The Marshmallow Test Was Wrong About Your Kids

You almost certainly know the story. A small child is left alone with a marshmallow and a promise: wait fifteen minutes without eating it, and you will get a second one. Some children hold out, some give in, and - so the legend goes - the ones who waited went on to better grades, better careers, better lives. The marshmallow test became one of the most repeated parables in modern psychology, a tidy proof that willpower at age four sets the trajectory of a life. It is a wonderful story. It is also, in the strong form everyone remembers, wrong - and the correction matters for how we think about our children.

I want to be precise here, because this is a case where the headline outran the science and then hardened into folklore. The original studies were real and the researchers careful. What failed was the popular conclusion built on top of them: that a single act of childhood self-restraint is destiny. When researchers went back to test that conclusion with a larger, more representative sample, much of the magic dissolved - and what remained pointed less at willpower and more at the world around the child.

What the original marshmallow test actually claimed

The work began with Walter Mischel and colleagues at Stanford. In the best-known follow-up, Mischel, Shoda, and Rodriguez (1989) reported that preschoolers who were able to delay gratification tended, years later, to show stronger outcomes on measures such as academic performance. That correlation is the seed of the entire legend.

But two features of the original work rarely survive into the popular version. First, the samples were small and drawn from a fairly narrow community connected to Stanford - not a broad cross-section of children. Second, and more importantly, a correlation in that group was never evidence that willpower alone caused the later outcomes. It was always possible that something underneath both the waiting and the success - a stable home, educated parents, a sense that promises get kept - was doing the real work. The 1989 findings were genuine. The leap from "is associated with" to "determines" was added later, mostly by people retelling the story.

The replication that rewrote the legend

In 2018, a team set out to test the famous claim properly. Watts, Duncan, and Quan (2018) revisited the marshmallow test using a sample far larger and more diverse than the original, and they did something the legend never did: they statistically accounted for the children's family background and socioeconomic circumstances.

The result is the heart of this article, and it deserves to be stated carefully. The predictive power of the marshmallow test was much weaker than the legend claimed, and once family background and socioeconomic status were taken into account, the association between a preschooler's patience and later achievement shrank substantially. The willpower of a single four-year-old in a single moment was not the engine of later success it had been sold as. Most of what the marshmallow appeared to predict was better explained by the conditions surrounding the child. This is the caveat that should travel with the marshmallow test every time it is mentioned, and it almost never does: never present it as proof that willpower alone determines a child's future.

Why environment beats willpower here

Once you see the result, the mechanism makes intuitive sense. Waiting for a second marshmallow is not only a test of self-control. It is also a test of whether the child has reason to trust the promise. A child from a stable, resourced home has usually learned that adults who say "wait and you will get more" tend to be telling the truth. Waiting is the rational choice. A child whose experience has taught them that a treat in hand is safer than a promise in the future is behaving sensibly too, given their world. The behaviour reflects the environment as much as any inner trait.

That reframing changes the parenting implication completely. If the marshmallow measured destiny-grade willpower, the task would be to drill self-denial into a toddler. But if much of what it captured was the stability and trustworthiness of a child's environment, the lever is somewhere else entirely. You build the conditions - predictability, kept promises, a calm and reliable home - and the patience tends to follow as a reasonable response to a trustworthy world. This is the same theme that runs through everything in parenting and child success: the environment is not the backdrop to the trait, it is much of the cause.

What still holds about self-control

It would be an overcorrection to conclude that self-control is a myth. It is not. The honest position is narrower than the legend and more robust than the backlash. Self-control, measured well and across childhood rather than in one tense moment with a marshmallow, does carry real predictive weight.

The strongest evidence comes from outside the marshmallow paradigm. Moffitt et al. (2011), following about 1,000 people in Dunedin to age 32, found that childhood self-control predicted adult health, personal finances, and fewer legal problems - and that the gradient held even after accounting for intelligence and social class. The contrast with the marshmallow story is instructive. Moffitt's measure was a sustained assessment of self-control across a child's early life, not a single snapshot, and it was tested in a large, representative cohort. Self-control matters. What does not hold is the specific marshmallow legend: that one moment of toddler willpower, divorced from environment, writes the future. The trait is real; the parable oversold it. The same nuance applies when patience is linked to money, which I take up in delayed gratification and wealth.

This article is for informational and educational purposes only and does not constitute financial, legal, tax, medical, or professional advice. Individual results vary.