Four Decade Long Dunedin Study Shows How to Predict Future Social Costs of Individuals
Print Friendly and PDF

From The Guardian:

‘High social cost’ adults can be predicted from as young as three, says study

20% of population uses majority of public services, research shows, indicating long-term importance of early years investment for disadvantaged children

By looking at socioeconomic background, experience of maltreatment, IQ and self-control it was possible to predict which children would grow up to become part of this “high- cost” segment of society, the study found.

Nicola Davis Monday 12 December 2016 11.00 EST

A small fraction of the population is likely to account for the majority of societal costs, according to new research into the impact of childhood disadvantage on later life.

The research – based on New Zealand data but involving an international team – also revealed that such “high cost” adults can be predicted when as young as three years old from an assessment of their brain health.

The study, which followed around 1,000 children from birth, found that at 38 years of age just 22% of the group accounted for 81% of its criminal convictions, 78% of pharmaceutical prescriptions, and 66% of welfare benefits.

“About 20% of the population is using the lion’s share of a wide array of public services and we found that the same people use most of the national health service, the criminal courts, insurance claims for disabling injury, pharmaceutical prescriptions and social welfare benefits,” said Terrie Moffitt, co-author of the research from Duke University, North Carolina.

But she cautioned that the results should elicit compassion rather than being used to stigmatise individuals.

“Very often when we think of the people who are the greatest burden on society we can jump to the conclusion these are lazy layabouts who are happy to live off the public purse,” she said. “But in fact what this research suggests is, yes, there is a very high cost fragment of society, but these are people who weren’t very well prepared as pre-schoolers for making their way into the really modern, fast-paced, higher technical, education-dependant job markets.”

Josh Hillman, director of education at the Nuffield Foundation, who was not involved in the research, agreed, adding that while the study was based on children in New Zealand, the findings are likely to apply in other developed countries, including the UK.

“The implications are that we have really underestimated the long-term benefits of investment in early years education for disadvantaged children, both in terms of the benefits for the children themselves but also in terms of the payback for the public purse,” he said.

Writing in the journal Nature Human Behaviour, the team from the US, UK and New Zealand, described how they analysed data collected from 1,037 children born in Dunedin, New Zealand, who took part in what is known as the Dunedin Longitudinal Study.

Around 95% of the children were followed up until the age of 38, when data was collected from personal interviews and a host of national administrative databases, allowing the team the rare chance to explore connections between factors in childhood and a host of outcomes in adulthood.

The results reveal that 20% of the group accounted for the majority share for each of the different resources or services considered, ranging from criminal convictions to welfare benefits and cigarettes smoked.

But they also found that many individuals cropped up as heavy users of multiple services. “They did tend to be the same individuals who were showing up again and again and again in each of the different government databases,” said Moffitt.

Indeed, analysis of heavy users of three of more services revealed that the same 22% of the cohort accounted for 81% of the group’s criminal convictions, 77% of fatherless children, 36% of injury insurance claims, 78% of prescriptions, 66% of welfare benefits and 40% of excess obese kilograms, as well as more than half of cigarettes smoked and nights spent in hospital.

It was possible to predict which of the children were most likely to grow up to become part of this high cost segment of society from measures of their socioeconomic background, experience of maltreatment, IQ and self-control, which were taken repeatedly from birth to age 11.

But the team also discovered that a rating of “brain health”, based on the combined results from a 45 minute-long assessment of motor skills, understanding of language, social behaviour and IQ at three years of age, was almost as accurate a predictive tool.

“Given two pre-schoolers about 80% of the time we can accurately predict which group they will end up in,” said Avshalom Caspi, another author of the research, also from Duke University.

Commenter res writes:

Thanks for that The Times link!

The full paper behind that The Times article is available at

Childhood forecasting of a small segment of the population with large economic burden

Here is an iSteve article from two years ago about the Dunedin Study:

And here is a link to James Thompson’s blog (mentioned by Steve) calling out an early keynote speech about this paper:

Hopefully Dr. Thompson and/or Steve will comment on the current paper.

