Biting your nails over a very tight presidential race three weeks before Election Day?
Turnout is where precision targeting really works, says Sasha Issenberg, author of “The Victory Lab: The Secret Science of Winning Campaigns.” Political scientists are using the tools of behavioral economics and behavioral psychology to understand why people vote and how to enforce that habit; campaigns are building voting scores for people in their databases like the financial industry builds credit scores.
Does it work? More than 1.1 million people already have voted, according to early voting statistics collated by George Mason University political scientist Michael McDonald, and plenty of them were prodded to do so without ever knowing it.
We talked to Issenberg about the modern get-out-the-vote operation and who’s ahead in that game, Democrats or Republicans.
Q: What do we know about why people vote, and how do we know it?
A: The traditional messaging from the League of Women voters was to talk about how low turnout was, so don’t be part of the problem. And everything we knew from behavioral psychology was that to make an activity sound sad and lonely and unpopular was not a particularly good way to encourage other people to do what you want them to do.
Q: So along came randomized controlled experiments for winning elections?
A: Behavioral economics realized early on that there are things that you can control, and that gives you the ability to measure.
The best experiment was done in Michigan in 2006 before the GOP primary — these two guys from Yale and one of their post-docs and this crazy direct-mail guy from East Lansing who sent out a series of letters before the election, and one of them said something like, “Dear Ann, your history as a voter is a publicly available document through the Board of Elections; here is your history as a voter. You did not vote in the mayoral election, you did not vote in the school board, and here are your neighbors’ history and whether they voted. . . . There is another election coming up and we’ll follow up with an update after the election.”
And this increased turnoutby eight percentage points. The guy who sent it got death threats. They called it social pressure. It’s been demonstrated in every other aspect of human behavior. We are afraid of being outed as bad citizens. We are afraid of being shamed in front of our peers. We will adjust our actions to uphold whatever standards we uphold for our communities.
Q: Is there a difference between Democrats and Republicans regarding voting, since there is also this emerging body of research into genetics and liberalism and conservatism?
A: We know a lot about personality differences, but I don’t think we are seeing what they call heterogeneous treatment effects among partisan lines, on the real behavior of voting. . . . However, we now in the 21st century have come to the acceptance that every other aspect of your personality or character can be shaped by genetics or heredity. Except we thought that somehow politics lived apart from that, and we were purely rational when it came to supporting the death penalty or [whether] taxes should be more evenly distributed.
What is radical [in all the new science of campaigning] is that all these things that we have learned in other disciplines that animate human behavior might work in politics, too. And we test them and it does.
Q: So a campaign can take all the microtargeting data it has collected and build these models to predict which voters will respond to which of these motivational cues?
A: Yes, from a thousand categories, and create a predicted individual likelihood of performance. The campaign will have a turnout score, for how often you have voted; a support score, for the percentage likelihood of whom you would support; a pro-choice score, etc., and those all are framed as probabilities out of 100.
Q: For each voter?
A: For 170 million people, yes, for the Obama campaign. And they have them for non-voters to determine whom they want to register. Now they can make a list of the people you want to register because they look like your kind of voters, and now they don’t have to stand outside a supermarket and wait for whomever comes along.
Q: Does the Romney campaign have this?
A: Yes, they structure it differently, but basically, everybody does this.
Q: What is your assessment of how these two campaigns are doing, side by side, in how they can exploit all this information they accumulated on people?
A: The Obama campaign is significantly ahead. I think we tend to underappreciate how radically different the reelection campaign is. We tend to look at it as, the advantage is you can send out the secretary of the interior to go out and give grants, and I don’t want to preclude that, but you also have four years to plan your budget, hire your staff, lay out a research agenda, and have all these local elections or uncontested primaries to muck around in.
Karl Rove in 2001 and 2002 could do testing for the 72-hour project [to get out the vote in the 2004 reelection campaign for President George W. Bush], and build in the structure. The Obama campaign built on what it learned in 2008.
It used to be that every one of those interviews that volunteers did went on the clipboard, and the clipboard went onto the file cabinet, and the file cabinet went into the dumpster and after Nov. 10 into the landfill. Now, it’s going on a Republican or Democratic party server forever.
Democratic candidates for years to come will benefit from the huge volume of contacts that the Obama campaign was able to make — whether Obama wins or loses in November.