The Marshmallow Test: What It Shows About Self-Control
Author: Ukrainian Psychological Hub · Published: October 10, 2026 · Editorial Policy
The marshmallow test is a classic experiment in developmental psychology in which a young child can accept a smaller reward immediately or wait for a larger reward. It became famous because some early follow-up studies linked preschool waiting time to later academic and social outcomes. The contemporary evidence tells a more detailed story: children's waiting reflects attention, reward preferences, expectations, trust, social context, and circumstances as well as developing self-control. A single waiting time is an inadequate measure of a child's character or future.
The central scientific result is twofold. In carefully designed experiments, changing how children attend to a tempting reward can change how long they wait. In longitudinal research, the ability to wait has sometimes correlated with later outcomes, but those correlations shrink or vary when researchers consider family background, early cognitive ability, and the home environment. A preregistered study tracking participants to age 26 found that preschool performance did not reliably predict adult functioning after adjustments Sperber et al., 2024.
These findings illuminate a broader problem in psychological measurement. Observing a choice is straightforward; deciding exactly which psychological processes caused it, and how far it predicts behavior years later, requires much more evidence. This article reconstructs the original experiment, the major replications and reanalyses, the psychology of waiting, the influence of reliability and social conditions, and what parents, educators, and adult readers can responsibly learn.
What Is the Marshmallow Test in Psychology?
The marshmallow test is an experimental delay-of-gratification task associated with psychologist Walter Mischel and colleagues at Stanford University. In a commonly described version, a preschool child sees one treat and is told that waiting for the experimenter to return will earn a larger reward, often two treats. The child can obtain the smaller reward sooner by signaling to the experimenter or eating the available treat, depending on the particular procedure. Researchers record the child's behavior and waiting time.
Delay of gratification refers here to postponing an attractive immediate reward to pursue a more desirable outcome later. It can involve resisting the immediate stimulus, shifting attention, thinking differently about the reward, or deciding whether the promised future reward is worth waiting for. The task is one controlled observation of such a decision, not a general intelligence test, clinical examination, diagnostic instrument, or universally valid measure of willpower.
Popular retellings often describe every participant as facing precisely one marshmallow now versus two marshmallows after 15 minutes. The original research program included different rewards, waiting intervals, instructions, and experimental conditions. Some studies examined which strategies helped children wait; others followed subsets of participants over many years. These components should be distinguished when interpreting claims about the experiment Mischel & Ebbesen, 1970; Shoda et al., 1990.
How Did the Original Marshmallow Experiment Work?
The laboratory arrangement made the conflict between two outcomes concrete. A young child encountered a treat that was available now and a more appealing treat promised after a wait. The experimenter left the child in a room under specified conditions. Researchers observed when the child ended the wait, whether the child completed it, and what the child did to manage temptation. Procedures differed across the studies, so the precise stopping rule and maximum interval must be reported with the relevant experiment.
The best-known study is a family of experiments rather than one moment captured in a viral video. In the 1970 paper, Mischel and Ebbe Ebbesen manipulated whether children could look at the reward objects. In a 1972 series, Mischel, Ebbesen, and Antonette Zeiss investigated attentional and cognitive strategies. Both papers demonstrated experimentally that the arrangement of a waiting situation can matter greatly for delay behavior Mischel & Ebbesen, 1970; Mischel et al., 1972.
Why were children asked to wait?
The scientific problem was to examine voluntary delay in a conflict between an immediately accessible temptation and a preferred but postponed outcome. This is a revealing conflict because the larger reward is attractive, yet pursuing it requires managing a period in which the smaller reward remains tempting. Researchers could then manipulate the features of that period and identify processes that make waiting easier or harder.
What counted as success in the laboratory?
In a particular protocol, reaching the specified time limit counted as successfully waiting, while ending early yielded the smaller reward. This laboratory criterion allowed researchers to compare experimental groups. It should not be carried over as a moral criterion for children. A child might prefer the immediate treat, doubt the promised return, be uncomfortable in an unfamiliar room, or have had different experiences with adult promises. The observed choice cannot by itself distinguish all of these explanations.
What Did the First Experiments Actually Demonstrate?
