The Victory Lab (29 page)

The twenty-four-year-old had experienced revelations like this before.
As a teenager living in Philadelphia’s western suburbs, Rogers learned just after Thanksgiving of his junior year of high school that he had been diagnosed with non-Hodgkin’s lymphoma. He scheduled his trips for chemotherapy around his team’s lacrosse schedule, and every two weeks would check in to the oncology ward of a local children’s hospital. Most of Rogers’s fellow patients were much younger kids, and he kept a distance from them, and from the Happy Meals and video games their presence drew. Rogers further rebelled against the frivolity by proudly refusing to watch television and choosing instead, for the first time in his life, to read for pleasure. The first book he picked was Henry David Thoreau’s
Walden
, which led him to other works of American transcendentalism. By the next Thanksgiving, Rogers had beaten cancer, completed his junior year on schedule, played in fourteen of his team’s sixteen games, and developed a new sense of the self. “Even when you are healthy again you don’t feel normal,” says Rogers, who as an adult kept a fit physique and a gleaming shaved head. “You carry some secret differences.”

Now Rogers was driving back to Williams, where he had played Division 3 lacrosse and been elected student body co-president. He met with one of his favorite professors, Al Goethals, whose Introduction to Social Psychology class Rogers realized had fertilized his now-blossoming interests. Goethals had introduced his students to behavioral psychology, the burgeoning subdiscipline that took a pessimistic view of the human brain as a flawed instrument for navigating life’s most challenging decisions. Rogers had voraciously consumed writings by some of the field’s most prominent scholars, including psychologist Robert Cialdini, an expert in the way that consumers were simply unable to make rational choices, and economist Richard Thaler, who explored how that flawed “mental accounting” warped markets involving everything from auctions to savings accounts. As he sat down again with Goethals years later, Rogers expressed the frustration he had with the polling profession, and its inability to bring rich insights to understanding the political brain. “I knew psychology was powerful, and wanted to learn more,” he says. Goethals assigned Rogers
further reading, and the two started to regularly meet in what amounted to a personal tutorial in the behavioral sciences. Eventually they came to the same conclusion: that Rogers should apply to a graduate school that would allow him to study the psychology of decision making.

When Rogers told Abacus’s partners, Janet Grenzke and Mark Watts, that he was planning to leave politics for the academy, neither was surprised. They had seen the curiosity on his face when he asked questions like “how do we know this?” in response to some piece of received wisdom, and his disappointment when he heard their replies. “The answer was,” Watts says, “ ‘In our field, in general, we learn as we go along. We have a lot of beliefs that aren’t well-tested because no one can afford to test them.’ ” Both Grenzke and Watts had earned Ph.D.s from political science programs, and yet they both discouraged Rogers from entering one. “Political science as a discipline just isn’t asking applied questions,” says Watts. “Psychology is getting at this in a better way than political science.”

In 2003, Rogers enrolled at Harvard in the social psychology program, but after a semester he became frustrated that professors kept trying to steer him away from studying questions with broad, real-world consequences. Rogers transferred to an integrated psychology program at Harvard Business School and quickly aligned himself with a professor, Max Bazerman, famous for his research on negotiations. Bazerman had helped to shape the field of behavioral science in the 1980s, and like many of his peers with backgrounds in psychology or economics he ended up at a business school, where insights on decision making had immediate, and potentially lucrative, applications.

Bazerman had just published an article in the
Journal of Behavioral Decision Making
titled “What We Want to Do Versus What We Think We Should Do.” This “want-should” research, as Bazerman labeled it, examined the conflict in how people struggled with multiple goals: there are things they think they should do (stay on a diet) and those they want to do (order the banana split). Along with five coauthors, Bazerman had designed an experiment in which they tweaked the “dictator” ultimatum
game used by game theorists to demonstrate that test subjects who said they would treat their opponent fairly didn’t always do so when faced with an immediate game situation where they could punish opposing players.

One day in class, to illustrate this conflict, Bazerman mentioned that he had noticed that some movies his wife had placed in their Netflix queue lingered there seemingly forever, leapfrogged by new additions she pushed to the head of the line. Foreign documentaries might be the things people thought they should see, Bazerman speculated, but romantic comedies and action flicks were always what they wanted to watch. Rogers and a classmate, Katharine Milkman, thought the movie example could give them the chance to yank the study of want-should conflicts out of artificial lab settings and into the real world. When Milkman persuaded an Australian online video rental company, Quickflix, to share its customer data, they had discovered the type of naturally occurring experiment upon which economists love to stumble. Milkman was fascinated by consumer decision making, but Rogers liked the subject because he thought its lessons would be applicable for understanding how people approach policy questions. He already had global warming in mind: people know they should reduce energy use, but don’t want to drive any less. Could policymakers build support for controversial proposals by committing to them well in advance, while people are still thinking about the trade-offs in “should” terms rather than “want”?

Rogers was surprised that applying the sensibility and methods of behavioral psychology to politics remained a largely virgin endeavor, as political scientists looked past the nonconscious mind and psychologists explored other dimensions of human experience. Rogers started to compile an open-ended list of topics he wanted to research, and he found them plentiful and self-evident—many either topics that psychologists had examined in laboratory settings, but never in the field, or that had been demonstrated in situations other than politics. While most of his peers at Harvard saw behavioral psychology as a means to commercial ends, Rogers was looking for ways to make people better citizens, and impressed
colleagues as an unlikely humanist. “There are people in the experimental world who think that people are silly and we understand the truth,” says Mike Norton, a professor who taught and collaborated with Rogers. “He’s more interested in how do people respond, how do people work, rather than ‘let’s change their actions.’ ”

Rogers struck those around him as combining—in a way unusual among grad students—a restless mind that spotted a lot of potential research avenues with a practical sense for how to quickly head down most of them. Norton saw Rogers’s charisma as a motivating tool, noting his “ability to get people excited so they’ll work.” But Rogers also ran up against the limits of graduate school, especially when his experimental agenda fell outside the standard purview of business school research.

