A single message sent to 60 million people on Facebook Inc.’s social networking site by University of California researchers may have encouraged 340,000 more votes in the 2010 election, a study suggests.
The work, co-authored by two Facebook data scientists, tracked the effects of the nonpartisan message from the point when it was first seen by individuals, through their personal network and into the voting booth. When the first recipients passed on the message, it included their pictures, a note that they had voted and a link allowing the next-in-line to find a nearby polling place.
The research, reported today in the journal Nature, is the first in which scientists directly intervened in the lives of online subjects to measure social effects, said James Fowler, at the University of California, San Diego. It found that 80 percent of those who voted were influenced to do so by someone they knew online. The result may be important in the current presidential election, the researchers said.
“If we’d only measured the direct effect on the recipients, we’d have missed the whole story,” said Fowler, a study author and political scientist, in a conference call. “The network is key. The online world and the real world affect one another. This message that started online, that spread online, actually affected real-world behavior.”
While it’s not clear from the data how many of the 340,000 voters might have done so whether they received the message or not, the statistical evidence, backed by voting records, strongly suggests the influence of the message grew and had a concrete effect as it was handed along, Fowler said.
Only 37 percent of eligible voters went to the polls in the 2010 congressional election, according to the U.S. Census Bureau. Today’s finding suggests that political strategists should aim their outreach more at networks than individuals in the future, according to Fowler, who said he isn’t affiliated with either of the current presidential campaigns.
The researchers’ message was displayed at the top of the “News Feed” to 60 million Facebook users, encouraging users to vote. It provided a link to their local polling place and a clickable button that said “I Voted.” This post also showed as many as six small photos of other so-called Facebook “friends” who had also clicked “I Voted.”
The study was broken down into three groups. A second group, involving 600,000 people, received a similar notice that didn’t include pictures of their friends, and a third group of about the same size didn’t receive any post.
The researchers then checked the publicly available voter rolls to see what effect their postings had on individual turnout. The announcement featuring friends’ photos were linked to 60,000 votes, and the follow-through messages were tied to another 280,000, the study found. There was no statistical difference in turnout between those who got no message and those who got the purely informational message.
Close friends mattered most for the effect, the research suggested. The average Facebook user has 150 friends, and about 10 of those -- the ones the user interacts with the most -- accounted for the difference in voting, Fowler said. The effects were equal among self-described liberals and conservatives.
Cameron Marlow, the head of data science at Facebook, said that the posts were put up on Election Day in 2010.
“Each campaign cycle brings new technologies that enhance civic engagement and we’re excited that this research suggests that social influence and the power of friends may impact voter turnout,” Marlow said.
Facebook -- which didn’t contribute funding to the study beyond the time of their two employees, according to an e-mailed statement -- has grown to become the world’s largest social-networking service with more than 950 million users. Members of the service can share messages, photos, videos and other activities on the service.
Still, the company, which makes most of its money from advertising, has come under investor scrutiny since its initial public offering in May amid concerns about its growth prospects. Since its first trading day, the stock has failed to close above its IPO price of $38.
With the presidential election in full swing, the company is benefiting from political advertising from both Democrats and Republicans. At the same time, the company has been trying to more closely tie advertising with interaction from friends on the service. It’s rolled out advertising features called “Sponsored Stories” in which buyers pay for content in the News Feed after members or friends indicate they like a company.
Marcus Messner, an assistant journalism professor who teaches on social media at Virginia Commonwealth University in Richmond, said the study shows why social media is so important in the presidential campaign.
“It’s a heavy emphasis for both campaigns, and the fight for the Senate and House of Republicans in high-profile races,” Messner said in a telephone interview. “There’s hardly an hour where you don’t get a posting from Obama and Romney. It’s especially important in this campaign, which is so close, and because it comes to turn out and who can get people to vote.”
Obama has had an advantage, he said, because he has more younger supporters and has had a longer, established social media campaign, Messner said. Obama also engages social media audiences, by posting a “this seat’s taken” picture after Clint Eastwood talked to an empty chair, pretending to address Obama, during a Republican National Convention speech.
In a Sept. 4 survey by the Pew Research Center’s Internet and American Life Project, about a third of social networking site users say social networking sites are very or somewhat important to them in keeping up with political news.
Today’s study builds on other work that’s shown behavior, such as obesity, smoking and drinking, and emotional states spread through networks of friendships.
“The Holy Grail is how do you find evangelizers who pass it along,” said Todd Van Etten, managing director of the Washington office of Crowdverb, a Seattle-based social mobilization company for corporations and campaigns. “How do you sway people on the fence? You always want to be part of something larger than themselves.”
The study was supported by the James S. McDonnell Foundation and the University of Notre Dame and the John Templeton Foundation as part of the Science of Generosity Initiative.