The strongest causal evidence from the classic experimental work concerns the immediate conditions of waiting. When rewards were highly salient, waiting could be harder. When children were distracted from the reward, they often waited longer. Researchers altered attention and cognition within the task, so these differences support the claim that the waiting situation and the child's strategy can change behavior Mischel & Ebbesen, 1970; Mischel et al., 1972.
This is a meaningful finding for the psychology of self-control. It shows why behavior can improve when temptation is made less prominent or a person has another way to occupy attention. The inference is specific: strategies can change performance on a delay task under studied circumstances. It does not follow that any brief distraction exercise permanently changes personality or causes later academic success.
Attention and distraction
Attention determines what information remains psychologically available during the wait. Watching a tempting reward can repeatedly refresh the desire to consume it. Playing, looking away, singing, or imagining something unrelated can reduce the frequency of those reminders. In the original studies, experimental differences in attentional focus produced differences in waiting. These results make attention a plausible mechanism rather than treating waiting as a pure test of stubborn endurance.
How children represent the reward
Another line of experiments examined the way a reward is mentally represented. Thinking about a treat in terms of abstract or less arousing features can differ from dwelling on its appealing taste or texture. Mischel's research linked these cognitive representations to the ease of waiting. The lesson is that how a child construes a temptation can affect behavior even while the external reward remains unchanged Mischel et al., 1972.
The Hot–Cool Model: A Theory of Self-Control
Janet Metcalfe and Walter Mischel later proposed a hot–cool model of delay of gratification. In this account, relatively immediate, affectively charged responses are described as hot, while more reflective, cognitively mediated processes are described as cool. Their interaction helps explain why stress, attention, development, and representations of rewards might affect waiting Metcalfe & Mischel, 1999.
The model is a theoretical interpretation of a family of experimental findings. Hot and cool refer to psychological functions in the framework, not a literal pair of brain compartments that cleanly divide human decisions. Describing a child as possessing a permanently strong or weak cool system on the basis of one test goes beyond what this model establishes.
Modern accounts can also consider reward valuation, executive attention, expectations about the future, learned social rules, and how a decision fits the child's goals. The marshmallow task offers a useful window into these interacting processes precisely because multiple routes can produce the same waiting time.
The Famous Long-Term Claim: Did Waiting Predict Future Success?
The marshmallow test entered popular culture chiefly through follow-up findings. In 1988, Mischel, Yuichi Shoda, and Philip Peake reported that preschool waiting behavior was associated with some adolescent competencies assessed later, including parental ratings. The published analysis concerned 95 participants for whom those ratings were available Mischel, Shoda, & Peake, 1988.
A 1990 follow-up examined conditions under which delay time predicted adolescent cognitive and self-regulatory outcomes. Importantly, the authors emphasized that particular experimental conditions made a difference to the predictive value of waiting. The result was not an unconditional claim that every version of the task measures the same stable trait Shoda, Mischel, & Peake, 1990.
Those early associations inspired a widely repeated narrative: children who waited for the second treat would grow into more successful adults. That narrative stretches the original evidence in three ways. First, association is not proof that waiting caused later outcomes. Second, the early follow-up samples were selective and limited. Third, a childhood performance measure may reflect stable aspects of the environment as well as a child's own self-regulatory strategies.
Prediction is different from causation
A longitudinal correlation asks whether people who scored differently as young children also differ on an outcome years later. It does not demonstrate that changing the childhood score would change the future. Family opportunities, social support, early learning environments, cognitive development, and many other variables can shape both the child's willingness to wait and later life outcomes. Researchers therefore need careful adjustment, appropriate comparison groups, and ideally interventions with long-term follow-up before making a causal claim.
How much does an individual score tell us?
Even a statistically meaningful average association can be weak as a prediction for one particular person. Children who waited for a shorter period can achieve excellent outcomes; children who waited longer can later face setbacks. Waiting time also has a ceiling and can bunch participants together at short or long intervals. A test may identify some group-level variation without providing a reliable ranking of children's futures.
The 2018 Replication: Smaller Associations After Adjustment
A pivotal conceptual replication by Tyler Watts, Greg Duncan, and Haonan Quan analyzed a larger and more diverse cohort than the original Stanford follow-ups. Their 2018 paper focused on children whose mothers had not completed college and examined associations between delay at approximately age four and outcomes at age fifteen. They found a modest unadjusted association with later achievement, but it was substantially smaller than the correlations emphasized in classic retellings Watts et al., 2018.
