The Marshmallow Test Reconsidered: Trust and Rational Waiting in Childhood
The famous marshmallow test presents a child with a deceptively simple choice: eat one treat now, or wait for an adult to return and receive two. The longer a child waits, the familiar story goes, the more self-control that child possesses.
However, this classic experiment carries an implicit assumption: waiting is worthwhile only if the adult is likely to keep the promise. If the second marshmallow may never arrive, taking the certain reward is not necessarily a failure of willpower. It may be the sensible choice.
A revealing study from the University of Rochester made this assumption visible. Before children faced the marshmallow task, an adult either kept two small promises or broke both. Children who had just seen a reliable adult waited an average of 12 minutes, while those exposed to an unreliable adult waited only about three minutes. This significant difference highlights that the marshmallow test does not purely measure self-control but also a child’s judgment about the trustworthiness of the adult’s promise.
How a Nursery-School Task Became a Character Test
In the late 1960s and early 1970s, psychologist Walter Mischel and colleagues developed the classic delay-of-gratification experiment at Stanford’s Bing Nursery School. Preschoolers were presented with an immediate small reward or a larger reward after waiting alone. The original research investigated how attention, distraction, and mental strategies influenced children’s ability to wait.
Over time, the test became widely interpreted as a measure of inner willpower and discipline, symbolizing future success. This popular narrative suggested that children who waited demonstrated durable self-control, while those who did not were impulsive. Yet, the choice involves more than resisting temptation; it also requires evaluating the reliability of the adult’s promise to return with a larger reward.
The Adult First Established Whether Promises Meant Anything
The Rochester study by Celeste Kidd, Holly Palmeri, and Richard Aslin involved 28 children aged between three and five and a half years. They were randomly assigned to either a reliable or unreliable condition, balanced by age and gender.
Before the marshmallow task, each child took part in an art project using unattractive crayons. The experimenter left the room, promising to return with better supplies. In the reliable condition, the adult delivered on this promise; in the unreliable condition, she returned empty-handed, apologizing for the mistake.
This sequence was repeated with stickers. Again, the reliable adult fulfilled the promise while the unreliable one did not. This setup was a randomized manipulation designed to provide immediate evidence about the adult’s reliability rather than a test of the child’s general trust or background.
Three Minutes After Broken Promises, Twelve After Kept Ones
Following the art project, the marshmallow test was administered. The experimenter placed a single marshmallow within reach and explained the choice: eat one now or wait for two once she returned. The maximum wait was 15 minutes.
Children in the unreliable condition waited an average of just over three minutes (181.57 seconds), whereas those in the reliable condition waited about 12 minutes (722.43 seconds). Notably, only 1 of 14 children in the unreliable group waited the full 15 minutes, compared to 9 of 14 in the reliable group.
This dramatic nine-minute difference arose from a short sequence of kept or broken promises, demonstrating that the children’s behavior was influenced by their assessment of the adult’s trustworthiness rather than a sudden change in personality.
While some interpret this as evidence that the test measures trust more than self-control, the reality is nuanced. Waiting still requires attention control, emotional regulation, and coping strategies. Reliability significantly impacts behavior but is not the sole determinant.
A Rational Choice Still Requires Self-Regulation
From the child’s perspective, the decision weighs a guaranteed single marshmallow against the prospect of two marshmallows later—if the adult returns as promised. When the adult has demonstrated reliability by delivering better crayons and stickers, waiting is a sensible investment. But if the adult has broken promises, waiting carries real risk, and eating immediately secures the only available reward.
This rationality does not imply that four-year-olds consciously compute probabilities but reflects learned expectations about people and environments. Children’s decisions reveal a practical assessment of trustworthiness rather than pure impulsivity.
Self-control remains critical after a child decides the future reward is worth waiting for. To succeed, a child may need to employ strategies such as looking away, singing, or distracting themselves. Thus, the marshmallow test intertwines two questions: “Can I wait?” and “Is it worth waiting?”
This insight aligns with broader research on self-control, emphasizing that behavior emerges not just from willpower but from habits, environmental cues, and the perceived value of effort in context.
A Larger Extension Found the Effect Again, but Smaller
The Rochester study’s small sample (28 children) and large effect size called for replication. In 2020, researchers at the University of Michigan repeated the reliability manipulation with 60 children aged three to five in familiar school settings.
The study again found that children exposed to a reliable adult waited longer than those exposed to an unreliable one. However, the effect was smaller, and analyses revealed interaction effects between gender and condition, reflecting a more complex reality than a simple binary outcome.
This replication supports the key finding that perceived reliability influences waiting behavior but cautions against interpreting specific wait times (e.g., 12 vs. 3 minutes) as fixed childhood traits.
Importantly, the experiment tested an unfamiliar adult’s reliability over a short period and did not directly measure broader factors like family stability, socioeconomic status, or general trust. These variables may influence self-control but require separate, comprehensive study.
The Prediction-of-Success Story Also Became Less Tidy
The marshmallow test gained fame partly due to claims that early waiting predicted later life success. More recent longitudinal studies have nuanced this narrative.
A 2018 replication involving 918 children found that longer waiting at age four was associated with better academic achievement at 15. Yet, this association was roughly half the size of earlier reports and diminished substantially after controlling for family background, early cognitive ability, and home environment.
Associations with behavioral outcomes at age 15 were weaker and often not statistically significant. Additionally, the primary difference was between children who waited at least 20 seconds versus those who didn’t, rather than a smooth gradient where every additional minute predicted more success.
A 2024 preregistered study following 702 participants to age 26 found modest unadjusted relationships between waiting time and educational attainment or body mass index. However, after adjusting for childhood factors, most associations with adult achievement, health, and behavior lost statistical significance.
These findings do not negate the importance of self-control. Indeed, broader and repeated measures of childhood self-control—assessed over multiple years—have demonstrated links to later health and financial outcomes, as previously reported by Silicon Canals.
Such multi-year behavioral assessments provide different, arguably more robust evidence than a single laboratory test involving one adult, one reward, and one afternoon.
One Marshmallow Cannot Diagnose a Child
The Rochester experiment primarily serves as a caution against overinterpreting the marshmallow test outcome. A child who eats early may be motivated by hunger, may value the second treat less, may struggle with attention, or may distrust the adult. They may have learned that “later” is not always dependable.
The test records a final choice but does not reveal the underlying mechanisms. Labeling a child impulsive based solely on this outcome mistakes behavior for explanation.
For adults, the practical takeaway is that delayed rewards become easier to choose when the world demonstrates consistency. Patience is less speculative when promises are reliably kept.
This perspective reframes the marshmallow test without discarding it. Children still regulate attention and desire but also make decisions under uncertainty. Waiting can reflect self-control; eating can reflect justified skepticism.
A second marshmallow is valuable only if it appears. Before judging a child’s willpower, the Rochester experiment urges us to inspect the promise.
Source: Here
