Originally published in June 2013 and comprehensively updated/edited in July 2026.
From the famous marshmallow test to modern neuroscience, discover how research on delayed gratification transformed our understanding of self-control, executive function, and long-term success.
In the early 2010s, a strange cultural obsession took hold. Videos of small children sitting alone at a table, staring down a single marshmallow, spread across the internet by the thousands. Some kids poked the marshmallow suspiciously. Some sniffed it. A few dismantled it slowly, savoring tiny pieces as if that somehow didn’t count as eating it. Others simply gave in within seconds. The clips were funny, but they were also doing something more substantial: introducing millions of people to one of the most cited experiments in psychology, and to a question that has occupied scientists for more than fifty years. Why can some people wait for a better reward later, while others can’t help but take the smaller one now, and what does that difference actually predict about a life?
The timing mattered. The early 2010s were a period of growing public fascination with willpower, productivity, and self-optimization. Popular books promised that discipline could be trained like a muscle. Employers wanted employees who could resist distraction. Parents wanted children who could sit still, finish homework, and delay the impulse to grab a toy from a sibling. Self-control increasingly came to be seen as the missing ingredient in success, more important, some argued, than raw intelligence. That argument was not just folk wisdom. It was rooted in decades of research, most of it originating with a single, deceptively simple experiment involving a marshmallow, a child, and an empty room.
More than a decade later, the science has matured considerably. Researchers have revisited old assumptions, run larger and more rigorous studies, and in some cases overturned parts of the original story. What has emerged is a richer, more nuanced picture of self-control, one that still confirms its importance but complicates the tidy narrative that made it famous. This is the story of how psychology’s understanding of delayed gratification evolved, and what that evolution reveals about the science of human behavior itself.
A Marshmallow and a Waiting Child
The marshmallow test was developed in the late 1960s and early 1970s by psychologist Walter Mischel and his colleagues at Stanford University. The setup was simple. A child was seated in a bare room with a single treat in front of them, often a marshmallow, sometimes a cookie or pretzel. The experimenter explained that the child could eat the treat right away, or wait alone in the room until the experimenter returned, in which case the child would receive two treats instead of one. Then the adult left, and a hidden camera recorded what happened next.
The results varied widely. Some children waited the full period, which could stretch to fifteen or twenty minutes. Others lasted only seconds. What made the study famous wasn’t the waiting itself, but a follow-up years later. Mischel and his colleagues tracked down some of the original participants as adolescents and found that children who had waited longer tended to have higher SAT scores, better social skills, and fewer behavioral problems as teenagers.<cite index=”0-1″>A 1989 report in Science by Mischel, Yuichi Shoda, and Monica Rodriguez linked longer delay times in preschool to higher cognitive and social competence in adolescence.</cite> That single follow-up study, small by modern standards, became one of the most influential findings in developmental psychology. It suggested that a simple behavior observed in early childhood could forecast outcomes more than a decade later, an almost irresistible idea for both scientists and the public.
Over the following decades, the marshmallow test became shorthand for the entire concept of self-control. It appeared in parenting books, business seminars, and, eventually, viral video compilations. But the original studies involved small, relatively homogenous samples, mostly children connected to Stanford’s campus community. As the test’s fame grew, so did questions about whether its conclusions would hold up in more diverse and rigorously designed research.
From Willpower to Wiring: Executive Function Explained
To understand why psychologists cared so much about a child’s ability to wait for a second marshmallow, it helps to understand the broader mental machinery involved. Researchers now generally discuss self-control as part of a larger set of mental abilities called executive function.
Executive function is not a single skill but a collection of related ones, all managed largely by the brain’s prefrontal cortex, the region behind the forehead responsible for planning and judgment. Developmental psychologist Adrienne Diamond, in an influential review of the field, described three core components: inhibitory control, the ability to resist an impulse or override an automatic response; working memory, the capacity to hold and manipulate information in mind over short periods; and cognitive flexibility, the ability to shift attention between tasks or perspectives.
Resisting the marshmallow required all three. A child had to inhibit the reach-and-eat impulse, remember the promised reward for waiting, and often had to mentally reframe the situation, thinking of the marshmallow as a cloud or a picture rather than food, to make the waiting easier. This reframing turns out to be central to the science, and it points to something important: self-control isn’t primarily about gritting your teeth and enduring temptation through sheer force of will. It’s about strategy.