In that sample, another minute waited corresponded to about one tenth of a standard deviation in achievement in the unadjusted analysis. Adding controls for family background, early cognitive ability, and home environment reduced the association by roughly two thirds. Much of the achievement variation was associated with whether the child waited at least approximately twenty seconds, rather than a simple linear reward for waiting longer and longer. Associations with adolescent behavioral measures were smaller and rarely statistically significant Watts et al., 2018.
The study mattered because it tested an influential claim in a population with a different social and educational composition and used detailed covariates. It challenged the idea that preschool waiting time is a powerful stand-alone indicator of later success. It also showed how the apparent predictive strength of a measure can change once other plausible explanations are considered.
The 2018 findings should be interpreted with the same care demanded of the earlier work. Statistical adjustment helps address measured confounding but cannot prove that every source of confounding has been removed. The results characterize that study's sample, specific measures, and outcomes. They establish a serious limit on broad public claims rather than proving that every possible version of delay measurement is uninformative.
The 2020 Reanalysis: Why Scientists Reached Different Conclusions
Laura Michaelson and Yuko Munakata published a preregistered secondary analysis of the same large data set. Using an analytic approach based more closely on aspects of the original studies, they found statistically significant associations for three of five adolescent outcomes. In particular, they reported some relationships with problem behavior that had not emerged as significant in the 2018 analysis. Their paper provides an unusually clear case of researchers obtaining different conclusions from overlapping data because outcomes and analytic choices differ Michaelson & Munakata, 2020.
Crucially, these authors did not establish that willpower alone explained those outcomes. Their analyses suggested that social support better accounted for some of the predictive relationships than the isolated interpretation of self-control. This makes the apparent disagreement scientifically productive: one analysis emphasizes attenuated associations after adjustment, while another identifies selected associations and interprets them through children's environments.
The combined evidence favors a measured conclusion. Preschool delay behavior can carry information about current psychological and social conditions, and some associations with later outcomes may appear under certain analyses. The size, robustness, and interpretation of those associations depend on outcome definitions, sampling, model choices, and background factors. None of these studies validates using an individual marshmallow score as a life forecast.
The 2024 Follow-Up: What Happened by Age 26?
A more recent preregistered longitudinal study followed 702 participants who had completed a marshmallow-style task at approximately 54 months of age in 1995–1996 and provided measures at age 26 in 2017–2018. The paper, by Jessica Sperber, Deborah Lowe Vandell, Greg Duncan, and Tyler Watts, appeared in Child Development in 2024. It evaluated adult achievement, health, and behavior rather than stopping at adolescent outcomes Sperber et al., 2024.
The unadjusted associations included a modest correlation between early waiting and educational attainment (r = .17) and an inverse correlation with adult body mass index (r = −.17). Nearly all adjusted regression coefficients were nonsignificant after accounting for demographic background, the early home environment, and relevant developmental abilities. The authors concluded that marshmallow task performance did not reliably predict adult outcomes in their analyses Sperber et al., 2024.
This is a stronger test of some exaggerated claims because it examines later adulthood with a relatively large sample and preregistered analyses. It does not imply that self-control, supportive environments, or long-term planning are irrelevant. It means that this particular preschool behavioral measure had limited long-range predictive validity after the factors considered in the study were taken into account.
The authors also examined differences by socioeconomic status and sex and did not find a clear, consistent moderation pattern for adult outcomes. That finding should not be rewritten as evidence that socioeconomic conditions never matter: other experiments show that trust and the reliability of the immediate setting can change waiting performance, and broader social conditions shape development in multiple ways.
Does the Marshmallow Test Measure Self-Control or Trust?
Both constructs can matter, along with additional factors. A child who believes an adult will return with the promised reward faces a different decision from a child who considers that promise unreliable. In the second situation, accepting the immediately available option may protect against losing both. The capacity to resist temptation cannot be inferred cleanly without considering expectations about whether delayed rewards will arrive.
Celeste Kidd, Holly Palmeri, and Richard Aslin tested this proposition in 2013. Children first experienced an experimenter who either fulfilled or failed to fulfill promises about art materials. Children exposed to the reliable experimenter subsequently waited longer in a marshmallow task. The study involved 28 children, a small experimental sample, but it demonstrated that an experimentally manipulated reliability cue can alter waiting Kidd et al., 2013.
