Laurie Wang

According to a recent co-authored Yale study, children may think just as much about the future as adults do.

A study designed by experts in cognitive psychology, economics and statistical methodology at Boston University, Harvard and Yale found that elementary school children, like adults, are more likely to cooperate with people they know they will have to interact with again.

The researchers created a virtual “Prisoner’s Dilemma” game in which children played against an anonymous opponent and tried to maximize their own winnings. The children played either one round or several rounds, known as “one-shot” or “repeated” games, respectively. Data revealed that schoolchildren are able to strategize the way that adults do when working with other people.

“Adults cooperate more when they know that they will interact again with the same partner — ‘I have to cooperate with you today, so that you will cooperate with me tomorrow,’” said David Rand, co-author of the study and professor of psychology, economics, cognitive science and management at Yale. “The idea of the present study was to see if younger children also engaged in the same kind of strategic reasoning.”

Past studies about the prisoner’s dilemma rarely involved children, but researchers were able to design a situation that 11-year-olds could understand, co-author and Boston University psychology professor Peter Blake said. In the study, students from five fifth- and sixth-grade classrooms used a computer interface to play out the prisoner’s dilemma. Players, who remained anonymous to their opponents, could either push or pull a virtual tray of coins. Players could deliver one coin to themselves by pulling the tray, or three to their partners by pushing it. Players who pulled the tray were classified as “defectors” while those who pushed the tray were called “cooperators.”

Blake said that designing an experiment that children would understand was difficult.

“The main challenge we faced was how to design an interface that was intuitive for children and still retained the logical structure of the game,” Blake said. “We put our heads together to try to link a game theoretic approach with development … The resulting interface was very much a creative collaboration among the co-authors.”

According to the results, children understood that if they cooperated with the same partner over several rounds of the game, their own payoffs would be greater. Researchers found that children even used conditional strategies, such as the “Win-Stay-Lose-Shift,” a strategy in which one player’s defection immediately led the other player to stop cooperating.

The study found that girls were generally more cooperative than boys, even when girls had successfully defected in a prior round. But this was not the only striking result, researchers said. Before the simulation, parents filled out a survey about their child’s behavior, and results showed that children whose parents reported worse behavior used different strategies than children who reportedly behaved better.

“When we looked at children whose parents rated them as having more conduct problems, they used different strategies in the repeated game,” Blake said. “They were more likely to defect on the partner and less likely to forgive.”

Children with behavioral issues appeared to employ strategies such as “Grim,” in which after one defection by their partner, children defected until the end of the game. According to the paper, this result showed how conduct problems in childhood can manifest in more complex social situations involving trust and awareness of future social interactions.

However, the researchers indicated several times that the small sample size, among other factors, may have led to a skewed data set. For example, the research was conducted mostly within isolated genders; girls played against girls and boys played against boys, and the children were aware of this gender separation.

“There’s something that I find a little unfortunate in his study,” economics professor Johannes Hörner said. “Namely, girls were matched with girls. The fact that girls cooperate –—that’s great … but I would have loved to know whether they cooperate because they are against girls or whether they cooperate because they are girls.”

Making inferences about psychological approaches to games based on behavior can be tricky, Hörner said, as a great deal hinges on children’s motives, which researchers are unable to observe.

Moreover, a certain event may not occur during the game, not because it is impossible for it to occur, but because the child’s strategies might be successful in preventing the outcome from happening, Hörner said. In other words, one outcome could be the result of a strategy or complete chance.

Despite these challenges, Hörner said he has great hopes for the future study of game theory. In particular, he said researchers would do well to conduct research in the field, especially in situations where subjects are unaware they are taking part in a study.

Still, researchers said they are interested in the psychological aspect of the study. In the future, researchers will aim to learn more about development and how game strategies chosen change over a participant’s lifetime, Rand said.

“The most natural next step is [to] try the same setup with children of various ages and see how strategic reasoning changes over the course of development,” Rand said.

The study was published on Sept. 29, 2015, in Scientific Reports.