Political scientists and organizers have known for years that most effective way to get someone to vote was through personal contact: a knock on the door from a canvasser, or even better, a nudge from a friend. Now a new study by researchers led by U.C. San Diego that is being published tomorrow in the journal Nature offers detailed evidence that a non-partisan get-out-the-vote reminder on Facebook can also increase voter turnout--especially if they come with evidence that your real friends are also voting. As first reported in the U-T San Diego News, "about 340,000 more people turned out to the polls two years ago because of a single Facebook message posted" on Election Day.

Political science professor James Fowler, the lead author, says, "This really I think is the first study to show that online social networks can effect these real-world behaviors at a scale that's potentially important."

Not only were Facebook users influenced by the get-out-the-vote message, the study argues that their friends, and the friends of their friends, were also influenced to vote.

The study buttresses a recent report by the Pew Internet & American Life Project that found that people whose friends post some or a lot of political content on social networks are more likely to become involved politically or to say they've changed their mind about a political issue.

Working with Facebook's cooperation, the researchers had access to 61 million Facebook users who were at least 18 years old by Election Day 2010. (Hence the study's startling title: "A 61-million-person experiment in social influence and political mobilization.")

They were randomly divided into three groups: a) about 60 million (the "social message" group) who were shown a GOTV message at the top of their News Feed, which included a link to local polling places, an "I Voted" button, a counter tallying how many other users had reported voting, and up to six profile pictures of their own Facebook friends who had reported voting; b) about 600,000 (the "informational message" group) who were shown all of those things except the faces of their friends; and c) about 600,000 in a control group who were not shown any GOTV message at all. For about 10% of the overall group, the researchers were able to validate claimed voting behavior from publicly available voter records.

Users who received the social message were 2.08% more likely to say they voted than those who received the informational message without the pictures of their friends (that is, 20.04% vs 17.96%). They were about .4% more likely to have actually voted. More than 12 million people overall clicked on the "I Voted" button in 2010, compared to about 5.4 million in 2008, by the way.

Cameron Marlow, the head of data science at Facebook, said, ""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."

The authors note that "seeing faces of friends significantly contributed to the overall effect of the message on real-world voting. In fact, turnout among those who received the informational message was identical to turnout among those in the control group…which raises doubts about the effectiveness of information-only appeals to vote in this context." People who saw pictures of friends that they were deemed to be closer to, based on the amount they interacted with each other on Facebook, were more likely to be influenced positively. Party identification had no effect either way.

In all, the study says that "friends generated an additional 886,000 expressed votes" and "close friends generated a further 559,000 votes," but based on actual validation of people's voting, the verified impact of the Facebook message--the 340,000 additional votes noted above--was entirely due to the influence of close friends. "There is no evidence that ordinary friends had any effect on either [validated voting or polling place searches]. In other words, close friendships accounted for all of the significant contagion of these behaviours, in spite of the fact that they make up only 7% of all friendships on Facebook."

The authors note that while these "contagion effects"are small, they only sent one message on Election Day, and the pool of people receiving it might have already voted by absentee ballot, or not logged into Facebook that day, or logged in too late to influence others. They also note that they probably undercounted the number of actual votes influenced by the message because they couldn't match people's Facebook names to voter records due to typos, nicknames and the like.

But the bottom line of this report will probably make both sides of the "weak ties" vs "strong ties" argument happy. A nudge online can cause a statistically significant number of people to be more likely to vote. But at the same time, they are most likely to be influenced in doing so by the behavior of the people who are actually their stronger friends. A big and shallow social graph isn't as influential as a small but intense one. Or, when it comes to using online tools to spur voting, real friends beat "friends" hands down.