With nearly 1 billion users, Facebook has clearly become a feature of many people’s lives worldwide. A new study suggests that the social network has the potential to get hundreds of thousands of people to engage in a single behavior – namely, voting.
Researchers report in the journal Nature that one Facebook message may have gotten 340,000 additional people to the polls for the 2010 United States Congressional elections.
The team, led by James Fowler, professor at the University of California, San Diego, designed the experiment with the cooperation of Facebook. Cameron Marlow of the data science division of Facebook collaborated on the study, too.
“This really, I think, is the first study to show that online social networks can affect these real-world behaviors at a scale that’s potentially important,” Fowler said at a news briefing Tuesday.
Fowler and Harvard’s Nicholas Christakis are prominent in social network research; their book “Connected” gathers copious research on how a myriad of behaviors spread from person to person. For instance, you may be happy as a result of your friend’s happiness and your friend’s friend’s happiness. Bad habits such as smoking can also spread in this way.
“The network is key,” Fowler said. “If we want to make the world a better place on a massive scale, we should focus not just on changing a person’s behavior, but also on utilizing the network to influence that person’s friends,” he said.
This study’s scope was huge: more than 60 million people received a statement on the top of their News Feed that encouraged them to vote and offered a link for finding polling locations. The item also displayed a clickable “I Voted” button, a counter showing how many other Facebook users also said they voted, and the profile pictures of up to six randomly-selected Facebook friends who had also clicked “I Voted.”
There were two other groups: about 600,000 people saw all of the above with out the pictures of Facebook friends, and an additional 600,000-some didn’t receive the message in their News Feeds at all. All participants were randomly assigned.
People who got messages were more likely to vote than those who did not, the researchers found – in fact, the percentage difference between the groups in the experiment suggests that an additional 60,000 people were motivated to vote as a result of the message.
Seeing the photos of Facebook friends as part of the message appears to be critical in getting people to vote, the researchers found. Users who got the message without photos were no more likely to vote than people who didn’t get any message. But those who got the message with pictures had higher voting rates.
“The messages not only influenced the users who received them, but the user’s friends, and their friends of friends as well,” Fowler said.
Moreover, the friends of the people who saw the messages were also more likely to vote than friends of participants who didn’t see the messages, Fowler said. Multiplying this effect per friend by the number of friends by the number of people who saw the message, researchers determined that 280,000 additional votes were cast. All of these resulted from ties in the social network, not the message itself, Fowler said.
Close friends – with whom Facebook users likely have a face-to-face relationship – were extremely influential in this contagion of voting, the researchers found.
How do we know that these 340,000 people wouldn’t have voted anyway?
Fowler compares this to a drug trial, in which some people get randomly assigned a medical treatment or a sugar pill. If the people in the treatment group tend to show more positive effects than the control group, that means the drug may cause a positive effect.
Researchers didn’t just rely on the self-reported information on Facebook about who voted, they also used publicly available voting records. It appears that 4% of people who said they voted on Facebook did not actually vote.
The effect of the message appears to be nonpartisan: As many Democrats as Republicans seem to have gone to the polls as a result of the message, Fowler said, although many people do not state their affiliations on their Facebook profiles.
There are still unanswered questions about the implications of this research, however. It’s not clear whether age is a factor in who voted. Researchers have yet to explore other characteristics of people who are most influenced by Facebook messages.
All in all, between the direct and indirect effects found in the study, the total percentage increase per person in voting behavior as a result of a Facebook News Feed message is about 2.2%. That’s on the lower end of the spectrum compared to what other “get out the vote” studies have found with regards to other means of communication.
But the real effect could be even greater than what the study observed, Fowler said. A lot of people who saw the message had probably already voted, through early voting and absentee ballots – in fact, one-third of voters in the 2010 election cast their votes before Election Day.
Many people who saw the message were also unable to vote because they logged into Facebook too late in the day to go to their local polling station - as high as 20% may have been in this situation, Fowler said. Also, younger people are less likely to vote, and are less susceptible to such appeals, he said.
“I think if you added all of these things up, we’d find that this was actually one of the stronger ‘get out the vote’ messages that we’ve seen in the literature,” he said.