15:01 PM

Ohio State researchers among first to study Facebook data

Focus will be on “problematic” sharing on social media site

Researchers at The Ohio State University will be among the first to have access to privacy-protected Facebook data to study social media’s impact on democracy in the United States.

The Ohio State-led project was among 12 inaugural recipients of the Social Media and Democracy Research grants from the Social Science Research Council and its partner, Social Science One.

The research team, which also includes scholars from three other U.S. universities, will have unprecedented access to anonymous data from Facebook on the sharing of online content. The team will use these data to examine a variety of behaviors that may pose harmful influences on what people learn about science, politics and their community.

“We know that rumors and misinformation spread on social media platforms like Facebook can undermine citizens’ ability to make thoughtful decisions that affect their lives,” said Kelly Garrett, associate professor of communication at Ohio State and principal investigator of the project.

“But we can’t stem these problematic behaviors if we don’t understand them. That’s what this study is all about.”

Kelly GarrettGarrett will lead the study with colleagues from Ohio State, Stony Brook University, the University of Michigan and Cornell University.

More than 60 researchers from 30 academic institutions around the country were chosen to receive grants through a competitive peer review process organized by SSRC. Facebook, SSRC and Social Science One have built a first-of-its-kind privacy-protecting research infrastructure for the scientists to use for their work.

While the sharing of news that fact-checkers have labelled as false will be one focus of the study, the researchers will also be examining other sharing behaviors that may pose issues.

For example, some people share headlines to stories they haven’t read, and some content is only shared among a highly partisan network of Facebook users.

“That doesn’t mean that information shared by strong partisans is always necessarily harmful. But it is a behavior that is prone to being abused and can have harmful consequences,” Garrett said.

“So it is important to understand when it happens, how often it happens and whether we can explain its emergence.”

One question the researchers will examine is whether people share problematic posts more often at specific times of day, or times of the year, and whether they are more likely to come from specific parts of the country.

The study will explore how changes that Facebook makes in its platform – such as how messages with questionable content are flagged – affect sharing behavior.

In addition, the researchers will examine how real-world events, like terrorist attacks or candidate debates, influence what people are sharing.

Garrett said he hopes the results of the study will lead to ways to reduce the posting of fake news and other kinds of problematic sharing.

“We believe our findings could help system designers create early warning systems that predict when people are likely to share problematic posts,” he said.

“That could help Facebook design changes to help keep problematic sharing in check.”

Garrett’s colleagues on the study are Robert Bond, assistant professor of communication and political science at Ohio State; Jason Jones, assistant professor of sociology at Stony Brook University; Ceren Budak, assistant professor of information at the University of Michigan; and Drew Margolin, assistant professor of communication at Cornell University.