5.03 · The Marshmallow Test

The Marshmallow Test
A famous experiment with four-year-olds, and what it does and does not show.

Narrator: In the late 1960s, the psychologist Walter Mischel began a simple experiment at Stanford University. Between 1968 and 1974, more than 600 children at the Bing Nursery School on campus took part. Each child, aged four to six, sat alone at a table with one treat: a marshmallow, a cookie or a pretzel stick. The researcher made an offer: “Eat this one now, or wait until I come back, and you can have two.” Then the adult left for up to fifteen minutes. A bell on the table let the child call the adult back, but then there was no second treat.

Some children ate the treat as soon as the door closed. Others tried everything to resist: they covered their eyes, turned their chairs around or sang songs. A child might whisper: “Don’t look. Don’t look.” In the best-known version of the test, about a third of the children waited long enough to win the second treat.

The surprising part came years later. In 1990, Mischel and two colleagues studied some of the same children as teenagers. Those who had waited longer scored higher on the SAT, a U.S. college entrance exam, and coped better with stress, their parents said. For years, the test seemed like a crystal ball: wait at age four, succeed later.

But one question was missing: why do children wait? In 2013, researchers at the University of Rochester tested 28 children, aged four and a half on average. An adult made a small promise to each child. For half of them she was reliable and kept it; for the other half she broke it. Then each child took a waiting test like Mischel’s. The children who had seen a broken promise waited about three minutes on average. The others waited about twelve. Picture a child who has learned that promises get broken. She might think: “Why wait? Nobody is coming back with two.” That is not weak willpower. It is a reasonable guess.

In 2018, the researchers Watts, Duncan and Quan repeated the test with a larger group. They followed 918 children whose mothers had not finished college, comparing a waiting test at age four with school results at fifteen. Children who waited longer did a little better at school, but the link was only half as strong as in the original studies. And when the team accounted for family background, early thinking skills and the home environment, the link shrank by about two thirds.

So what does the marshmallow test teach us? Waiting is a real skill, and it helps, but it does not decide your future. Much of what looks like self-control may also be trust, and trust depends on a child’s world. To follow through on a plan, make the first step small and add support that keeps its promises: a fixed time, a friend who expects you.