A later extension by Lillie Moffett, Carol Flannagan, and Priti Shah studied 60 children in a more familiar school environment. Waiting was again longer in the reliable condition, although the observed difference was substantially smaller than in the original manipulation. The authors' Bayesian analysis supported a reliability effect, while conventional analysis offered a less decisive result. The study illustrates both the plausibility of the mechanism and uncertainty about how large and robust it is across settings Moffett et al., 2020.
Trust does not replace every other mechanism. Children can trust an experimenter yet find the treat unusually tempting, struggle with boredom, or have difficulty shifting attention. Children can also wait for social reasons even when the reward itself is less appealing. The evidence supports studying interacting causes rather than attributing all behavior to a single trait or circumstance.
How Social Norms and Cooperation Change Waiting
Children learn what behaviors are valued by observing other people and understanding group expectations. In experiments reported by Sabine Doebel and Yuko Munakata, preschoolers assigned to a group waited longer when told that their group typically waited and an outside group did not. Beliefs about group behavior changed both task performance and, in one experiment, children's evaluation of waiting Doebel & Munakata, 2018.
A 2020 study by Rebecca Koomen, Sebastian Grueneisen, and Esther Herrmann used a modified cooperative marshmallow task involving 207 children in Germany and Kenya. Children's success depended on a partner also delaying gratification. In these interdependent circumstances, children were more likely to wait than in a standard individual version. The finding supports the importance of social goals and cooperative meaning for delay decisions Koomen et al., 2020.
These findings expand the scientific question. Waiting may express concern for a partner, adherence to a group norm, or a belief that a joint goal is worthwhile, as well as management of private temptation. Social context is part of the behavior being measured. Group studies also caution against treating the marshmallow task as culturally neutral: the meaning of waiting and the trustworthiness of promises can vary with the setting and relationships involved.
What the Marshmallow Test Can and Cannot Tell Us
The task can inform developmental research when procedures and outcomes are specified carefully. Experimental variations demonstrate that attentional strategies, salient reward cues, perceived reliability, and social expectations can affect waiting. Longitudinal analyses can test whether observed waiting is associated with later measures and how those associations change with adjustment. These are substantive contributions to psychological science.
The task cannot establish a child's global self-control level with the certainty of a diagnostic test. It cannot determine whether a child has ADHD or any other condition. It does not directly measure intelligence, conscientiousness, moral worth, resilience, or guaranteed academic achievement. It cannot tell caregivers that a child who eats the first reward lacks discipline or is destined to struggle in adulthood.
A more precise way to read a score is to ask what behavior occurred under what set of rules, with which rewards, at what age, in what setting, and with what expectations. The same child could behave differently after a change in instructions, fatigue, reward preference, room environment, trust, or peer norms. Researchers can use such variation to understand processes; parents should avoid collapsing it into a label.
Is the marshmallow test a reliable personality test?
It is primarily a behavioral paradigm devised for psychological experiments, with multiple versions. To treat any version as a standardized personality assessment would require evidence about test–retest stability, validity across contexts, normative interpretation, and prediction at the individual level. The longitudinal and experimental literature discussed here does not support using an informal one-off marshmallow task as a clinical or educational ranking tool.
Can waiting be a rational decision rather than a character trait?
Yes. A decision to wait depends partly on the probability that the promised reward will materialize, the relative value of each option, and the burden of waiting. When a child expects promises to be honored, delaying may be attractive. When experience indicates uncertainty, consuming what is available now may be a reasonable response. The reliability experiments directly demonstrate that such beliefs can matter, without settling every question about each child's private reasoning Kidd et al., 2013.
The Main Scientific Limitations
The original Stanford follow-ups involved relatively small and selected groups. This constrains how confidently the results can be generalized to children from other communities. Different task versions create additional difficulty because visibility of rewards, waiting duration, scoring rules, and instructions can change behavior. Researchers must avoid comparing waiting times as though these procedures were interchangeable.
Confounding is a central problem in the longitudinal evidence. Family resources, home stability, cognitive development, and supportive relationships may influence both the childhood task and later outcomes. Including these factors in statistical models changes estimates, but observational models still depend on what was measured and how. The 2018 and 2020 analyses also demonstrate the consequences of choosing different outcomes and model specifications Watts et al., 2018; Michaelson & Munakata, 2020.
