No, Scientists Cannot Predict Whether Toddlers Will Become Criminals

IMAGE: Getty Images
IMAGE: Getty Images

Earlier this month, a new study came out suggesting that it’s possible to predict whether a toddler will become a criminal. Based on neurological exams, scientists correlated the brain health of people at age three with whether they went on to commit any crimes as adults. And for those with poor brain health, 80 percent of the time, it turned out that they did.

The study nabbed headlines around the world suggesting that we’d arrived at a “Minority Report” future in which a person could be condemned based on a 45-minute brain test alone. Cue the outrage.

The only trouble is, that’s not actually what the study found.

The headlines in question come to us via the Dunedin Study, a massive research effort that began in New Zealand in the 1970s to follow 1,000 people beginning at birth. Over the past four decades, the study has netted many important findings, among them revelations about why people age differently, what leads to substance abuse and how early childhood experiences shape adult life. In the most recent research, the study’s authors were looking to test the idea that a small percentage of the population accounts for a large percentage of social service recipients, and then figure out whether there might be early identifiers to help pinpoint segments of the population most in need of preventative services.


Researchers dug through decades of health and government records for the study’s 1,000 participants, finding that within the cohort, 80 percent of economic burden was attributable to just 20 percent of its members. That small slice accounted for 81 percent of the group’s criminal convictions, 66 percent of welfare benefits, 78 percent of prescription fillings and 40 percent of those considered obese. Researchers also looked at the results of a pediatric examination each study member had done at age three, back in the 1970s. The examination included a neurological evaluation assessing verbal comprehension, language development, motor skills and social behavior. Poor scores on that test, they found, were a good indication that a person would wind up a member of the “high cost” group. In fact, by looking at those scores, they could guess who would end up in which group about 80 percent of the time.

The end goal, study director Richie Poulton told me earlier this week, is to help health and education providers better target their services at those most in need.

“We have all of this education and preventative programming,” he said. “But so often those services do not get to the people who need them the most.”


But online, this research got boiled down, essentially, to one click-worthy pull quote: “Future criminals revealed at three,” as one headline put it. Poulton said the headlines were “unfortunate,” distracting from the findings with Orwellian fear-mongering.

The researchers were conscious of how their results might be interpreted. In their paper, published in Nature Human Behavior under the much less buzzworthy heading “Childhood forecasting of a small segment of the population with large economic burden,” researchers wrote that they were “aware of the potential for misusing these findings, for stigmatising and stereotyping.” It’s implications, though, suggested the opposite: “There is no merit in blaming a person for economic burden following from childhood disadvantage,” they wrote.


Fake news, in all its varied iterations, is high on the list of our current moment’s trending topics. But it is an issue that has long plagued science. Studies get twisted or simplified to the point of no longer being true, reduced to a headline that each time it’s retweeted or reblogged gets further and further from the truth. Sometimes pure fiction makes its way into the public consciousness as fact. Every year, the British Medical Journal publishes an issue of joke science, but years later, those papers wind up being cited as real. Earlier this year, when I was reporting on the biotech company Oxitec’s quest to release genetically modified, disease-fighting mosquitoes in the Florida Keys, project scientist Derric Nimmo told me he had taken to going door-to-door in the Keys in an attempt to fight rumor with fact. Rumors included one that the modified mosquitoes could make children sterile and another that the company was actually the cause of Zika’s spread.

The reason this matters is that the science that is permitted to progress is the result of not just technological progress, but also politics and ethics. The headlines heralding that Dunedin Study researchers have created an early-warning system for criminals aren’t just wrong—they could impact, say, whether those researchers get precious funding dollars in the future.


Among scientists, the notion that childhood risk factors can be useful predictors of events later in life is controversial. Dunedin Study researchers sought to use their decades of data to weigh in. This new research is another point in the debate, suggesting that certain childhood risk factors could indeed be useful to public health policy-makers in determining who is most in need of preventative services.

