Self-control in childhood predicts health and wealth in adulthood

By Ed Yong | January 24, 2011 3:00 pm

“Control! Control! You must have control.” – Yoda

Pay attention. Put that down. Stop doing that. Eat that later. Would you, just, behave? These phrases are a familiar part of family life, as parents try to drum a sense of self-control into their children. Right from the start, they are taught to restrain their impulses, focus on their goals, and control their choices. This seems like a wise move, but how could you tell if such instruction actually affects a child’s fate?

Ideally, you would follow a group of children into adulthood, to see how their degree of self-control affects the course of their lives. You’d need to catch up with them at regular intervals to look at their health, mental state, finances and more. You’d need to meticulously plan the study decades before the important results came in, and you’d need to keep in close touch with the volunteers so they stick with the study. In short, you’d need to have set up the Dunedin Multidisciplinary Health and Development Study.

The Dunedin Study was the brainchild of Phil Silva, and its wide-ranging team include  Terrie Moffit and Avshalom Caspi, a husband and wife duo who work at Duke University and King’s College London. The study began way back in 1975, with 1037 children who were born in Dunedin, New Zealand between April 1972 and March 1973. The researchers became their occasional companions through most of their lives, up till the age of 32. At 11 separate points, Moffit and Caspi measured the recruits’ health, wealth and more. And amazingly for a study of this sort, 1014 of the children are still alive and involved.

Thanks to their unique study, Moffit and Caspi have found that children who show high levels of self-control within their first decade of life do better in adulthood. Even after accounting for things like intelligence and social class, those who had a tighter grip on their behaviour as children are now in better health as adults. They’re also less likely to be abusing drugs, have a criminal record, or suffer from financial problems.

Moffit and Caspi assessed the children’s self-control at the ages of 3, 5, 7, 9 and 11, by looking at their hyperactivity, attention, impulsiveness, aggression, and more. The children did the same evaluations, as did their parents and teachers. All the scores were a good fit for one another, so Moffit and Caspi combined these numbers into a single measure of childhood self-control. Finally, they adjusted the value according to the volunteers’ family background and IQ.

This childhood score proved to be a portent of things to come. Children with lower scores (poorer self-control) had poorer health at the age of 32. Their lungs didn’t perform as well. They were more likely to have gum disease, be overweight, or depend on drugs like tobacco, alcohol or cannabis. Among those with the highest levels of self-control, 11% had multiple health problems, compared to 27% of those with the lowest levels (see graph below).

Those with poorer self-control were also more likely to run into financial or social problems. As teenagers, they were more likely to start smoking, leave school with no qualifications, or have unplanned pregnancies. As adults, they had more credit problems and troubles with money, and fewer tangible assets like a home, savings or a pension. They were more likely to have been convicted of a crime, and their own children were more likely to be raised in a single-parent household. And in fact, their childhood self-control was a better predictor of these financial worries than either their IQs or social backgrounds.

Of course, these are just correlations. Moffit and Caspi did their best to account for differences in background and IQ. Even so, the thousand children still came from families with many other differences that could have affected the course of their lives, including parenting style and home environment.

To deal with that, the duo turned to a second group of children – 509 pairs of British twins who they have followed since their birth in 1994-95. Each pair of twins has lived with the same parents in the same house. And despite their shared environments, the twin with poorer self-control at the age of five was still more likely to start smoking at the age of 12. Their behaviour was more antisocial and their school grades were worse than their more restrained siblings.

Moffit and Caspi think that this research has big implications. If poor self-control underlies a lot of important societal problems, such as drug abuse, crime and poverty, you could do a lot of good by boosting this skill during childhood or adolescence. Many other things correlate with these social problems, including IQ and social background, but few of these are as malleable as self-control.

For a start, Moffit and Caspi say that their results support the use of “opt-out programs”, which aim to steer people’s decisions by making the “right” choices easier. People could be automatically enrolled onto organ-donor programs, with the option to opt out if they wish. Canteens and supermarkets could put healthy options front and centre, instead of the sweets and snacks that currently line the check-outs. Specific amounts could be automatically deducted from their salaries and transferred into savings, unless they say otherwise.

