In 'Victory Lab,' A Concoction Crafted To Move Voters Political campaigning is increasingly driven by data. Journalist Sasha Issenberg says voter outreach has shifted from a precinct-centered game to one focused on individuals' behavior. In his new book, The Victory Lab, he says the smallest changes in tactic have had the biggest impact on politics.

In 'Victory Lab,' A Concoction Crafted To Move Voters

  • Download
  • <iframe src="" width="100%" height="290" frameborder="0" scrolling="no" title="NPR embedded audio player">
  • Transcript


There was a time when campaigns relied on old standbys to find voters. Rich neighborhoods would vote Republican; union members were Democrats; city pols could deliver their sections of the city. But that was then, and this is a very new now with the tiniest variations in voter behavior noticed and stored in politicians' computers. Now, campaigns collect information that goes way beyond demographics - data points as disparate as the soda you prefer the catalogs you peruse, the car you drive all make up a picture of a voter that campaigns can then use to find common ground with their candidates. It's a data-driven world now. And journalist Sasha Eisenberg describes it in his new book "The Victory Lab." He joins me to talk about it. Welcome.

SASHA EISENBERG: Thanks for having me.

WERTHEIMER: Well, let's talk about how campaigns got modern.

EISENBERG: So, there are two major innovations, both of which take place around 2000. One is these two political scientists from Yale, run a field experiment on the streets of New Haven in the fall of 1998 and it's basically a drug trial. Instead of giving people drugs and placebos, they're giving them pieces of direct mail and phone calls and door-knock reminders to vote. And for the first time, people are really able to measure cause and effect in these basic things the campaigns have been spending money on. And so there you find out that the phone reminder has very little impact, that the mail has a small but impressionable impact and that sending somebody to knock on a door has a significant, measurable impact on mobilizing them to vote. And all of the sudden campaigns realize that these things that they have been unable to sort of measure or test can be approached empirically. At the same time, people in politics realize that they could take the very limited amount of information that exists on a voter registration record, which is typically your...

WERTHEIMER: Name and address.

EISENBERG: and address, gender, date of birth, party registration, which elections you voted in, maybe a couple dozen data points at most, and can combine it with the huge databases that have been created by commercial firms to send catalogs or magazines to prospective subscribers or charities (unintelligible)...

WERTHEIMER: Or sell soap.

EISENBERG: ...donors or sell soap. And campaigns now run these sort of complex statistical algorithms that troll through these sometimes thousands of data points that they have for every voter in the country and come up with a prediction of how they think you're going to vote, if you're likely to cast a ballot at all, if you're going to be pro-choice, if you're going to be a gun owner. And those will have become the sort of staple of campaigns sorting through the electorate in a far more nuanced granular way.

WERTHEIMER: I mean, why do you want to do that? What do you do with that?

EISENBERG: Campaigns used to target by precinct. Democrats would go into Democratic precincts and would more or less treat everybody as somebody that they wanted to mobilize, even if a Democratic precinct was only 75 percent Democratic. They would sort of indiscriminately mobilize voters there through GOTV - get out the vote - activities. And a quarter of the votes they'd be mobilizing would be Republicans. By reducing this to an individual level, campaigns can interact with and mobilize one voter and not their neighbor. In a state like Virginia, Democrats can put together a coalition that's a little more far-flung than they were looking for because you can go out and harvest a few votes here and a few votes there, even if it's generally unfriendly turf.

WERTHEIMER: Is one party better at this than the other party? Is one, I mean, can you sort of pick out stars of the new technology?

EISENBERG: Yeah, I think generally the left is way ahead of the right now in terms of use of analytics and data and politics. A lot of it has to do with the fact that there's just be more of sort of migration of talent and ideas from parts of academia, particularly political science, some behavioral psychology. And so there are a lot of academics who are collaborating on these experiments with institutions or campaigns or the Democratic Party. I think that a lot of the research that's informed by thinking from psychology has moved our idea of what goes on in a campaign from just changing people's opinions to modifying their behavior.

WERTHEIMER: Modifying their behavior.

EISENBERG: Well, voting is, you know, there's sort of an underlying question of why do some people vote and other people not vote. And a lot of what we've learned through these experiments have been there's a sort of social dynamic around voting. You know, the most successful tool ever tested for use - turning a non-voter into a voter was this in this experiment done in Michigan in 2006 where a mail consultant sent voters a postcard with a copy of their own vote histories, which is the list of the recent elections in which you voted and not voted. And then it said this is a publically available document from your county Board of Elections, and here are your neighbors' vote histories too. And it had other people on the block and the elections in which they had voted and not voted. Said there's another election coming up. Afterwards, we'll send everybody an updated set. And this increased turnout among the people who received it in this randomized experiment by 20 percent.

WERTHEIMER: They didn't want to be seen by their neighbors to falling down and not doing their civic duty?

EISENBERG: Right. And it's an idea that in the behavioral sciences known as social pressure. It's been sort of tested in other areas but only recently do people try to apply it to voting. And I think it's sort of freed some people in politics from just thinking about this as about policies or ideology or leadership values, that the sort of ways in which you can nudge somebody to vote or register are things best understood in behavioral terms.

WERTHEIMER: What are the big truths of campaigns that you think you should no longer hear reporters talking about, being very positive that they know what is going on?

EISENBERG: I think big truths entirely should probably be dispensed with. I think that we're starting to understand that lots of little things have measurable impacts in campaigns. People's minds are largely made up whether they know it or not before an election starts. You know, the party you're registered with is highly predictive of how you're going to vote. There are far fewer moveable parts, I think when we get into the scrum of a campaign like this one, then the media coverage leads us to believe in far less ability for one great speech or amazing debate line to sort of change the architecture of a race. And so one of the reasons I focus on so many small marginal improvements, like changes to canvassing scripts as opposed to changes to talking points before somebody goes on stage at a convention is that we still - and by we, I mean the journalists and the people inside the campaigns who are selling these services - we still have a very little ability to understand how those big changes in message or narrative, or whatever we want to call them, are actually changing things. And it's hard to find big things that make a difference. And my guess is if we in the past found big things and thought they meant a difference, we were looking for a story to tell more than we were actually accounting for the things that were moving voters.

WERTHEIMER: Sasha Eisenberg's book is called "The Victory Lab." Thank you very much.

EISENBERG: Thanks for having me.


WERTHEIMER: You're listening to NPR News.

Copyright © 2012 NPR. All rights reserved. Visit our website terms of use and permissions pages at for further information.

NPR transcripts are created on a rush deadline by an NPR contractor. This text may not be in its final form and may be updated or revised in the future. Accuracy and availability may vary. The authoritative record of NPR’s programming is the audio record.