Friday, October 12, 2012
BROOKE GLADSTONE: As reporters relentlessly cover the horserace, they’re often missing what may be the biggest part of the story, why campaigns do what they do. That often is missed behind a curtain of privately gathered, staggeringly granular data. Two of our producers started digging into this, and they’re here now, I guess, to gripe about the lapse in coverage. Chris Neary, welcome to this side of the mic.
CHRIS NEARY: Hi there.
BROOKE GLADSTONE: PJ Vogt, welcome to you too.
PJ VOGT: Hey.
BROOKE GLADSTONE: So, starting with PJ, what’s behind this curtain?
PJ VOGT: So reporters still slice voters into archetypes, you know, middle age white males, college students, seniors, soccer moms. But campaigns, more and more, just look at you. We know that for the past five to ten years they’ve been doing what they call micro targeting. That just means that they’re setting their sights on actual individual voters. Sasha Issenberg is the author of “Victory Lab,” a book about what goes on behind the curtain. He said to us that each of us has a specific entry in these data bases.
SASHA ISSENBERG: Which is your name, your address, gender, age, in some cases race, ethnicity. Then there is information from the census about your block or the neighborhood in which you live, which allows the campaign to, to make projections about your socio-economic status, your education level, household type.
PJ VOGT: Then campaigns add more data from commercial firms. So, now they know that you subscribe to The New Yorker, that you have a weird obsession with sports, that you went on a vacation with your wife and child.
CHRIS NEARY: Uh, I’d like t point out to listeners [LAUGHS] that PJ has stopped describing –
- a hypothetical person and is now [LAUGHS] describing just me!
PJ VOGT: I know, I’m sorry but, honestly, for a campaign you aren’t a hypothetical.
SASHA ISSENBERG: They’re running algorithms that are basically looking for patterns between that big mass of data they have about each of you with the information that their polling tells them about specific attitudes about the election that’s underway.
PJ VOGT: Here’s a case in point. Back in 2002, Mitt Romney was running for governor of Massachusetts. His team was buried in data points, and they were searching for clues about those voters who didn’t plan to vote for Romney but could be convinced.
SASHA ISSENBERG: One of the attributes that unified them was a likelihood of being a premium cable subscriber, basically having HBO.
PJ VOGT: The campaign didn’t care why premium cable subscribers were more persuadable; they just wanted to find a correlation. But journalists care, so they struggle to find a reason. You know, swing voters seem to have responded to Romney’s tough on crime policies because it reminded them of “The Wire.”
BROOKE GLADSTONE: That’s a stretch.
PJ VOGT: Yeah but in 2004, when Bush won reelection, there were a lot of data stories that were nearly that crazy. Here’s an example from Eisenberg:
SASHA ISSENBERG: This idea that Karl Rove figured out that Republicans drink bourbon and Coors, and then he went out and found them, and then he got them all to vote a lot more, almost every part of it is, is false or exaggerated in ways that totally misrepresent it.
PJ VOGT: Sure, the Republicans looked at consumer buying habits but that’s just a small part of what’s going on behind the curtain. Campaigns are targeting groups of voters that we, as reporters, don’t even think of as a group. And if we don’t understand the campaign strategy, it’s really hard to evaluate if it’s working or not.
CHRIS NEARY: Okay, so PJ’s story is about the magnitude of data. My story is about how it’s actually used in the ground game, you know, phone banks and canvasing and, most boring of all, direct mail. Reporters don’t report on direct mail unless it’s horribly, juicily offensive. But in Michigan, in 2006, a group of researchers randomly selected potential voters from a study group and then sent them one of four different letters. Here’s one of those researchers, Columbia’s Don Green, explaining how the letters worked:
DON GREEN: Each, more strongly than the next, enforced the norm of civic duty.
CHRIS NEARY: The first letter basically reminded you that yeah, as a society we’ve decided voting is good, and please remember that. Then the second letter told you that researchers were actually tracking whether you voted.
DON GREEN: Telling people that whether they complied with this norm was going to be monitored as a matter of public record. The point of the study wasn’t to make people feel bad. It was basically to test this very basic social psychological proposition about the importance of norms.
CHRIS NEARY: Okay, now something from letter three, which was addressed to an entire household and gave the voting history of each member. Here Brooke, read it.
BROOKE GLADSTONE: Okay.
CHRIS NEARY: It does sort of make you feel bad.
BROOKE GLADSTONE: “This chart shows your name from a list of registered voters showing past votes, as well as an empty box which we will fill in to show [LAUGHS] whether you vote in the August 8th primary election. We intend to mail you an updated chart when we have that information.” Yikes.
CHRIS NEARY: [LAUGHS] All right, and finally, the fourth letter.
BROOKE GLADSTONE: Okay. “We’re sending this mailing to you and your [LAUGHS] neighbors to publicize who does or who does not vote. The chart shows the names of some of your neighbors, showing which have voted in the past. After the August 8th election, we intend to mail an updated chart. You and your neighbors will all know who voted and who did not.” Chris, that sounds like blackmail!
CHRIS NEARY: Uh, well each of the letters boost the turnout, but guess which one worked the best?
BROOKE GLADSTONE: The last one, right?
CHRIS NEARY: Oh yeah, three times more effective than any other piece of mail ever tested.
SASHA ISSENBERG: That’s certainly enough to flip a close election, so behavioral psychology is increasingly used as a weapon in campaign wars. And, by the way Brooke, you’ll never get that last letter. Campaigns hate to send out anything that prompts virulent hate mail in return, and one of those researchers got some of that mail. But they are running – other tests. Let’s bring in Hal Malchow here. He says tiny adjustments can dramatically change behavior, like in political fundraising letters, where once a shop accidentally misprinted the stamps on their envelopes and found out that that actually mattered.
HAL MALCHOW: They tilted the stamp slightly to the right and got a better response rate on the mailing. I once did a test where every single word of the letter was exactly the same, except that on the carrier envelope and at the top of the letterhead I put a little campaign graphic and it said, Fowler for Senate, something like that – 30% difference.
CHRIS NEARY: Malchow was one of the first political consultants to harness data for direct mail. Those crazy desperate-sounding emails you get from Obama or Romney asking for money, Malchow says if they’re doing it, they’ve tested it and it works.
BROOKE GLADSTONE: Okay guys, that actually is fascinating, but what’s your gripe?
PJ VOGT: Campaign reporters need to focus on this stuff. This is the real battlefield.
CHRIS NEARY: And look, we know the reporters don’t have access to those data bases or the results of the tests the campaigns are running.
PJ VOGT: But Issenberg says that’s no reason to give up reporting on campaign data.
SASHA ISSENBERG: I, I think of it as a sort of useful analogy to military coverage. People who cover the Pentagon realize they don’t understand how drones work, they don’t understand how drone targeting is done. And yet, military correspondents have figured out how to write that ambiguity and uncertainly and unknowability as part of the story, and not confuse our readers into thinking that we know everything about what they’re doing with drones.
PJ VOGT: So look [LAUGHS], we don’t want reporters to break into the data centers. We just want them to be honest about what they don’t know.
CHRIS NEARY: And to adjust their narratives to the age of algorithms. I mean, Brooke, is that so much to ask?
BROOKE GLADSTONE: Yeah, probably. [LAUGHS] Thanks, Chris.
CHRIS NEARY: Thanks, Brooke.
BROOKE GLADSTONE: Thanks, PJ.
PJ VOGT: Thanks.