The Marshmallow That Was Never Only About Willpower

In the late 1960s, at the Bing Nursery School on the Stanford campus, the psychologist Walter Mischel began leaving small children alone in a plain room with a treat, usually a marshmallow, and a single rule. If the child waited until the researcher returned, a second treat would follow. If the child rang a bell, the researcher would come back at once, but the child would keep only the first. Some ate within seconds. Others held out for about fifteen minutes.

Years later Mischel and his colleagues followed up with some of the children, fewer than a hundred in all, most of them from families connected to the university. They reported that those who had waited longer tended to score higher on the SAT and were described by their parents as better able to cope with stress. The finding travelled well beyond psychology, into parenting books and school programmes, as proof that willpower at four shapes a life.

In 2018 Tyler Watts, Greg Duncan and Haonan Quan tested the idea again on roughly nine hundred children from a far more varied range of families. They reported that the link between waiting and later achievement was about half as large as before, and that it shrank further, to a small size, once they accounted for family background and early thinking skills.

Another experiment suggested why. In 2013 Celeste Kidd and her colleagues at the University of Rochester let children meet an adult who either kept a promise about better art supplies or broke it. Those who had met the unreliable adult waited far less time for the marshmallow afterwards. Eating the marshmallow can be a sensible bet by a child who has learned that promises are often broken.

Eating the marshmallow can be a sensible bet by a child who has learned that promises are often broken.

Mischel himself never claimed that waiting was a fixed inheritance. His work reported that children who waited were using strategies, such as looking away, singing, or imagining the treat as a picture rather than a sweet, and that such strategies can be taught. The popular version kept the number and discarded the method.

Even so, the replication did not erase the effect, which remained small but not nonexistent, and researchers still disagree about what the remainder reflects. It may be patience, or privilege, or something no one has yet measured. The child in the room is still deciding, and no one is quite certain what the decision measures.