Selection and attrition also matter. Long-term cohorts retain only a subset of the people first recruited, and follow-up respondents may differ from nonrespondents. A study's outcome measures may be more or less sensitive to the behavior it seeks to explain. Statistical significance should be evaluated alongside effect size, confidence intervals, preregistration, replication, and the practical usefulness of prediction.
Finally, the task's construct validity is contested when interpreted too broadly. Waiting combines cognitive and motivational factors with situational conditions. A useful behavioral measure need not represent a single stable psychological trait. The 2024 adult follow-up provides particularly important evidence against interpreting preschool wait time as a powerful, context-free forecast of adult achievement, health, or behavior Sperber et al., 2024.
What Does This Research Mean for Parents and Educators?
The most responsible application is to create circumstances in which children can learn, trust, and practice manageable ways to pursue goals. Reliable promises matter: if adults repeatedly fail to deliver promised outcomes, waiting becomes a different decision. Clear expectations, consistent follow-through, opportunities for play, and developmentally appropriate support can help children understand that waiting sometimes produces worthwhile benefits.
The classic attention experiments offer a practical principle: redirecting attention away from a salient temptation can make a difficult wait easier in the moment. With young children this may involve engaging in an absorbing activity, leaving a treat out of sight when appropriate, or using age-suitable distraction. Such strategies should be offered as supports rather than turned into an endurance competition Mischel & Ebbesen, 1970.
Social settings matter too. When children see cooperative behavior, understand shared goals, and have reliable partners, their willingness to delay can change. Adults can model patience and communicate clear, realistic expectations. Experimental results from group and cooperation tasks provide evidence for social influence on waiting behavior, while leaving open questions about whether specific parenting programs based on these tasks lead to enduring benefits Doebel & Munakata, 2018; Koomen et al., 2020.
Caregivers should also consider immediate needs. Hunger, fatigue, distress, and the child's understanding of a task can affect willingness to participate, even if a particular study did not measure each factor. A child may need nourishment, reassurance, or a shorter, more comprehensible goal. It would be inappropriate to use food deprivation, public comparison, humiliation, or coercive waiting exercises to test supposed potential.
Evidence that specific laboratory manipulations affect waiting does not establish that deliberately raising a child's marshmallow score will improve adult earnings, relationships, health, or academic attainment. The more durable educational objective is to support healthy development through consistent care, learning opportunities, predictable environments, and a repertoire of flexible strategies.
What Can Adults Learn from the Marshmallow Test?
Adults routinely face versions of the same temporal conflict: spending versus saving, immediate entertainment versus a valued assignment, convenience versus a long-term health goal, or a short-term emotional response versus a relationship priority. The analogy is useful when it draws attention to the interaction among immediate rewards, future values, beliefs, and environment. Adult decisions involve additional obligations and forms of reasoning that are absent from a preschool snack task.
One practical implication is to redesign the moment of temptation. Instead of relying exclusively on effortful resistance, a person might place a phone outside the workspace, prepare tomorrow's materials in advance, or make a valued action the easiest available next step. The marshmallow studies provide conceptual support for the importance of cue visibility and attention; whether any specific adult strategy works should be judged using research on that strategy, rather than extrapolated directly from a childhood experiment.
A second implication is to examine the credibility of future rewards. It is easier to commit effort when the outcome is realistic, meaningful, and supported by dependable circumstances. A delayed benefit that is uncertain or inaccessible may not be worth the same investment as a dependable one. Good self-regulation often involves evaluating competing priorities and changing circumstances, not waiting automatically whenever a larger future reward is promised.
For a broader understanding, see the English Hub's overview of motivation, habits, and self-control and its specialized explanation of self-control mechanisms. Research on the marshmallow experiment addresses a particular evidence question within this larger domain.
Marshmallow Test, Delay of Gratification, and Related Concepts
The marshmallow test is one experimental method. Delay of gratification is the broader psychological phenomenon of postponing an immediately available benefit for a preferred future benefit. Self-control refers to processes that help coordinate impulses and actions with more valued objectives. Self-regulation is broader still and can include adjusting goals, emotions, attention, and behavior as circumstances change. These terms are related but occupy different explanatory levels.
