What Does the Marshmallow Test Actually Test?

Photograph by Tom Cockrem/Getty Images

Walter Mischel’s marshmallow test is one of the best-known studies in the history of psychology. In the 1960s, Mischel, then a professor at Stanford, took nursery-school students, put them in a room one-by-one, and gave them a treat (they could choose a cookie, a pretzel stick, or a marshmallow) and the following deal: They could eat the treat right away, or wait 15 minutes until the experimenter returned. If they waited, they would get an extra treat. Tracking the kids over time, Mischel found that the ability to hold out in this seemingly trivial exercise had real and profound consequences. As they matured and became adults, the kids who had shown the ability to wait got better grades, were healthier, enjoyed greater professional success, and proved better at staying in relationships—even decades after they took the test. They were, in short, better at life.

Mischel’s work has been enormously influential, making its way into popular culture (most recently in this year’s romantic comedy The Five-Year Engagement) in a way that few academic studies have. It has changed the way educators and psychologists think about success: The lesson is that it’s not just intelligence that matters, but self-control and patience and being able to tame one’s impulses—from the desire to eat the marshmallow to the desire to blow off an exam or have an affair.

A new study (PDF), however, suggests that we may be taking, at best, an incomplete lesson from Mischel’s work. Celeste Kidd, a cognitive science graduate student at the University of Rochester, is the lead author on the paper. When she was younger, Kidd spent some time working in shelters for homeless families. She began to wonder how growing up in such a setting, full of change and uncertainty, might shape the way kids responded to the sort of situation Mischel’s study presented. “Working there gave me some strong intuitions about what kids who were in that situation would do, given the marshmallow task,” she says. “I’m fairly sure those kids would eat the marshmallow right away.” Not because they were weak-willed, but because very little in their upbringing had given them much reason to believe that adults would do what they said they would. What was missing from Mischel’s famous experiment, Kidd argues, was trust.

Kidd’s own version of the marshmallow study was designed to test the effect of trust. First, the three- to five-year-olds in the study were primed to think of the researchers as either reliable or unreliable. In the first part of the study, the researchers handed over a piece of paper and a jar of used crayons, then told a child to either use those crayons or wait for a better set of art supplies. In the second part of the study, the experimenter gave the child a small sticker and told the young subject to either use that one or wait for bigger, better stickers. For half the kids, the experimenter kept the bargain, returning with a loaded tray of markers, crayons, and colored pencils, then several big stickers. For the other half, the experimenter returned a few minutes later to say, apologetically, that there weren’t in fact any better art supplies or any better stickers.

After this, the kids were given the marshmallow test. The results were dramatic: Nine out of the 14 kids in the reliable condition held out 15 minutes for a second marshmallow, while only one of the 14 in the unreliable condition did. If kids were unsure they were going to get a second marshmallow, they didn’t bother to wait.

As it turns out, Mischel himself has looked at the role trust and confidence play in a person’s ability to delay gratification. Reached while traveling in Europe and asked about the new study, he responded with an e-mail linking to three of his early papers. One of them, from 1961, looked at whether coming from a fatherless household affected a child’s willingness to wait for a reward.

But the descriptions of Mischel’s work have focused mainly on determination and grit, and many of the charter schools and educational researchers that have taken the marshmallow results to heart tend to see self-control as a unitary quality that can explain both our childhood decisions and our adult outcomes. In Kidd’s study, the willingness to wait is more of a situational trait. Rather than being engaged in a desperate struggle against their own appetites, the young subjects of her study were carefully calculating the likelihood that they would actually get a second marshmallow. Her work suggests that getting kids to be better at waiting—in the lab and in life—is a matter of persuading them that there’s something worth waiting for.