“We sought to inform a question that nags at the behavioral and social sciences, and that has strategic consequences for national policy on children: how strong is the connection between childhood risk and future costly life-course outcomes?” the researchers wrote. “Results reported here suggest that the importance of childhood risks for poor adult outcomes has generally been underestimated.”

A shows the burdens on society of multiple-high-cost users. B shows that in contrast, a substantial segment of the cohort did not belong to any high-cost group. IMAGE: Nature
A shows the burdens on society of multiple-high-cost users. B shows that in contrast, a substantial segment of the cohort did not belong to any high-cost group. IMAGE: Nature

Some of the findings were in line with what other researchers have suggested. They found that members of the high-cost group had the same four childhood disadvantages: they grew up poor, were mistreated, scored low on IQ tests, and exhibited a lack of self-control. But while those numbers led to a statistically significant prediction when looking at the group as a whole, they didn’t help predict what might happen to a certain individual. That’s where the neurological assessments came in. A simple, 45-minute neurological exam, they found, could be an extremely accurate predictor of whether an individual would heavily rely on social and health services later in life. This could, in theory, help providers target specific individuals from a young age with services to help them stay on the straight and narrow.


“Ameliorating the effects of childhood disadvantage is an important aim,” the researchers wrote, “and achieving this through early-years support for families and children could benefit all members of a society.”

The research of course still needs to be replicated and put through real world tests to see whether identifying such people early on might have a positive effect. In the meantime, though, turning science fact into science fiction benefits no one.


Senior Writer, Gizmodo.

Share This Story

Get our `newsletter`


This has been a problem with people “interpreting” science without really understanding it. Everyone is guilty to some level of doing this when they step outside their field, but its especially troublesome when people have social currency and use their platform to create “fake science” that supports their personal ideals and desires.

I think that people who don’t understand society’s pressures are more likely to misinterpret such scientific findings - and then to lose trust in the science because it is seemingly telling them something they don’t want to hear. In this case, the study was identifying poor brain health - something I could have fallen under as a child depending on how I was diagnosed. ADHD is certainly one of those issues that reduces learning capacity if you don’t have the resources to treat it. If someone is having a problem learning when very young - it creates a situation where they will not have as many opportunities in life. That lack of opportunity equates to crime, as they seek ways to achieve what their peers achieved. And lets face reality, there is a political class system in place that makes “criminalization” far more common for minorities and those who use drugs.

While this study identifies a learning problem, taking that correlation out of context and distorting the science around it causes more harm than good. But that doesn’t mean that there are not signs of serious issues that are often ignored because some people have fixed social and ideological biases.

What follows is not a scientific observation, but I think its important to note that just because something hasn’t been scientifically proven or disproven, there are things in society that we should recognize as right and wrong. The example I will use is the following: Every kid has killed things like insects and such - usually seeing something injured or hurt causes a person to empathize - and then recoil at what they’ve done. I might be able to kill a stinging/biting insect, but I couldn’t torture it. I certainly couldn’t kill even a mouse or rat by intention.

But show me the child that tortures small creatures and works their way up to mammals and dogs and cats, and you can bet your bottom dollar that these people will have little problem doing the same thing to human beings - so long as they’re smaller and weaker than they are. Its one thing to say that you can “predict” an activity. But it still amazes me how blindly society can discard obvious signs of issues - especially children that fall into this category - and say “Well, you know,its just a dog/cat/pet, it doesn’t mean anything.” Yes... yes it does. I’m not saying these children are evil, I’m saying they need help from society - and if they don’t get it, they have a very good chance of becoming evil. And all the superiority complexes and religious idealism isn’t going to change the rot at the core if it goes untreated. They might never kill a fellow human being, but they might be able to do worse and have it go unnoticed. And before you say that there’s nothing worse than murder - then you might want to check your concept of good and evil - because you don’t have to kill a person to destroy them and everything around them.