More about the Dunedin Study and its members:

Regarding the paper:

I thought Figure 4 did a good job of showing the dramatic cost differences between the high cost (22% of subjects) and the low cost (30%) groups.

Screenshot 2016-12-13 15.07.05

There is more detail in the supplementary materials PDF including this:

The dataset reported in the current article is not publicly available due to lack of informed consent and ethical approval, but is available from the corresponding author on reasonable request by qualified scientists.
Page 9 of the SM had an interesting table of social cost category burden by sex. Not surprising that men dominated (~75%) crime and injury claims, but I was a bit surprised by how much women dominated hospital stays and prescription fills (by 71/63%).

Page 7 of the SM has details of the age 3 test(s) they used.

As res says, back in 2014 I blogged about an early version of this Dunedin study, which I had read about on James Thompson’s Psychological Comments. I wrote two years ago:

The Dunedin Study: Nature v. Nurture Across 40 Years


Whenever I read denunciations of The Bell Curve, I’m struck by how little subsequent data are cited. We hear a lot of a priori arguments that were musty when the late Stephen Jay Gould was trumpeting them and a lot of ad hominem anger, but few references to new data that have emerged in the 20 years since.

It’s not as if there isn’t a lot more data today.

What made The Bell Curve of particular importance was that around 1990 Murray and Herrnstein got their hands on a particularly rich data set: the federal 1979 National Longitudinal Study of Youth tracking of 12,000+ nationally representative youngish people combined with the Pentagon’s 1980 administration of the AFQT/ASVAB enlistment test to the NLSY sample.…

So, if Herrnstein and Murray are wrong, we now have 20 more years of data from the NLSY79 tracking, which now includes thousands of children of females in the original study, including both mother and child IQ-like scores. … Plus we have the NLSY97 tracking study from 18 years later that has now been running for 17 years.

And we have lots of other long-term tracking samples, such as ADD Health.

Overseas, there is the Dunedin sample

The medical and dental school in the New Zealand city of Dunedin enrolled virtually every child born in Dunedin over a 12-month period in 1972-73 into this lifelong study. The subjects are now in their early 40s and their children are being enrolled as they reach age 15. A sizable documentary is being prepared on the results called “The Science of Us.” (Above is a brief trailer.)

Here is a link to the 44 minute documentary.

And here’s the home page of psychologists Avshalom Caspi and Terrie Moffitt who have written some of the most ambitious papers based on the Dunedin population.

At James Thompson’s Psychological Comments, he summarizes an upcoming Caspi and Moffitt paper based on 40 years of Dunedin data where they look at who are net contributors and who are net consumers of the common weal.

This strikes me as potentially hugely useful in immigration policy. My opinion has been that our immigration system ought to try to exclude individuals (and thus their descendants) who are likely to cost far more than they pay in taxes. Here’s an upcoming study of a first world city over the last 40 years. Dr. Thompson wrote in response to being given a preview of the big Dunedin Study:

Are you a nuisance?

What if we were to take an objective measure? Track a thousand newborns, and keep a close account of the profit and loss ledger. At this point you may feel a trifle uneasy. Who are we to judge these matters? What price the jocular remark of a mute inglorious Milton? How could one possibly assess the wit of someone who lacks a Twitter account?

Furthermore, you may recoil at the possible results of such an enquiry. If some individuals turn out to be a nuisance and a high cost to society, what then? Should they be exiled to some other land whether the natives are even more generous and gullible, or should we intervene as best we can to make them into productive citizens? …

When you find yourself in a hole, stop digging. Taking care of your own citizenry’s problem children is one thing, taking care of other peoples’ is something else.

The big problem with this kind of research has been that it comes up with pretty much the same conclusions as social sciences had before their findings got memory-holed. The old learning is ably summed up in the 1985 book Crime and Human Nature by James Q. Wilson and Richard Herrnstein. But perhaps we are getting so far down the road that the people like Stephen Jay Gould who read the old scientists out of polite society are dying off and thus the new researchers have a little more freedom to present their findings in the mainstream media?

[Comment at]

Print Friendly and PDF