Psychology › Emotion, motivation & happiness › Self-control and willpower
The marshmallow test
1972; re-tested 2018· Finding or experiment
Mixed evidenceSome support; effects smaller or less consistent than once thought.
Walter Mischel's experiment offering preschoolers one treat now or two if they could wait alone, often for 15 minutes.
See it on the Psychology map →What it says
From the late 1960s at Stanford's Bing Nursery School, Walter Mischel left children aged about four or five alone with a treat, often a marshmallow, and a bell to summon the adult. Waiting meant two treats. Children who distracted themselves, by singing or covering their eyes, waited longer. Follow-ups in 1988 and 1990 linked longer waiting to higher SAT scores and better coping in adolescence. The finding became shorthand for 'self-control predicts success'.
Why it matters
It made delayed gratification famous and influenced parenting and schooling. Its revision shows how background and trust shape what looks like willpower.
The evidence
Status: mixed evidence. A 2018 study of about 900 more diverse children found the link to later achievement half as large, shrinking further once family background was counted. In 2013, children waited longer when the adult had proved reliable.
Open questions
Why children in the 2000s waited longer than children in the 1960s, according to a 2018 comparison, despite fears about screens and instant gratification.
Connections (3)
▸The replication crisis weakened claims of The marshmallow testWell replicated
The connection
In 2018 researchers revisited the test with about 900 children from a US national study, far more diverse than Stanford's preschool. Waiting time still predicted achievement at 15, but half as strongly, and counting family background, early skills and home life cut that by about two-thirds. A 2020 follow-up of Mischel's original children found little link to midlife outcomes.
Who showed it and when
Tyler Watts, Greg Duncan and Haonan Quan, New York University and UC Irvine, 2018; Daniel Benjamin and colleagues' midlife follow-up of the original Stanford participants, 2020.
The evidence
Larger, more representative samples with statistical controls the originals lacked. But the 2018 study was a conceptual replication, with a 7-minute limit rather than Mischel's exact setup.
Open questions
How much the remaining link reflects self-control itself versus the advantages of a stable, well-off home.
▸The marshmallow test tests children's resistance to Present bias (hyperbolic discounting)Mixed evidence
The connection
The test is a child-sized version of the choice present bias describes: a small reward now versus a bigger one later, with the immediate option in plain sight. Mischel found that making the treat less vivid, such as imagining it as a picture or a puffy cloud, extended waiting, an early sign that how temptation is framed shapes impatience.
Who showed it and when
Stanford's Bing Nursery School, late 1960s–1970s; Janet Metcalfe and Mischel's 'hot and cool' model, 1999; economists studying delay discounting from the 1980s on.
How solid is it?
Mischel's strategy experiments show framing changes waiting. The link to adult present bias is indirect, since the task also measures trust and hunger.
Open questions
Whether children's waiting reflects stable impatience or a sensible bet about whether the adult will really return with the second treat.
▸The marshmallow test measured children's Self-control and willpowerMixed evidence
The connection
The waiting task was treated as a window on self-control. Mischel's follow-ups linked longer waiting at four to higher SAT scores and better coping years later, and the test became shorthand for willpower predicting success. Mischel himself emphasized learnable strategies, like distraction and 'cool' thinking, over a fixed trait.
Who showed it and when
Stanford's Bing Nursery School, hundreds of children tested 1968–1974; follow-ups by Mischel, Yuichi Shoda and Philip Peake, 1988 and 1990; Mischel's book The Marshmallow Test, 2014.
How solid is it?
Follow-up analyses rested on small subsets of the original children; broader self-control ratings across many situations, as in the Dunedin study, predict outcomes more reliably.
Open questions
Whether a single short task can capture a stable trait, when children's waiting varies with hunger, trust, the specific treat and who runs the test.