The Prefrontal Cortex Is a Late Bloomer
One reason young children struggle so much with delay is biological. The prefrontal cortex is among the last brain regions to fully mature, continuing to develop into a person’s mid-twenties. This is not a flaw in the system; it reflects the fact that this region handles some of the most computationally demanding tasks the brain performs, integrating information from memory, emotion, and sensory input to guide long-term decision-making. A four-year-old’s difficulty waiting for a second marshmallow says less about character and more about a still-developing neural architecture. This developmental lens matters for how the research is interpreted, and it recurs throughout later chapters of this story.
The Dunedin Study: Following Self-Control Across a Lifetime
If the marshmallow test asked whether early self-control predicted adolescent outcomes, a much larger and longer-running project asked whether it predicted an entire adult life. The Dunedin Multidisciplinary Health and Development Study followed roughly a thousand children born in Dunedin, New Zealand, in the early 1970s. Researchers tracked them from birth into midlife with remarkable consistency, retaining more than ninety percent of the original cohort over the following decades.
<cite index=”1-2″>Terrie Moffitt, Avshalom Caspi, and their colleagues analyzed this cohort and found that children with lower self-control went on to experience more health problems, greater financial struggles, and higher rates of substance dependence and criminal conviction by the time they reached their early thirties</cite>, even after accounting for intelligence and social class. The self-control measure in this case wasn’t a single marshmallow test but a composite drawn from observations by teachers, parents, and researchers across childhood, covering traits like impulsivity, persistence, and the ability to control frustration.
What made the Dunedin findings particularly compelling was their gradient structure. This wasn’t simply a story of children with poor self-control ending up worse off than everyone else. Outcomes tracked incrementally across the entire spectrum. Children in the middle of the self-control distribution fared worse than those at the top, and better than those at the bottom. This gradient suggested something closer to a general capacity, distributed across the population in the way that height or reaction time might be, rather than a trait that only mattered at the extremes.
The study also offered a hopeful note. Individuals who improved their self-control between childhood and adulthood showed correspondingly better outcomes than their childhood scores alone would have predicted, evidence that this trait, while stable, isn’t fixed. That distinction, between something being consistent and something being unchangeable, turns out to matter a great deal.
Self-Control Versus Intelligence: Untangling the Predictors
One of the more provocative claims to emerge from this research area is that self-control might matter more than intelligence for real-world success. <cite index=”2-3″>Psychologists Angela Duckworth and Martin Seligman tested this directly in a study of eighth-grade students, measuring self-discipline through a combination of self-report, parent and teacher ratings, and behavioral tasks involving delayed reward.</cite> <cite index=”2-4″>They found that self-discipline measured in the fall predicted final grades, school attendance, and standardized achievement test scores more strongly than IQ did, and that self-discipline explained roughly twice as much variance in academic performance as intelligence.</cite>
This finding resonated widely, partly because it offered a more democratic story about achievement. Intelligence, as commonly measured, is difficult to change substantially. Self-control, by contrast, seemed like something a person could build through practice and habit, an appealing idea for educators and parents alike. It’s worth noting, though, that self-control and intelligence are not entirely separate systems. Both draw on overlapping neural resources, particularly in the prefrontal cortex, and children with stronger working memory, itself a component of executive function, also tend to perform better on traditional intelligence measures. The two traits are correlated, not independent, which complicates any simple claim that one “outdoes” the other. What the research more cautiously supports is that self-control adds meaningful predictive power beyond what intelligence alone can explain, not that it operates in isolation from it.
Replication and Revision
Science does not stand still, and one of the most important developments in this field arrived decades after the original marshmallow studies, when researchers subjected the classic finding to a far more rigorous test. <cite index=”3-4″>In 2018, Tyler Watts, Greg Duncan, and Haonan Quan published a conceptual replication of the marshmallow test using a sample more than ten times larger and far more socioeconomically and racially diverse than Mischel’s original Stanford cohort.</cite>
The results complicated the original story considerably. <cite index=”3-5″>The researchers found that delay of gratification at age four was only weakly associated with achievement at age fifteen, and that this association shrank to near insignificance once the researchers statistically controlled for the child’s family background, including maternal education, income, and home environment.</cite> In other words, a large portion of what the original marshmallow test appeared to predict may have actually reflected the circumstances children were born into, circumstances that independently shape both a child’s ability to wait for a treat and their later academic success.
