Resume Reading — The Famous Anonymous


The Famous Anonymous

The problem with Western test subjects.

Why are you bringing white people in here?” a man bellowed at us in Swahili. He was referring to me. I was following Monica, a…By John Bohannon

Why are you bringing white people in here?” a man bellowed at us in Swahili. He was referring to me. I was following Monica, a Kenyan research assistant, as she led her team through the maze of dirt paths and metal-roofed huts that made up Kibera, a lawless swath of Nairobi that is Africa’s largest slum.

We arrived at a bar, in the room of a hut that was just big enough to fit a pool table and a counter behind a steel cage. Johnny, the bar’s owner, soon arrived, wearing a huge grin, a T-shirt emblazoned with the words “Lost Soul,” and a pair of immaculate white tennis shoes that seemed immune to the slum’s pervasive dull-red dust. He was well-off by Kibera standards, thanks to the proceeds from his bar, a nearby boxing gym, and this gig as a fixer for scientists doing research in the slum.

Johnny’s client on this occasion was the Busara Center, a research institute supported by the American non-profit organization Innovations for Poverty Action. “Busara” is Swahili for “wisdom.” Drawing on a pool of some 4,000 test subjects from Nairobi’s slums, Busara is striving to rectify what has come to be known as the “WEIRD” problem: the fact that much of the social science performed today uses research subjects that are almost exclusively Western-Educated people from Industrialized, Rich, Democratic countries.

The anonymous, random test-taker plays a vital role in behavioral science. By making decisions in one way and not another, these subjects sway entire research efforts, whose conclusions are then folded into everything from policy to product development. Yet the peccadilloes of WEIRD subjects, often presumed to be universal, may not apply in a different cultural context. By serving as stand-ins for humanity as a whole, they unwittingly enjoy a kind of fame.

An experiment in progress at the Busara Center’s lab in Nairobi, Kenya.

Johnny led us deeper into the ramshackle labyrinth, to the home of his friend where a table had been set up in the courtyard. Monica sat down and opened her laptop, and Johnny got on his cell phone and began dialing. Within 10 minutes, people began to arrive: women with infants tightly swaddled in blankets, snug as caterpillars; young men with knitted caps pulled low; older men dressed in sweaters, despite the heat. When the courtyard was nearing capacity, Monica launched into her speech. “We are recruiting people to participate in the studies at our center,” she said. “The purpose of the study is to understand how people think and make their economic decisions.”

The Truth About Auto Safety

A 1959 Sport Cars Illustrated review of Chevrolet’s new Corvair describes the car as “not without sin.” A throwaway comment on comfort soon became an appropriate description. Assailed with lawsuits for safety defects, the Corvair rolled to infamy in Ralph...READ MORE

As people signed up with Monica, they left a mobile number and a scan of their thumbprint on her laptop. After the recruitment drive, some of the subjects would be randomly chosen, at a point in the future, to visit the Busara Center’s headquarters, 2 kilometers away in one of Nairobi’s business districts. There, at a laboratory with 20 cubicles containing touch-screen computers, they would play economic games designed by scientists.

Johnny, the bar’s owner, soon arrived, wearing a huge grin, a T-shirt emblazoned with the words “Lost Soul.”

A few days earlier, I had tried my hand at one of the games that the Kibera residents would soon have the chance to play. It was a version of what is known as the “centipede game,” a classic method for studying trust and altruism. I was randomly paired with a partner, with whom I shared a common reserve of cash. The money in this pot doubled every few seconds, up to a stated limit of 256 shillings—around $3, a decent daily wage in Kibera. If both of us waited patiently for the cash to grow to its maximum and then shared it at the end of the round, everyone would win. But that assumed that we trusted each other. At any point, I could grab the cash myself, and the other player could do the same, which would bring the round to an end. Each pair of players shared 10 rounds before being randomly reassigned. For most of Busara’s games, real money was at stake. Counting a participation bonus of 200 shillings, the Kibera subjects could walk home with 400 shillings or more.

Challenges like this have a long history in behavioral economics. In 1992, a group of Caltech researchers published a seminal paper in which they analyzed the choices of 136 Caltech and Pasadena Community College undergraduates. The students played a version of the centipede game, in which pairs of students alternately got the chance to snatch the larger share of a growing pile of money. The researchers compared the subjects’ responses to computer simulations of the game. If the human players were aiming to make as much money as possible, the logical move seemed to be to selfishly take the largest pile. But interestingly, the human subjects did not seem to follow a single rational strategy. In general, they were far more generous early in the game than the strictly rational computer players. And about 5 percent of the humans behaved completely altruistically, always sharing the pot with other players even in the face of exploitation. No one made the selfish decision of taking the pot on the first move.

