds. “If we don’t understand the fake-news problem, the solutions we’re developing might not be the right ones,” he says.

doi: 10.1038/d41586-019-01447-5
Nature Briefing

Sign up for the daily Nature Briefing email newsletter

Stay up to date with what matters in science and why, handpicked from Nature and other publications worldwide.

Sign Up