Criminal trajectories: Antisocial behavior across the lifespan

What makes a young person commit criminal acts? By the time Terrie Moffitt, PhD, began her research career in the 1980s, scientists had reported a long list of risk factors that were associated with antisocial behavior and delinquency. “But the effect sizes of each of them were quite small,” recalled Moffitt, now a professor of psychology and neuroscience at Duke University and a professor of social development at King’s College London. Through her research, she acquired the insight that there were two different causal processes going on—"and they were washing each other out,” she said. “If you separated participants into two groups, then, in fact, causal factors in one group would be quite big.”

According to this dual taxonomy model, the behaviors of these two groups had distinct causes and trajectories. Factors such as a family history of substance abuse notably increased the odds of LCP behaviors, for example. And although interaction with delinquent peers was likely to encourage antisocial behavior in AL offenders, those in the LCP category were apt to act in delinquent ways regardless of what their friends were doing.

Moffitt’s model generated findings that have informed interventions and influenced education, public policy, and the law. Along the way, her work has inspired psychologists, criminologists, and other social scientists worldwide. “The paper was pivotal—in fact, I still have my original copy from 1993,” said Alex Piquero, PhD, a professor of sociology and criminology at the University of Miami. “It really changed a way of thinking for a lot of criminologists.”

Moffitt, T. E. (1993). Adolescence-limited and life-course-persistent antisocial behavior: A developmental taxonomy. Psychological Review, 100(4), 674–701.

By the 1970s, criminologists recognized that most crimes, and the vast majority of serious offenses, were committed by a small fraction of habitual offenders. With her paper, Moffitt opened a door to understanding those lifelong offenders as well as the larger proportion of troubled youth who emerge as law-abiding citizens on the other side.

Fresh out of graduate school in 1985, Moffitt had spent 2 years in New Zealand working on the longitudinal Dunedin study, when the participants were 13 years old. She collected police records and interviewed the youth about their self-reported delinquent behaviors. As she followed their exploits through adolescence and early adulthood, she was also reflecting on her own experiences. “I had gone to quite a rough high school where most of us engaged in some kind of law-breaking. I’d started to notice that some of my friends who didn’t have the cognitive abilities to get a further education seemed to be trapped,” she said.

She hypothesized that individuals with executive functioning problems and other cognitive deficits would be more likely to persist in crime, while those with stronger cognitive abilities would back off from delinquent behavior as they reached adulthood. Those personal and professional observations led to her first influential paper on the topic. First she submitted the paper to a leading criminology journal, but they asked her to cut its length in half—effectively removing much of the content most relevant to psychologists. Moffitt recognized that the theory had important implications for both criminology and developmental psychology. Though she did trim the paper a bit, she ended up publishing it in Psychological Review, where she was able to include enough details to be useful to scientists in both disciplines. “At the time these were two quite separate fields. I tried to write an integrative paper that pointed out how each field could help each other,” she said.

That effort paid off, as Moffitt’s theory led to important new work in criminology and in psychology. “It was exciting because her paper laid out a really nice roadmap for people to test empirically,” Piquero said. “A fun part of being a social scientist is that you see research questions everywhere—and her paper just handed them out on a platter.”

Moffit’s initial paper described two groups of antisocial adolescents: AL offenders who outgrow antisocial behaviors and LCP offenders who continue to commit crimes at high rates into adulthood. As researchers set about testing her theory, they started to find other possible patterns of deviant behavior that didn’t necessarily fit neatly into those two groups. Longitudinal research from Piquero helped clarify those findings.

Along with psychologists Edward Mulvey, PhD, at the University of Pittsburgh School of Medicine, Larry Steinberg, PhD, at Temple University, and other colleagues, Piquero helped develop the Pathways to Desistance study, which recruited 1,354 14- to 18-year-olds in Philadelphia and Phoenix who had been found guilty of committing serious crimes and followed them for 3 years. Using advanced methods to identify distinct trajectories, they supported many of Moffitt’s predictions about AL and LCP offenders and also described additional trajectories beyond those two broad categories (Developmental Psychopathology, Vol. 22, No. 2, 2010opens in new window).

That research showed two groups of youth with high levels of self-reported offending at the study’s baseline. One group showed a marked decline in self-reported criminal behavior over time, while the other continued to offend at significant rates—patterns that seemed to align with Moffitt’s AL and LCP predictions. But there was also a group of chronic offenders who reported moderate levels of offending that continued steadily across the 36-month study period. “Our research suggested there were two groups of chronic offenders, one very high-rate and one lower rate, but which both continued to offend through a good portion of the life course,” Piquero said.

Other researchers were also accumulating evidence for those lower-level chronic offenders. And as the body of work expanded, Moffitt welcomed the new ideas, Piquero said—an openness that he believes contributed to the model’s ongoing influence. “A lot of people are protective of their theories. She has always been willing to modify aspects of the theory in response to new findings,” he said.