Inhibitory control refers more specifically to suppressing or interrupting an inappropriate or dominant response in certain tasks. Temporal discounting refers to how the subjective value of an outcome changes with delay; it is commonly studied through choices about sooner and later rewards. Habit refers to learned tendencies for situational cues to elicit relatively automatic behavior. A marshmallow waiting time cannot serve as an interchangeable score for all of these processes.
The Hub's separate article on delay of gratification explains the general construct across everyday decisions. The self-control versus self-regulation comparison addresses the conceptual boundary. This article remains focused on the historical marshmallow experiments, their methods, competing analyses, and evidential meaning.
Can the Marshmallow Test Diagnose ADHD or Other Conditions?
No. An informal marshmallow task cannot diagnose ADHD, depression, anxiety, or any other psychological or medical condition. Difficulties with waiting, task initiation, frustration, or impulse management may occur for many reasons, including age, circumstance, stress, reward preference, and ordinary individual variation. A clinical assessment uses a much broader history, pattern of symptoms, functional impact, and established diagnostic standards.
A child who quickly eats a treat has not thereby demonstrated a disorder. A child who waits for the maximum time has not thereby ruled one out. Concerns about persistent impairment across settings should be discussed with a qualified pediatric or mental-health professional rather than inferred from a viral experiment or a home-made score.
The same distinction applies to adults. Struggling to postpone an attractive activity or to begin a demanding task does not identify a specific diagnosis. Educational strategies can be useful for ordinary goal conflict; continuing and significant difficulties warrant context-sensitive assessment rather than a moral or neurological label.
Frequently Asked Questions About the Marshmallow Test
Who invented the marshmallow test?
Psychologist Walter Mischel and colleagues developed and investigated classic delay-of-gratification procedures, including influential experiments with Ebbe Ebbesen and Antonette Zeiss at Stanford University. The best-known papers date to 1970 and 1972 Mischel & Ebbesen, 1970; Mischel et al., 1972.
What does the marshmallow test measure?
It measures observed willingness or ability to delay a particular reward under a particular set of experimental conditions. Interpretation may involve attention, temptation management, trust, valuation, and social expectations. It is not a comprehensive or clinically validated score of a child's self-control across life.
Is the marshmallow test debunked?
The experiments still demonstrate that waiting behavior responds to psychological and contextual manipulations. The exaggerated claim that one preschool waiting time reliably determines future success is unsupported by the strongest later follow-up evidence. The accurate conclusion is that the test remains scientifically informative for certain questions, while its long-term individual predictive value is limited and contested Sperber et al., 2024.
Did the 2018 replication find no relationship at all?
No. It reported a modest bivariate achievement association that became much smaller after adjustment, and weaker or generally nonsignificant associations for several behavioral measures. A different preregistered analysis of the same data found selected significant outcome relationships but suggested that social support helped explain them Watts et al., 2018; Michaelson & Munakata, 2020.
What did the 2024 study find?
The preregistered follow-up of 702 participants to age 26 found weak unadjusted associations for some outcomes and mostly nonsignificant adjusted coefficients. The authors concluded that preschool marshmallow performance did not reliably predict adult achievement, health, or behavior in their analyses Sperber et al., 2024.
Does poverty explain why some children eat the marshmallow sooner?
Economic and family circumstances may affect experiences of reliability, opportunity, and future rewards, but a single socioeconomic variable cannot explain every child's choice. Experimental studies provide direct evidence that manipulated trust cues affect waiting, while longitudinal studies examine multiple correlated background factors. Socioeconomic circumstances deserve attention without turning them into deterministic predictions about children Kidd et al., 2013; Watts et al., 2018.
Can children learn to wait longer?
Laboratory experiments show that changing attention, instructions, and reliability cues can increase waiting in particular situations. This supports teaching manageable strategies and creating predictable settings. Evidence from a short laboratory improvement does not establish permanent personality change or guaranteed life benefits Mischel et al., 1972; Moffett et al., 2020.
How long are children supposed to wait?
Some famous protocols offered a waiting opportunity of about 15 minutes, while others used different intervals and stopping rules. There is no clinically meaningful pass-or-fail time for children generally. The interval is part of a particular experiment's design, not a developmental threshold for worth, health, intelligence, or future success.
Should parents try the marshmallow test at home?