This doesn’t mean the original findings were wrong, exactly, but it does mean they were incomplete, and arguably overstated in the popular retelling. A child raised in a household with reliable access to food and consistent promises kept by adults has less reason to distrust a stranger’s offer of “two marshmallows later.” A child raised amid scarcity or unpredictability has learned, reasonably, that a bird in hand beats one that might never arrive. Self-control, in this light, isn’t purely an internal trait waiting to be measured. It’s also a rational response shaped by a child’s environment and lived experience.
Beyond the Marshmallow: Trust, Reliability, and Rational Waiting
This environmental angle was explored directly in an elegant study by Celeste Kidd and her colleagues, published around the same period as the growing skepticism toward the original marshmallow paradigm. <cite index=”4-5″>Kidd, Holly Palmeri, and Richard Aslin exposed young children to an experimenter who was either reliable, following through on a promise to bring better art supplies, or unreliable, failing to deliver as promised, before administering the standard marshmallow task.</cite> <cite index=”4-6″>Children who had just experienced an unreliable adult waited less than half as long, on average, as children who had experienced a reliable one.</cite>
This study reframed the entire debate. Waiting for a marshmallow isn’t purely a test of willpower; it’s also a bet on whether the promised reward will actually materialize. A child’s decision to eat the treat immediately can reflect sound judgment about an unpredictable world rather than a deficiency in self-control. This finding has quietly reshaped how psychologists talk about delayed gratification, shifting the emphasis from an isolated character trait toward a more contextual understanding that accounts for trust, environment, and lived history.
The Ego Depletion Debate
Parallel to the marshmallow research, another influential idea shaped public understanding of self-control throughout the 2000s and into the 2010s: the notion that willpower functions like a muscle that tires with use. <cite index=”5-6″>Psychologist Roy Baumeister and his colleagues proposed this “ego depletion” model, arguing that exerting self-control on one task consumes a limited psychological resource, leaving less available for subsequent tasks.</cite> Their early experiments seemed to show exactly that pattern: people who resisted eating cookies performed worse on a subsequent difficult puzzle than those who hadn’t faced the temptation.
The ego depletion model dominated the field for roughly two decades, but it eventually became a central case study in psychology’s broader “replication crisis,” the growing recognition that many well-known findings failed to reproduce reliably when tested again with larger samples and stricter methods. <cite index=”6-7″>A large, preregistered multi-laboratory replication effort involving dozens of research teams around the world, led by Martin Hagger and Nikos Chatzisarantis, tested the ego depletion effect using a standardized protocol and found no significant evidence of the effect overall.</cite>
This doesn’t mean self-control never fluctuates; most people know from experience that a long day of hard decisions can leave them feeling mentally worn down. But the tidy “muscle” metaphor, with its implication of a single, quantifiable resource being steadily drained, has not held up well under scrutiny. Current thinking leans toward a more flexible explanation, one involving shifting motivation and attention rather than a depleting fuel tank. This shift illustrates something important about how science actually progresses: not through a single decisive study, but through years of accumulated testing, some of it disproving ideas that once seemed settled.
Self-Control as Strategy, Not Just Suppression
Perhaps the most useful shift in this research area has been a move away from thinking of self-control as pure endurance, gritting one’s teeth against temptation, and toward thinking of it as a set of learnable strategies. Mischel himself, in his later career, described a “hot and cool” systems framework, distinguishing between an emotional, reflexive “hot” system that responds instantly to temptation, and a more reflective “cool” system capable of planning and delay.
The children who waited longest for their marshmallow generally weren’t suppressing desire through raw effort. They were distracting themselves, singing quietly, covering their eyes, or mentally transforming the marshmallow into something less appealing, like a picture rather than food. This is a strategic reframing of attention, not brute willpower, and it turns out to be a skill that can be taught. Later interventions built on this idea, teaching children to use similar attention-shifting techniques, produced measurable improvements in delay ability. The implication is encouraging: self-control, at least in part, is less a fixed trait than a toolkit of techniques, some of which can be handed to a child directly.
Self-Control in the Digital Age
The world into which these findings have landed looks very different from the one in which Mischel first ran his experiments. Smartphones, algorithmically curated feeds, and notification systems are, in many cases, designed to maximize immediate engagement, creating environments that can compete with the sustained attention involved in delayed gratification. Where a 1970s child faced an isolated marshmallow in a quiet room, a modern teenager faces a device engineered by teams of designers specifically to make immediate reward, a like, a message, a new video, more appealing than whatever long-term task competes for their attention.