After comparing the results to simulations of the game on one of Caltech’s supercomputers, the team found an explanation. The humans seemed to be aware that a small fraction of the players would be altruistic. So their mix of seemingly illogical moves early in the game was actually rational in light of this uncertainty. Once you know what kind of person you’re dealing with, you alter your strategy accordingly.

The human subjects did not seem to follow a single rational strategy.

However, it soon became clear that this result was sensitive to cultural bias. In 1995, an anthropology Ph.D. student at UCLA named Joseph Henrich started running human subject experiments among tribes in the Amazon basin. He wanted to test the assumption that everyone’s mind is basically the same under the hood, from an evolutionary and game-theoretical perspective. The prevailing wisdom was that, if about 5 percent of Caltech undergrads played a game as altruists, then so should a group of people who lived in the Amazon jungle.

Heinrich’s results told a different story. People in the Amazon basin used very different strategies to those of their American counterparts. They made moves that Americans would never make, such as rejecting offers of free money. Henrich theorized that gifts always came with strings attached in Amazon societies, and so nothing was ever truly free. And when he repeated the experiment with other populations in the developing world, he found that the patterns of Western research subjects—replicated and assumed to be common to all humanity—simply did not hold up.

I was eager to meet the person who had won the most money in first set of centipede tests that Busara ran with Kibera residents: Alex, a shy, 19-year-old boy who lived at home with his single mother.

While Monica was recruiting subjects, I managed to reach Alex on the phone. I passed it to Johnny, who gave him directions to our location, in Swahili. Alex appeared in the courtyard, smiling politely in a clean button-up shirt and tan trousers. He was short by Kenyan standards, about 5 feet 6 inches. We moved inside to get away from the crowds. As we spoke, Alex listened and watched me intently, as if I were a magician about to pull something out of my hat. It was the look of someone who expects the unexpected, and who is ready for it.

We compared our experiences as research subjects. “It was frustrating,” Alex told me in excellent English. “I did not know when to trust people.” I explained how I had approached the task, making a generous move right at the start, sharing the pot at 8 or 16 shillings, and signaling to my unknown partner that I wanted to play nice. Then I would let the pot grow higher, sharing each time. Finally, on the last round, I would suddenly turn greedy, grabbing as large a pot as I could for myself. The trust that I had cultivated would be destroyed, of course. But the next player wouldn’t know my devious ways. It seemed rational, and it worked pretty well. I came in third from my cohort of 15.

It was the look of someone who expects the unexpected, and who is ready for it.

“My strategy was mixed,” Alex told me. He would wait to see how the other person behaved and then he acted accordingly. If they played nice, he would respond in turn; if they were selfish, he would return the favor.

This is close to a classic tactic in evolutionary game theory known as “Tit for Tat,” James Vancel told me. Vancel is an American social scientist with blond hair and pale blue eyes, who is fluent in Swahili and runs Busara’s lab. He is reluctant to make firm assessments at this early stage, but said that his “anecdotal conclusion,” based on preliminary data, is that Busara residents may be less trusting than WEIRD subjects. In a 2006 paper, in which business administration students at the University of Arizona played the centipede game, 35 seconds was the median “stopping time”—when one of the participants decided to abandon cooperation and grab the pot for themselves. The median for Busara subjects, by contrast, was a mere 2.8 seconds, suggesting that they were far more skeptical of the other players’ intentions.

Vancel is keen to emphasize, though, that only time and much more data will be able to clarify what role culture plays in economic decision-making. “One would think that the sense of solidarity and community would lead to respondents from Kibera being slightly more altruistic than similar Western subjects,” he said. “But it could also be the opposite, due to people’s limited trust for strangers and the relative stakes of the money.”

Whether or not Vancel’s anecdotal conclusion is later confirmed, Busara’s research will should make clearer the limits of a popular, convenient assumption: that the very particular, individual lives of research subjects don’t matter.

John Bohannon is a biologist, science journalist, and dancer based at Harvard University.

4 Comments - Join the Discussion