A playful choice between an immediate and later treat may prompt conversation about waiting, promises, and preferences. It should not be scored as a psychological diagnosis or used to compare siblings. A home version lacks the standardization, safeguards, and interpretation needed for scientific measurement. Avoid making access to food or affection contingent on passing an informal test.
Does the marshmallow test prove that willpower is a limited resource?
No. The marshmallow task concerns delay behavior and its determinants. Claims that exerting self-control depletes a fixed, general-purpose mental resource come from a different experimental literature and require their own evidence, including replication research. The marshmallow studies cannot establish that resource theory simply because waiting sometimes feels difficult.
What is the most important takeaway?
Preschool waiting behavior is shaped by both the person and the situation. Controlled experiments explain ways to alter waiting; longitudinal work places important limits on predicting later life from a single childhood performance. Development is better understood through changing capacities, environments, opportunities, and relationships than through a simplistic marshmallow score.
Conclusion: The Marshmallow Test in Modern Psychological Science
The marshmallow test remains an influential contribution to developmental psychology because it made delayed gratification observable and allowed researchers to study attention and reward conflict experimentally. Its scientific history also illustrates the importance of replication, follow-up samples, confounding, analytical choices, trust, and social context.
The most defensible interpretation is that self-control is expressed through strategies and circumstances, and that an isolated preschool waiting time cannot predict the full course of a person's life. The field gained knowledge from the classic experiments and also from the studies that questioned what their most famous correlations meant.
Related Articles
Motivation, Habits, and Self-Control: Psychology of Behavior Change — the broad conceptual hub connecting motivation, self-regulation, goals, and learned behavior.
Delay of Gratification: Psychology, Context, and Future Outcomes — the broader construct and its different behavioral and decision-making manifestations.
What Is Self-Control? Psychology, Mechanisms, and Examples — the wider psychological definition of self-control.
Self-Control vs. Self-Regulation: Differences and Connections — why overlapping concepts should be distinguished.
How to Improve Self-Control: Evidence-Based Strategies — intervention strategies whose evidence must be evaluated beyond the marshmallow paradigm.
References
Doebel, S., & Munakata, Y. (2018). Group influences on engaging self-control: Children delay gratification and value it more when their in-group delays and their out-group doesn't. Psychological Science, 29(5), 738–748. https://doi.org/10.1177/0956797617747367
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. https://doi.org/10.1016/j.cognition.2012.08.004
Koomen, R., Grueneisen, S., & Herrmann, E. (2020). Children delay gratification for cooperative ends. Psychological Science, 31(2). https://doi.org/10.1177/0956797619894205
Metcalfe, J., & Mischel, W. (1999). A hot/cool-system analysis of delay of gratification: Dynamics of willpower. Psychological Review, 106(1), 3–19. https://doi.org/10.1037/0033-295X.106.1.3
Michaelson, L. E., & Munakata, Y. (2020). Same data set, different conclusions: Preschool delay of gratification predicts later behavioral outcomes in a preregistered study. Psychological Science, 31(2), 193–201. https://doi.org/10.1177/0956797619896270
Mischel, W., & Ebbesen, E. B. (1970). Attention in delay of gratification. Journal of Personality and Social Psychology, 16(2), 329–337. https://doi.org/10.1037/h0029815
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. https://doi.org/10.1037/h0032198
Mischel, W., Shoda, Y., & Peake, P. K. (1988). The nature of adolescent competencies predicted by preschool delay of gratification. Journal of Personality and Social Psychology, 54(4), 687–696. https://doi.org/10.1037/0022-3514.54.4.687
Moffett, L., Flannagan, C., & Shah, P. (2020). The influence of environmental reliability in the marshmallow task: An extension study. Journal of Experimental Child Psychology, 194, 104821. https://doi.org/10.1016/j.jecp.2020.104821
Shoda, Y., Mischel, W., & Peake, P. K. (1990). Predicting adolescent cognitive and self-regulatory competencies from preschool delay of gratification: Identifying diagnostic conditions. Developmental Psychology, 26(6), 978–986. https://doi.org/10.1037/0012-1649.26.6.978
Sperber, J. F., Vandell, D. L., Duncan, G. J., & Watts, T. W. (2024). Delay of gratification and adult outcomes: The marshmallow test does not reliably predict adult functioning. Child Development, 95(6), 2015–2029. https://doi.org/10.1111/cdev.14129
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. https://doi.org/10.1177/0956797618761661