These strategies will be familiar to anyone who has been following the UK government’s fascination with the “nudge” theory, proposed by economists Richard Thaler and Cass Sunstein. They are effectively appeals to laziness. The idea is that people with poor self-control should take the easy option, because opting out requires them to expend effort and planning.

But I’m not convinced of the value of the long-term value of these initiatives. They play to poor self-control, rather than attempting to improve upon it. And improvement is important. In their study, Moffit and Caspi found that the more self-control people had as children, the better their futures, even for those at the high end of the scale. They also found that children who developed better self-control as they grew up fared better than those who stayed at the same level.

The good news is that there are plenty of more direct measures. In 2009, Alex Piquero published an overview of 34 clinical trials that tried to boost children’s self-control. The trials used a variety of methods including role-playing using video-taped scenarios, rewarding children for good behaviour and relaxation training, including meditation and deep breathing techniques. Some involved a complex program of parent support groups, social skills training, academic tutors, peer buddies and more. These diverse methods had similar goals: teaching children to monitor and control their own emotions, solve social problems and build solid friendships.

Despite the variety of techniques, Piquero found optimistic results. Overall, the programs he analysed were largely successful in improving self-control and warding off delinquent or problematic behaviour. Moffit and Caspi’s work suggests that such improvements will cascade throughout a child’s adult life. However, Moffit warns that many of the studies that Piquero analysed had high drop-out rates and many weren’t properly blinded.

“The decisive answer is not in yet, and more programmes should be designed, and evaluated rigorously,” she says. “The essential part will be to prove that there is a return on investment, a good cost-benefit ratio. All this research is going to require sustained willpower on the part of scientists and funders, and we hoped our paper would give them encouragement to do the necessary work. Our paper does not say “nothing can be done” it says “get going and do something.”

Reference: Moffitt, Arseneault, Belsk, Dickson, Hancox, Harrington, Houts, Poulton, Roberts, Ross, Sears, Thomson & Caspi. A gradient of childhood self-control predicts health, wealth, and public safety. PNAS

Image by Quinnaya

Correction: Originally, the article implied that Moffitt and Caspi started the Dunedin Study in the 1970s. The study’s founder was actually Phil Silva.

More on child development:


Comments (24)

  1. Andrea Kuszewski

    I love, love, love studies like this. The best part: other research shows that cognitive control (or as referred to in this study, ‘self-control’) *can* be improved with training—the best results gained the earlier in life the training begins. So this means some of these outcomes can change in time, if we are committed to early intervention.

    I have long believed that many psychological conditions manifest out of poor control over genetic traits. Gain better control—the severity of symptoms decrease, and you are able to make those traits work to your advantage. This is similar to finding strategies that help with focus for a child with ADHD; the influx of info is still coming in at lightning speed (the trait), but you learn strategies to recognize the relevant info & focus on that (practiced control). New data is supporting this idea and approach more and more, which means many good things in the future for preventative treatment of psychological conditions.

    Such a great piece, Ed.

  2. becca

    I am really happy you paired this report with data that they are finding ways to increase self-control. I do wonder about context-specificity of self control- it seems it would be easier to invent a test, train people, and have them get better. But much harder to invent a general battery of tests, train people, and have them do better in life. How close are we to that?
    As a kid, I was always very poor at controlling my temper, but very good at controlling my attention (at least in some ways- mostly the art training for ability to focus on tiny details for many hours). Now I fear I’ve just lost the later ability, due to being constantly able to seek and get new input (the internet is horrible for that).
    Anyway, I was recently talking to someone- I can’t remember the context now- and I said something about executive function being very important, and that being what I was struggling with. And they replied with something like “well yes, but you can’t just fix that”… I was stunned. You mean, I said, I can’t *will* myself to having more willpower? (darn it, there’s a *term* for this… the specific variety of catch-22 where you could get what you need if only you already had it… I wanna say resource trap but that is googling up odd references. curse you internet and my lack of memory!)
    But anyway, maybe ther is hope… maybe I can undertake a strenuous willpower exercise regime. That fits much better with the Puritan ideals that hard work is itself a moral virtue, right? What would one need to do (as an adult or to help a child) to increase self-control?