Rogers had become particularly enamored of the work of Cialdini, who began studying sales techniques in the 1980s after he concluded that he was “an easy mark for the pitches of peddlers, fund-raisers, and operators of one sort or another.” Cialdini documented how consumers followed bad cues or were drawn to faulty assumptions, and the ways marketers could exploit them. Eventually he turned his powers toward promoting good behavior with cynical mind games. It was Cialdini, for instance, who documented the success of hotels that encouraged guests to reuse their towels by informing them how many guests also did so, rather than by highlighting how disappointingly low recycling rates were or the general importance of environmental concerns.

Rogers proposed to Bazerman that he follow a similar tack to promote voting, another bit of good civic behavior that suffered from low rates of participation. Massachusetts didn’t have any elections scheduled in 2003, so Rogers decided to see if he could instead increase voter registration by talking up its popularity. He dispatched several students to Boston bus stations with stacks of registration forms and instructions to approach riders with alternating messages: one emphasizing how many people signed up to vote, the other emphasizing how few did. But the experiment stalled before Rogers could get the necessary scale to draw meaningful conclusions;
his researchers struggled to find enough who weren’t already registered to measure the relative effects of the different messages. Rogers tried to save face by referring to the failed experiment as merely a “pilot study,” and leapt quickly when a presidential campaign offered to let him test out his theories on its voters.

MOST WEEKENDS IN THE FALL OF 2003
, Rogers left Harvard Square in his grandmother’s old gray Subaru station wagon and embarked on the hour-long drive from Cambridge to Derry, New Hampshire. As fall became winter, the foliage was drained from yellows and oranges into dour browns, and the roadside palette enhanced by the patriotic tricolor of signs announcing Democratic candidates for president. Rogers would arrive at a suburban office complex housing the presidential campaign of Wesley Clark, a retired four-star general who had entered the race late and was counting on a win in New Hampshire to launch him nationally. Rogers was drawn to Clark as a credible, progressive voice on national security issues, and offered himself to the campaign as a volunteer. Clark’s gamble was that he could squeeze enough space for himself between the two flawed candidates fighting for first place: Howard Dean, who had emerged as the candidate of the party’s romantics, and John Kerry, who by default had become the choice of its realists.

After a year studying behavioral psychology, Rogers was no longer capable of being a romantic about politics, or for that matter any aspect of human activity. When he would arrive in Manchester, Rogers would either be pointed to a phone or handed a clipboard and local street map. Whether he was making calls or canvassing at doors, Rogers would get a script for interacting with a voter, and he always wondered what Cialdini would have made of it. The scripts would say things like
You can make the difference
or
We need your support
. These phrases had become so deeply enmeshed in the aural tapestry of elections that it was unlikely that anyone actually listened
to the words themselves anymore. But Rogers did, and he heard a begging whimper on behalf of a lonely cause. This was the same tone the hotels had used when they beseeched guests not to be part of the problem, before Cialdini came along and showed them experiments that demonstrated people would rather fall in line with something already popular.

The previous spring, Rogers had written a paper for one of Bazerman’s classes that applied what Cialdini called “social norms” to voter turnout.
Cialdini had found repeatedly that what he described as injunctive norms (“you should not litter”) were far less effective at changing behavior than descriptive norms (“few people litter”). Perhaps, as political scientists had always assumed, voting was a different type of activity from reusing one’s hotel towels, in which case the calls to civic duty might resonate with their audience. But what if the motivation to vote came from the same place that determined human decision making in other spheres where people had a clear choice, like what to buy or how to spend their time? When Rogers decided to answer the question with the only tool he trusted, he despaired over the difficulty of executing field experiments in politics. “I was just a grad student,” Rogers says. “I didn’t think I could scare up thirty thousand dollars for an experiment.”

So when he started volunteering for Clark in the fall of 2003 and realized that a friend ran the campaign’s Derry office, Rogers saw a chance to piggyback an experiment onto the tens of thousands of voter contacts that Clark’s volunteers would be doing in the last days before the late-January primary. Rogers had seen only one of the papers by Alan Gerber and Don Green, but he was familiar with the rich tradition of field experiments in psychology and he saw those as models for a large-scale randomized test of different turnout messages. He was introduced to a campaign official, who eventually contacted Rogers with good news: the campaign would let him conduct his experiment in the run-up to the New Hampshire primary.

Rogers plotted an experiment to measure if a Cialdiniesque
voting is popular
spiel would prove more successful at motivating voters to turn out than the usual
we need your support
appeals. He drafted a pair of scripts
that reflected different approaches, and developed a plan to randomize the call sheets and walk lists assigned to volunteers, leaving a control group that would get no contact at all. On his weekend trips to New Hampshire, Rogers started making fewer calls on Clark’s behalf and instead worked to engender enthusiasm for his experiment among the field staffers who oversaw the canvasses and whose diligence would be necessary to ensure the test’s purity.