A new article published by Mother Jones this morning, found via the Drudge Report, says Facebook is conducting “secret” voting tests on its users. According to the article, every six years the social network pushes out a tool know as the “Voter Megaphone,” which is a “high-profile button” that allows voters to say “I’m a voter” or “I’m voting.” Facebook says it is designed to encourage users to vote. But the report also says Facebook has been fine-tuning the tool, and that it also attempts to influence voter behavior.
According to Mother Jones, Facebook says it has finished fine-tuning the tool, and hopes on Tuesday “many of its more than 150 million American users will feel a gentle but effective nudge to vote.” The article says that research shows that a few million people nationwide will vote for no other reason than they were encouraged by their Facebook friends to do so through the “Voter Megaphone.”
So what’s wrong with that? Voter participation is a good thing, right? If encouraging people to vote was all it did, yes, but what if it is designed to influence who a user votes for, or if it targets friends of users who are more likely to vote for one party over the other?
Quoting the Mother Jones article: “In the 2012 election, Facebook proclaimed that it would again be promoting voter participation with the ‘I’m Voting’ button. On Election Day, it posted a note declaring, ‘Facebook is focused on ensuring that those who are eligible to vote know where they can cast their ballots and, if they wish, share the fact that they voted with their friends.’
But this wasn’t entirely true, because, once again, Facebook was conducting research on its users — and it wasn’t telling them about it. Most but not all adult Facebook users in the United States had some version of the voter megaphone placed on their pages. But for some, this button appeared only late in the afternoon. Some users reported clicking on the button but never saw anything about their friends voting in their own feed. Facebook says more than 9 million people clicked on the button on Election Day 2012. But until now, the company had not disclosed what experiments it was conducting that day.”
Facebook’s vice president for global business communications, Michael Buckley, said, “Our voter button tests in 2012 were primarily designed to see if different messages on the button itself impacted the likelihood of people interacting with it, For example, one treatment said, ‘I’m a Voter.’ Another treatment said, ‘I’m Voting.'”
The company also tested whether the location of the button on the page had any effect in motivating a user to declare that he or she were voting. “Some of [the different versions] were more likely to result in clicks,” Buckley said, “but there was no difference…in terms of the potential ability to get people to the polls. Bottom line, it didn’t matter what the button said.”
But is it possible that by learning which messages are more effective at encouraging people to vote, or by placing the tool in locations on the page their research shows is more likely to influence someone to vote, that Facebook could then more effectively target users who are more likely to vote a certain way?
Again quoting the Mother Jones article: “Since 2008, researchers at the company have experimented with providing users an easy way to share with their friends the fact that they were voting, and Facebook scientists have studied how making that information social — by placing it in their peers’ feeds — could boost turnout. In 2010, Facebook put different forms of an ‘I’m Voting’ button on the pages of about 60 million of its American users. Company researchers were testing the versions to understand the effect of each and to determine how to optimize the tool’s impact. Two groups of 600,000 users were left out to serve as a control group — one which saw the ‘I’m Voting’ button but didn’t get any information about their friends’ behavior, and one which saw nothing related to voting at all. Two years later, a team of academics and Facebook data scientists published their findings in Nature magazine.”
“According to Buckley, there were three reasons many Facebook users didn’t see the voter megaphone in 2012. The first was a variety of software bugs that caused a large number of people to be excluded from seeing the tool. This might have kept the button off of millions of users’ pages. In addition, some users missed the button because they were part of a control group, and others may not have logged in at the right time of day. Buckley insists that the distribution of the voter megaphone in 2012 was entirely random — meaning Facebook did not push the voting promotion tool to a certain sort of user.”
“We’ve always implemented these tests in a neutral manner,” Buckley insists. “And we’ve been learning from our experience and are 100 percent committed to even greater transparency whenever we encourage civic participation in the future.”
But as the Mother Jones article points out, Facebook has not always been known for her transparency. Micah L. Sifry, the author of the article, writes, “For the past two years, my colleagues at TechPresident.com and I have have made a steady stream of requests for details on Facebook’s 2012 voter megaphone research. We were met with silence or vague promises that someday, when the research was published in an academic journal, we’d get some information about what the company was doing. That suddenly changed this week when I started asking Facebook about a different but related experiment it conducted during the 2012 election.”
“In the fall of 2012, according to two public talks given by Facebook data scientist Lada Adamic, a colleague at the company, Solomon Messing, experimented on the news feeds of 1.9 million random users. According to Adamic, Messing ‘tweaked’ the feeds of those users so that ‘instead of seeing your regular news feed, if any of your friends had shared a news story, [Messing] would boost that news story so that it was up top [on your page] and you were much more likely to see it.’ Normally, most users will see something more personal at the top of the page, like a wedding announcement or baby pictures.”
“Messing’s ‘tweak’ had an effect, most strongly among occasional Facebook users. After the election, he surveyed that group and found a statistically significant increase in how much attention users said they paid to government. And, as the below chart used by Adamic in a lecture last year suggests, turnout among that group rose from a self-reported 64 percent to more than 67 percent. This means Messing’s unseen intervention boosted voter turnout by 3 percent. That’s a major uptick (though based only on user self-reporting).”
The full article, complete with analysis and and the results of the research conducted by Facebook can be found here.