  3. Andrea Kuszewski

    Here’s a simple solution: All school systems should agree to mandatory Jedi training beginning in Kindergarten. So easy. 😉

    But seriously, Becca: There *is* research out there on improvements in cognitive control (executive function, self control) via specific training, transferring to general improvements in overall control—but the names of specific studies escape me at the moment. There was a lot of work done on the mechanism of cognitive control by Jon Cohen at Princeton. You might want to start there; I bet you’ll find some answers.

    The person who told you that cognitive control can’t be improved—meaning no disrespect— that person was just wrong.

  4. David Manly

    Great article, great research, great comments.

    I like it when Andrea wrote, “I have long believed that many psychological conditions manifest out of poor control over genetic traits.”
    I am an identical twin, and my brother and I have been raised (almost) identically with similar values instilled in us by our parents, older sibling and family. And yet, we differ very much in terms of temperament.
    My brother has poorer attention, and exhibits many of the same characteristics as discussed in the British twin study above. That is not to say that we are so different, but through perseverance and training I was able to change my behaviour and self-control.

    Good show, Ed.

    Oh, and mandatory Jedi training? I’m SO there.

  5. Åse

    OOOooooh Jedi training! Would appeal to my 7 year old.

  6. annoying nerd

    it’s “Control! Control! You must learn control!”

    ironically and inauspiciously, i was not able to control myself from providing a correction for Yoda’s imperative.

  7. @Andrea – this post is dedicated to our #scio11 chat about striving to go beyond the paper.

    @Becca – the meta-analysis I linked to involve a wide variety of different but successful techniques, which suggests that there’s at least the potential for non-specific programmes. I think Andrea’s right – even a cursory look at the evidence base revealed several positive interventions, at least in children. It’s apparently harder after that, but this might either be because (a) it’s harder, or (b) the evidence base is sparser. If you want to start having a look, check out the review for a start.

    @Annoying Nerd – Sigh. You are in fact the *second* person to point this out to me. Fixed, wearily 😉

  8. @Becca – a response from Moffit:

    Our report focussed on childhood and adolescence as good times to train self-control, but we can’t rule out adulthood. In an 2003 publication from our team, we found that a small group of Dunedin Study members who got highly responsible jobs in their 20’s showed significant increase in their self-control skills thereafter. Of course, these young adults were conscientious to begin with, which helped them to get those supervisory jobs, but they showed even more improvement in self control across their twenties. So, practice makes perfect?

  9. Kate

    Great article. Echoing becca, nice to see the pairing of the research with data about interventions that increase self control. On that front, I recently read a chapter in NurtureShock ( about teaching self control. It mostly covers Tools of the Mind – an early childhood programme in the US delivering impressive results:

  10. el_lur

    Curious: While text(s) assumes that the picture of health is a self training by Yoda, the study outlined indicates that the key variable is a collective.

  11. As a parent, I find this study validating. It’s a relief that all those threats and bribes are not in vain. :)

  12. Ian

    Very interesting piece of research – and dare I say refreshing too. Tell me, do you see it contradicting research by the likes of Susan Blakemore (and others) who believe there is no such thing as free will?

    I’m also interested to know why government policies in treating alcoholism, crime, even reducing STDs and teenage pregnancy do not promote the exercise of self-control.

    I appreciate this is a broad set of questions but being a person of faith, I see this research supporting the role of faith in maintaining a healthy lifestyle … I’m not sure about the ‘wealth’ side of it, but surely health and happiness.

  13. Aurora

    Great piece, I liked how you did more than ‘just’ report on the paper.

  14. yogi-one

    Very, very good work by Moffit and Caspi. It would be great to see more similar studies, although, as you rightfully point out, the time-scale and numerous factors affectings children’s development make it a very big, long (and necessarily) expensove kind of study that only shows real results after decades.

    More could be done, perhaps, by working backwards. This would mean finding people, particularly pairs of genetic and fraternal twins, whose lives could be well documented (this is easier nowadays since we have all nature of recording devices, video cams, pix, and everybody who works with children (schools, doctors, etc) is required to keep records that note any developmental or behavioral issues with the kids.

    The Jedi training idea is really a good concept. Kids would love it and it would give the parents and teachers a great platform to bring out the qualities we love in a Jedi – honesty, discipline, practice, problem-solving, a sense of adventure and excitement about life, all tempered with the spiritual values of compassion, respect for others, and meditation.

