Voting in the Facebook Age

-- By AnnaScott - 22 Oct 2012

Compared to other established democracies, voter turnout in the US is very low. A vast amount of political campaign resources are focused not on relaying their candidate’s message to undecided voters, but on making sure their already-identified supporters actually show up to vote on Election Day. Political junkies know this as the “ground game” or GOTV (Get Out the Vote) operation, and in contests where the margin is less than 2-3% it can be decisive in determining the winner. The backbone of voter turnout operations has looked the same for decades, relying primarily on phone-banking and door-to-door canvassing of identified supporters (although targeting has gotten increasingly aggressive and micro-level). Even the much-discussed, extraordinarily massive GOTV operation of the 2008 Obama campaign kept mostly within this traditional mold. Recently, however, the internet has opened up new possibilities in voter turnout strategies – possibilities which have legal and practical limits far beyond what the average citizen would expect.

Nature recently published a study suggesting that one Facebook announcement generated 340,000 additional votes nationwide during the 2010 midterm election. The message reminded users to vote and showed pictures of friends who had clicked on an “I Voted” button. One control group saw an Election Day reminder with no photos of their friends; another group received no message. Researchers then matched users with voting records to compare turnout among the three groups; the first group had a statistically significant increase in their voter turnout rate. So far, most discussions of this study have focused on its implications for our understanding of social pressure and the growing power of social media on our behavior. (For example, much has been made about how showing photos of close friends had a greater influence than showing non-close friends.) I think that more attention should be paid to the question of how far Facebook’s involvement in voting could go -- both legally and practically -- given the fact that voting records are, to varying degrees, publicly available.

Because of this public availability of voting records, theoretically Facebook (or an organization who is able to partner with them) might not need to rely on voluntarily provided information about whether you have voted. In one experiment, political scientists mailed voters a copy of their voting history along with voting histories of some of their neighbors, and warned that they would be repeating the mailing after the upcoming election. Turnout went up 20 percent among voters who received the “public shaming” mailing—an enormous jump. (Studies show that email GOTV efforts lead to turnout increases of only 1%, while door-to-door canvassing sees jumps of 8%.) As discussed above, a slight advantage in voter turnout can sway an entire election; thus political groups have a very strong incentive to push the legal limits of proven turnout strategies. Voting-related nonprofits have used or attempted to use this “public shaming” tactic here and there, most recently in the Wisconsin recall election, but never on a very large scale. Could a kind of “public shaming” tactic ever operate via Facebook?

While it is unlikely Facebook could ever publicly post your voting history to your profile, I can imagine something more like a blend of the public shaming tactic with a real-time, Election Day tracker linked to users. What if Facebook could show you photos of friends who hadn’t been recorded as checking in at the polls yet (or whose absentee ballots hadn’t been received yet), ostensibly so you could remind them to vote before closing time but also implying that your photo might be showing up to your friends until you voted?

Sounds a bit far-fetched, until you remember that voting records are publicly available and increasingly digitized, and that in 2008 the Obama campaign attempted an aggressive, near-real-time updating of supporter lists on Election Day, called the Houdini project. Campaign volunteers stationed at polling places matched voters with precinct rosters that had each voter assigned to a numeric code, which was entered into a campaign database that then “cleaned up” contact lists for later in the day, removing supporters who had already voted. While imperfect and fraught with day-of technological issues, the aggressiveness of the project hints at what’s to come as voter records are increasingly digitized and realtime voter updates become easier to obtain.

Surely you are aware of the distinction between updating of the public record and the real-time collection of information by campaigns, and you're just hiding this distinction from the reader for some undisclosed reason? You don't actually think that local polling places are updating some real-time list of who has voted in today's election, right? You understand that compiling and acting on this data is what GOTV activity is now about, right? And that it costs much money and requires very complex organizational discipline?

One argument against the idea that Facebook and their potential partners would ever wade further into voter turnout tactics is that it’s not that easy to match someone’s Facebook profile to their voter profile. (In the Nature study, the researchers were only able to conclusively match about 10% of the users to a voter profile.) However, recent partnerships between Facebook and state governments suggest that in the future this will not be as difficult. Washington allows citizens to register to vote through an app on Facebook. (It prefills their registration form with their name and birthdate; the user then inputs the number of their state-issued ID.) Washington’s Secretary of State has assured the public that Facebook does not have access to the state database, but that its page merely “overlays” the app, but this example illustrates how voter records and Facebook profiles could become increasingly connected over the next few election cycles.

Another counterargument might be that legally, Facebook as a private company could never obtain voter records from state governments; all states limit access to voter records to some extent, usually to political candidates/parties, elected officials, voter-education nonprofits, and PACs. But of course, entities can get around that by using apps and working only indirectly with Facebook.

The rise of SuperPACs? addresses two other practical obstacles for any kind of large-scale, electronic “public shaming” tactic: cost and blame. They have lots of money to experiment with voter turnout strategies, and can step in where candidates might not want to be associated with a tactic that might anger some supporters who see it as an invasion of privacy.

Targeted social pressure is a potent voter turnout tool – how far campaigns can and will go to harness it remains to be seen.

Why the Facebook advertisement? The firm-specific form of argument here is just unnecessary complexity. An actual understanding of what might happen and why can be derived from consideration of the social processes involved, without regard to the particular behavior of one thug-driven database product. If you spent less time on Facebook here and more time on thinking about elections, the results would be very much more valuable. That seems to me to be much the most promising line of revision.