Another important finding from Piquero’s longitudinal study: Almost 60% of participants reported relatively low levels of offending across the 36-month study period. That was especially notable given that all the participants had been arrested for a serious offense. “When we started, the question was: Were kids who get into the system for serious offenses all life-course-persistent offenders?” Piquero said. On the contrary, his work showed most adolescent offenders desist from crime, even without intervention. “That was an important point, because many people think kids who end up in prison for serious offenses are the worst of the worst—but the great majority of them were not,” he said. “That research suggested that maybe we shouldn’t put kids in prison for long periods of time.”

The late psychologist Tom Dishion, PhD, dove deeper into the idea of adolescence-limited behaviors, focusing on the role of peer influence. For AL youth, spending time with deviant peers is one of the most common paths to becoming more deviant, Dishion found. Yet ironically, many interventions placed at-risk youth into group therapy or community-based programs that put them in closer contact with delinquent peers (Annual Review of Psychology, Vol. 62, 2011opens in new window).

“Dishion showed that those things could actually increase the delinquency of kids who would otherwise not have gotten very far into it,” Moffitt said. But thanks to her model, researchers could start to think about predicting whether an adolescent offender was likely to outgrow the behavior and tailor interventions accordingly. Dishion’s work suggested that for offenders likely to be adolescence-limited—identified by the absence of risk factors such as cognitive deficits and a history of early childhood aggression—individual and family therapy would be more beneficial than treatments involving peers or placement in juvenile detention settings.

Meanwhile, other scholars focused on the smaller group of LCP offenders and how to prevent people from developing lifelong antisocial behaviors. Moffitt’s paper linked antisocial behavior with verbal and executive functioning deficits that emerged early in life. That suggested the need for much earlier interventions to prevent future delinquent behavior. “Up until the time of my paper, most of the interventions were with teenagers. By then, I argued, the horse was out of the barn,” she said.

As the participants reached adolescence, however, Olds couldn’t help but notice that kids in the Nurse-Family Partnership program were less likely to engage in criminal behavior than those in a control group were. Yet those differences seemed less pronounced as the teens entered later adolescence. Moffitt’s paper helped him understand why. “It was like a lightbulb,” he said. It wasn’t that the program’s impact was fading as the teens got older. Rather, a spike of age-appropriate acting out among AL offenders was muddying the waters, making it harder to see differences between kids in the intervention group and those in the control. “Adolescence-limited antisocial behavior cuts across social classes,” he said. “But if you are born into a highly inequitable context, you’re more likely to experience adverse childhood experiences that predispose you to life-course-persistent antisocial behavior.”

Following up on his participants at age 15, Olds found the program had significant effects on antisocial behavior. Teens born to mothers who received nurse visits had fewer instances of running away, fewer arrests and convictions, and fewer behavioral problems related to alcohol and drug use (JAMA, Vol. 280, No. 14, 1998opens in new window). Still, the youth in the study faced many hardships, and the program’s effects on crime in boys were no longer significant by age 19. However, the impact on girls remained notable. Some 30% of girls in the control group were arrested, and 20% were convicted of crimes. Among girls in the intervention group, just 10% had a history of arrests, and 4% had convictions (Archives of Pediatric Adolescent Medicine, Vol. 164, No. 1, 2010opens in new window).

These impacts on criminal behavior helped draw some important early attention to the Nurse-Family Partnership program, Olds said. “Terrie’s work helped us dig into the data in ways we may not have otherwise, and it reinforced the messages we were uncovering: There’s no better time to intervene in the lives of mothers and families than during pregnancy and the very early years.”

The Nurse-Family Partnership continues to work toward that goal today, serving more than 54,000 families annually in more than 250 communities. Beyond crime, the evidence-based program has shown positive outcomes for maternal health, child health and development, and family stability and success. “We need to make sure that children get off to the very best start in life,” Olds said.

Over the past 3 decades, scholars have expanded on Moffitt’s work to illuminate the factors that contribute to deviant behavior and to find ways to prevent it. In 2018, Moffitt was awarded an international prize from the Belgium-based International Juvenile Justice Observatory for her contribution to more effective evidence-based policies for young people in conflict with the law. (Among many other rewards, she also received the 2024 APA Award for Outstanding Lifetime Contributions to Psychology.)

She continues to add to the body of knowledge on antisocial behavior across the lifespan and now serves as the associate director of the Dunedin study alongside her husband and research partner, Avshalom Caspi, PhD, who is also a professor at Duke University and King’s College London. Moffitt is also a cofounder of the Environmental Risk Longitudinal Twin Study, which follows a group of participants born in the United Kingdom in 1994. Using twin data, she and her colleagues are further probing the risk factors for LCP antisocial behavior, while other recent research uses neuroimaging tools to investigate differences in AL and LCP offenders.

Following study participants into middle age has also opened up new avenues to explore, including a host of topics related to health and aging. “As participants get older, you have to let go of studying things like risky sex that you studied when they were 15, and start studying things like heart attacks, which they now have in their 50s,” she said. “It keeps me on my toes, developing new expertise and bringing in collaborators to incorporate these new topics in their lives. What makes longitudinal studies so exciting is constantly getting to learn something new.”