    This is fantastic work, and I’m thinking there must be ways to build on it without having to wait 30 more years for someone to complete another birth-to-full-adulthood study. Of course, such studies should be done, just saying we shouldn’t wait another whole generation to get started using the benefits of Moffit and Caspi’s work.

  15. I think this paper is good, and it certainly the largest set of correlations available, but I actually wonder how much it adds to certain aspects of the literature like drug addiction. They referred to it as “self-control” here, but in the drug addiction literature this is often measured as “impulsivity” using the same tests. Drug addiction studies have been focusing for several years on traits like impulsivity and find the same correlations with criminal records and drug use. This IS the first study to look at actual physical health, though, and I think that particularly interesting.

  16. @Scicurious – “Impulsivity” is one of nine separate components used to calculate the “self-control” score. Also Moffit mentions other existing correlational studies, but I think the value of the long-term, prospective study adds to those.

    @yogi-one – You wouldn’t really need another 30-year study. If you can show that interventions improve self-control at an early age, you can use this data set to infer that those interventions would affect adult outcomes. “Backwards” studies – formally, retrospective ones – are plagued with problems of their own, including issues with recruiting two matching groups of people, and relying on their potentially faulty memories. See:

    @Aurora – Thanks! Was making a specific effort to do that.

  17. Andrea Kuszewski

    @ Sci: This paper recently came out and addresses the issue/problem of “impulsivity” being used as a specific defining trait in a number of separate disorders:

    “Borderline Personality Disorder, Bulimia Nervosa, Antisocial Personality Disorder, ADHD, Substance Use: Common Threads, Common Treatment Needs, and the Nature of Impulsivity”

  18. No Self Control

    First remember reading about the link between self-control and success via Mischel’s marshmallow test:

  19. I found this study (and the general approbation of its findings) a little disturbing. Isn’t ‘self-control’ merely social conformity – choosing to do what others ask of you? It’s no surprise then that children who chose to conform to the expectations of those around them should subsequently score highly on measures deemed ‘desirable’ by society (e.g. a responsible well-paying job, maintaining marriage due to social expectations, avoiding ‘risky’ behaviours, etc.). I can’t help but wonder whether the group of children with the least ‘self-control’ didn’t also include some of our most valued members of society (e.g. inventors, social reformists, scientists, artists, etc.). Aren’t there two-sides to the conformity coin?

  20. nancy brownlee

    Oh bravo, Lynda. It’s also a little disturbing that you’re the first to comment on the assumption of values implicit in the study.

  21. @Lynda – you’re conflating self-control with independent thought. The two are not the same thing. Self-control also includes things like resisting the temptation to overeat, or curbing an aggressive impulse. It’s about basic cognitive skills like holding one task in mind while parsing out irrelevant info. It has little to do with social conformity.

  22. @21
    Having now read the paper I agree that they weren’t measuring social conformity. Also – rather than independent thought – I was referring to a willingness to pursue one’s own desires, interests and values in the face of social disapproval.
    Nevertheless, I still found the paper disturbing. They refer to ‘self control’ as an umbrella concept, but it seems little more than a ragbag for socially inconvenient behaviours. The nine traits they used to quantify ‘self control’ (and note that the children weren’t tested experimentally but were rated by an observer, or a teacher, parent and self) were: lability, low frustration tolerance, hostility, roughness, resistance, restlessness, impulsivity, fleeting attention and lack of persistence. I don’t question the validity of research into impulsivity or aggression or attention span or frustration tolerance, but to lump all these very different things together and sanctimoniously label them ‘self control’ seems highly questionable.

  23. Matt B.

    Okay, #2 Becca has made my brain itch. I have to know what that term is for needing to have something in order to get it.

    I’ve thought of “vicious cycle”, “catch-22”, “what’s good for the goose is good for the gander”, “self-fulfilling prophecy” and “damned if you do, damned if you don’t”. None of those is it.

  24. Matt B.

    It’s close to the relevance paradox, wherein someone doesn’t seek information because, without having the information, he doesn’t realize it would help.

    Like a safe that contains its only key.


Discover's Newsletter

Sign up to get the latest science news delivered weekly right to your inbox!

Not Exactly Rocket Science

Dive into the awe-inspiring, beautiful and quirky world of science news with award-winning writer Ed Yong. No previous experience required.

See More

Collapse bottom bar