This context matters for interpreting the research. Executive function develops through practice and reinforcement, and environments that constantly reward immediate response over sustained attention may make the cool, reflective system harder to exercise, not because people have become inherently weaker-willed, but because the environment has shifted the balance of incentives. Some researchers studying attention and digital behavior have drawn a direct line between findings on delay of gratification and concerns about smartphone use, arguing that persistent interruption may function similarly to the unreliable-experimenter condition in Kidd’s study: an environment that trains rapid response rather than patient waiting. This remains an active and still-developing area of research, and claims connecting screen time directly to executive function deficits should be treated as an evolving hypothesis rather than settled fact.
What This Means for Raising and Teaching Children
Taken together, the modern research suggests a more balanced approach to fostering self-control than the popular “just teach kids to wait” narrative implied a decade ago. Consistency and reliability from caregivers appear to matter as much as direct training in patience; a child who trusts that promises will be kept has less reason to grab for the immediate reward. Teaching concrete attention-management strategies, distraction, reframing, breaking a task into smaller steps, appears more effective than simply instructing a child to “use willpower.” And because executive function continues developing well into early adulthood, patience with a child’s, or even a teenager’s, struggles with impulse control is not indulgence; it reflects an accurate understanding of how the brain matures.
None of this diminishes the value of self-control as a predictor of life outcomes. The Dunedin study’s findings on health, finances, and long-term stability remain among the most robust in this literature, replicated in various forms across other longitudinal cohorts. What has changed is the explanation for why self-control matters and where it comes from, a shift from viewing it as an innate, fixed trait revealed by a single childhood test, toward understanding it as a capacity shaped continuously by upbringing, social trust, neural development, and practice.
Conclusion: A More Complicated, More Useful Picture
The marshmallow test endures in public memory because it captures something intuitively true: the ability to wait for a better outcome does seem connected to how a life unfolds. But the science behind that intuition has grown considerably more sophisticated since those first grainy videos of children resisting temptation went viral. Researchers now understand self-control not as a single willpower reservoir but as a collection of cognitive skills, shaped by brain development, family environment, and the reliability of the world a child grows up in.
This more complicated picture is, in some ways, a more useful one. It suggests that self-control is neither a fixed inheritance nor a simple matter of moral effort, but a skill embedded in context, one that can be supported through consistent environments, explicit strategies, and patience with the slow maturation of the developing brain. As new technologies continue to reshape the environments in which attention and patience are tested daily, this line of research seems likely to remain as relevant in the coming decades as it has been over the last fifty years, a reminder that even our most cited experiments are rarely the final word, only an early chapter in a much longer scientific conversation.
Scientific Reference
Mischel, W., Shoda, Y., & Rodriguez, M. L. (1989). Delay of gratification in children. Science, 244(4907), 933–938.
Mischel, W., Ebbesen, E. B., & Zeiss, A. R. (1972). Cognitive and attentional mechanisms in delay of gratification. Journal of Personality and Social Psychology, 21(2), 204–218.
Moffitt, T. E., Arseneault, L., Belsky, D., Dickson, N., Hancox, R. J., Harrington, H., Houts, R., Poulton, R., Roberts, B. W., Ross, S., Sears, M. R., Thomson, W. M., & Caspi, A. (2011). A gradient of childhood self-control predicts health, wealth, and public safety. Proceedings of the National Academy of Sciences, 108(7), 2693–2698.
Duckworth, A. L., & Seligman, M. E. P. (2005). Self-discipline outdoes IQ in predicting academic performance of adolescents. Psychological Science, 16(12), 939–944.
Watts, T. W., Duncan, G. J., & Quan, H. (2018). Revisiting the marshmallow test: A conceptual replication investigating links between early delay of gratification and later outcomes. Psychological Science, 29(7), 1159–1177.
Kidd, C., Palmeri, H., & Aslin, R. N. (2013). Rational snacking: Young children’s decision-making on the marshmallow task is moderated by beliefs about environmental reliability. Cognition, 126(1), 109–114.
Baumeister, R. F., Bratslavsky, E., Muraven, M., & Tice, D. M. (1998). Ego depletion: Is the active self a limited resource? Journal of Personality and Social Psychology, 74(5), 1252–1265.
Hagger, M. S., Chatzisarantis, N. L. D., et al. (2016). A multilab preregistered replication of the ego-depletion effect. Perspectives on Psychological Science, 11(4), 546–573.
Diamond, A. (2013). Executive functions. Annual Review of Psychology, 64, 135–168.
