In 1972 and 1973, a group of researchers in New Zealand began recruiting participants for what would become one of the most valuable long-term studies in developmental psychology. They enrolled every child born in the town of Dunedin during a twelve-month window. There were 1,037 of them. Over the next four decades, they’d follow those children through repeated assessments at ages 3, 5, 7, 9, 11, 13, 15, 18, 21, 26, and 32. They would measure their intelligence, their family circumstances, their health, their school performance, their finances, their relationships, and, when relevant, their criminal records. And they’d ask a question most parents would find uncomfortable to think about clearly. Can you tell, from a child at age three, what kind of adult that child will become?
In 2011, when the participants were 32, the research team led by Terrie Moffitt and Avshalom Caspi of Duke University and King’s College London published their most striking finding in the Proceedings of the National Academy of Sciences. They had run the analysis on one trait measured across the childhood assessments. The trait was self-control, defined as the child’s ability to regulate their impulses, sustain attention, delay gratification, and persist through frustration. It was assessed through observer ratings, parent and teacher reports, and behavioural tests given when the children were 3, 5, 7, 9, and 11.
The finding was that children with low self-control in the first decade of life were, at 32, more likely to have physical health problems, measurable through blood tests and medical exams. More likely to be dependent on tobacco, alcohol, cannabis, or harder drugs. More likely to have serious financial difficulties, including missed payments, bankruptcy, and trouble with basic money management. More likely to be raising a child in a single-parent household. And more likely to have a criminal conviction on their record.
The Artful Age
A weekly letter on aging well, family across generations, and the creative life after the kids leave home.
If you find this article interesting, you might also enjoy this video below. What happens when an ordinary history teacher turns his class into an experiment to explain how the Nazi regime worked in Germany?
Chaos.
None of this, on its own, would have been remarkable. Poor self-control in childhood predicts these outcomes across many studies. What made the Dunedin finding land differently was what happened when the researchers controlled for the two variables most people would assume were doing the work. They adjusted for IQ. They adjusted for family social class. And the self-control effect remained. Children born into families of higher social class had, in general, better adult outcomes. Children with higher IQs had, in general, better adult outcomes. But within any tier of IQ and any tier of social class, the low-self-control children were doing worse at 32 than their higher-self-control peers from the same background. Self-control was doing something the other variables weren’t picking up.
Which brings up the difficult question the paper raised. If self-control at age 3 is doing this much predictive work, is it fixed? Are some children simply born with the neurological equipment to regulate their impulses well, and others without it, and is the rest of their life shaped by which side they landed on?
The researchers looked at this directly. They found that within the cohort, some children’s self-control improved substantially over the decade of childhood assessments. Others’ declined. Both movements mattered. Children whose self-control rose during the study period had better adult outcomes than children whose self-control stayed low, even when the starting points were similar. The trait was heritable, in the sense that some children came in with more of it than others. But it was also, on the current evidence, movable.
Subsequent research has extended this. In a 2016 meta-analysis published in the Journal of Experimental Criminology, researchers Alex Piquero, Wesley Jennings, David Farrington and colleagues pooled the results of dozens of evaluated programs designed to improve children’s self-control. The average effect size on self-control was 0.32, which is meaningful in the social sciences. The average effect on delinquency, a downstream outcome the programs were also trying to affect, was 0.27. Both effects were statistically robust across the studies included. Self-control, when targeted through structured intervention with young children, could be raised. And when it was, the downstream benefits followed.
Broader work has replicated this pattern. In a 2018 systematic review and meta-analysis in JAMA Pediatrics by Anuja Pandey and colleagues at University College London, researchers examined 49 randomised trials of self-regulation interventions in children and adolescents. The finding was that self-regulation could be reliably improved through intervention across both age groups, with positive downstream effects on academic achievement, social skills, mental health, behavioural problems, and substance use. The window for building this capacity, on the current research, is wider than the Dunedin finding might initially suggest. Age 3 isn’t a deadline. It’s the earliest point at which the trait is measurable well enough to predict adult outcomes with the level of confidence Moffitt’s team was able to achieve.
Which is the piece worth sitting with as a parent, a teacher, or a policy-maker. The Dunedin study wasn’t saying a child’s fate is sealed at age 3. It was saying that a particular psychological capacity, present in some children more than others by that age, does an unusual amount of work in shaping how the rest of the life unfolds. If a child has it in abundance, their other advantages compound. If they don’t, their other advantages, including a high IQ and a comfortable family background, are less protective than most people would assume. And on the intervention research, the capacity itself can be built.
The thing that lands hardest about this finding, for me, isn’t the predictive power. It’s the mechanism the study implies. What self-control seems to be doing across the decades isn’t magical. It’s the accumulated effect of thousands of small daily choices in which one option produces an immediate reward and the other produces a delayed one. The child who can consistently choose the delayed option builds, over time, a life in which the delayed options add up. Better health, because they made the harder choice on food and rest thousands of times. Better finances, because they resisted the impulse to spend often enough. Fewer entanglements with the law, because they walked away from provocation when a less regulated child wouldn’t have.
Which sounds obvious said all at once, but the study is showing something more targeted than an obvious moral. It’s showing that this capacity, this ability to make the harder choice a little more often than not, is measurable in a three-year-old, is predictive across decades, and is trainable if we choose to train it. Most cultures, in their public policy, don’t. What the Dunedin data suggests, without much fanfare for a piece of longitudinal research of this magnitude, is that this might be where a serious intervention would produce the largest downstream return.
By Kiran Athar. I write about psychology and human development for a general audience. I am not a psychologist, therapist, or clinician. What follows is my reading of the research on childhood self-control and adult outcomes, offered as an interpretation.