Getting Out the Vote: Sasha Issenberg Talks About the Science of Campaigns

October 10, 2012, 1:34 pm

By JOHN WILLIAMS

In his book “The Victory Lab: The Secret Science of Winning Campaigns,” Sasha Issenberg writes about the data-driven revolution in the methods of targeting potential voters and getting them to turn out on election day. In a recent e-mail interview, Mr. Issenberg discussed what motivates people to vote, whether Republicans have a structural advantage over Democrats and more. Below are edited excerpts from the conversation:

You write that scholars who study politics have been most frustrated by “the inability of their discipline to even justify the individual decision to vote at all.” What's the conventional wisdom right now about why people vote in the first place?

We often pose this question the wrong way: why are there so many people who don't vote? For nearly 15 years, academics have been running field experiments in campaigns to test what can increase someone's likelihood of voting. The things that work in these experiments often have nothing to do with the actual candidates or the issues at stake in the election. The successful motivators tend to be social. People who vote aspire to fit in with the people around them, or want to be seen as the type of person who does what's right.

In 1984, 25 percent of voters had split their votes on a ballot between Democrats and Republicans. By 2000, that number was 7 percent. Is this decline still happening? And does this trend make targeting voters less useful or even more important?

It's pretty evident from any poll now how hardened the partisan divide has become. Most of my narrative takes place in the shadow of 2000, which revealed the extent of that polarization. Since then, political strategists have b een interested in tools that can help them identify people who are already on their side but are not habitual voters - and discover what exactly can nudge them to the polls. Politics since 2000 has focused far less on changing people's minds than we assumed, and far more about modifying their behaviors.

What's the National Committee for an Effective Congress, and what role does it play in this story?

NCEC was the first modern institution practicing what is now described as data-driven politics. It was founded in 1948 by Eleanor Roosevelt as a PAC to fund liberal congressional campaigns, but that mission was effectively gutted by post-Watergate reforms limiting direct aid to candidates. In the mid-1970s NCEC decided that instead of handing out cash it would generate resources that all Democrats could use to improve their targeting. The goal was to create a profile for every precinct in the country, which turned out to be an amazingly arduous task before personal computers. But the result - a series of scores assessing each precinct's partisan behavior and volatility - was a marvel to campaign operatives who had been reliant on ward bosses and precinct captains for such intelligence about their turf.

As recently as 2000, you say, the Republican National Committee “had always been a more stable, centralized entity than the centrifugal [Democratic National Committee].” Is that still true?

Structurally I think that's still the case, and has remained so even under a Democratic president, because Obama has never shown himself interested in party-building. The institutions that have flowered in the Obama era - like the Analyst Institute, an odd amalgam of consulting firm, think tank and secret society that coordinates research projects across the left - exist outside the party's control but cultivate its expertise. Obama, for instance, has contracted with the Analyst Institute this year to admi nister persuasion experiments from within his campaign.

You write that after Obama was elected in 2008, “an industry that had been chronically unreflective about its failures was obsessed with learning from Obama's success.” Has Romney's campaign learned from that success?

They have certainly tried, although the Obama campaign has been quite successful at keeping private most operational details about what it did in 2008. For all that was written about the campaign's digital expertise, until I started reporting for this book barely anything about the campaign's astonishing use of voter data and analytics had come out. This remains the corner of campaign activity where the two sides have the least information about how the other side operates.

Do candidates ever feel hesitant about turning their campaigns into laboratories for political consultants, where experiments might backfire and cost them an election?

Yes, they almost always do. The whole id ea of a control group - designating a group of voters whom you don't contact for the purposes of an experiment - spooks campaign officials, as does the notion of investing in research that may not yield results until after election day. As a result, candidate campaigns have been slow to adopt testing techniques, even as they benefit from some of the insights that the experiments generate.

Is there a particular politician who's been more open than others about embracing these techniques?

I have a chapter about Rick Perry, and the decision of his top political consultant, Dave Carney, to bring four academic political scientists into his 2006 reelection with an invitation to test anything the campaign did. “The eggheads,” as Carney called them, plotted ways to measure everything from the scope of fundraising networks to the effectiveness of lawn signs and the value of seeking newspaper endorsements. For three weeks in the winter of 2006 all of Perry's television and radio ads - and several days of the governor's travel around Texas - were directed not according to his strategic interests, but completely at random, as part of an experimental design. It wasn't Carney's (or Perry's) goal to help produce scholarly journal articles, but they saw this as a way to keep political professionals honest: they were using academic tools to basically audit their consultants.

Studies have shown that potential voters respond far less to pleas about civic virtue than to messages that make them feel shame for not voting. How has that knowledge been put to practical use in campaigns?

In a few weeks voters from around the country will get calls from volunteers or visits from canvassers in which they are told that turnout in the coming election is expected to be very high. This is not really based on any particular knowledge of what turnout will be, but an insight from a sequence of behavioral psychology experiments that show people are mor e likely to cast a ballot when it's described as a popular activity rather than merely a dutiful one. If other people are doing something, experiments have shown, people are more likely to fall in with the crowd.