More than a half-century on from Delinquency in a Birth Cohort (Wolfgang et al., 1972), few scientific facts in criminology enjoy stronger empirical validation than the existence of a small core of offenders, roughly 5%–10%, who account for most crime and violence in a population. Regarding crimes like murder, rape, kidnapping and armed robbery, habitual, chronic or career criminals, as they are variously referred, are the most likely suspects. In fact, the asymmetry in offending is so established it appears in separate but related conceptual models in the social sciences, including the developmental taxonomy (e.g. Moffitt, 1993, 2018; Moffitt and Caspi, 2001), severe 5% model (Vaughn et al., 2011, 2014), and p Factor (Caspi et al., 2014, 2024), which themselves have strong research support.
The criminal career paradigm does not rest on these laurels and continues with advances in data, measurement and justice system application. Sometimes, newer research reminds us of lessons we learned long ago. For example, Gomes et al. (2025) examined chronic offending in the 1982 Pelotas Birth Cohort in southern Brazil. Among those 5623 individuals from South America – a long way from boys born in 1945 in Philadelphia – they identified 5% who were chronic offenders. Those individuals were responsible for 57% of total crime, 54% of violent crime and 68% of homicides. On every parameter of the criminal career, chronic offenders were the worst and more than 41% of them had prison history. I will return to the importance of prison later.
Another major recent achievement in the field is McCuish et al.’s (2021) splendid book The Life-Course of Serious and Violent Youth Grown Up: A 20-Year Longitudinal Study, which is based on the Incarcerated Serious and Violent Young Offender Study in Canada. Their exemplary research is a 20-year longitudinal study, which allows for observation of the twists and turns of the offending career across developmental periods. The most severe developmental trajectory in their study is called high-rate persistent, and it shows strong overlap with traditional chronic offending.
During emerging adulthood, nearly 82% of high-rate persistent offenders were chronic offenders. During full adulthood, the prevalence was nearly 72% and during mature adulthood through age 35, the prevalence was greater than 78%. Although it fluctuates across time, chronic offending also shows substantial stability. The worst offenders are almost always the worst offenders at different time points, they hold their place in the offending distribution.
Another refreshing feature of McCuish et al.’s (2021) book is the empirical assessment and specification of psychopathy to examine its association with various offending trajectories. Spoiler alert: the worst offenders are usually the most psychopathic. Consistent with general theories that confidently assert the importance of psychopathy for understanding antisocial behavior (DeLisi, 2016), they found that psychopathy during adolescence predicted chronic offending 20 years later (also see, McCuish and Lussier, 2025). The incorporation of a robust individual-level construct to instantiate the worst offending pathways is an exciting advance especially since the criminal career paradigm has historically been empirical, operating without a theoretical beacon.
Invariably, career criminals end up in prison, and although there are some exceptions, they generally fare very poorly in custody and are disproportionately responsible for the worst misconduct, including homicide. Several studies show that habitual criminals gravitate toward assaults, extortion and murder, and are keen to join security-threat groups, including the Aryan Brotherhood, Mexican Mafia and Black Guerilla Family, among others. These deviant associations usually produce trouble for correctional authorities and constitute fundamental threats to institutional safety (Butler et al., 2021; Cihan and Sorensen, 2019; DeLisi et al., 2021; Long, 2022, 2025; Sorensen et al., 2024). Career criminals disproportionate share of crime is effectively the same in custody as it is in the community. A few create havoc for many.
In a brilliant application of the Pareto Principle, the statistical regularity that a small number of inputs is responsible for a large number of outputs, Morgan et al. (2025) analyzed a national sample of US prisoners and found those at the 99th percentile of the distribution accounted for most of the problems. Among males, 1% accounted for 36% of misconduct in US prisons and among women, 1% accounted for 35% of the misconduct. Their findings are noteworthy on several grounds, but two points are important to consider. First, even though American prisons are filled with serious and chronic offenders, the majority of them are not terribly difficult to manage in custody. Based on their low levels of institutional misconduct, they are perhaps sending signals of a productive and prosocial reentry to society upon release.
Second, prisoners at the 90th percentile of misconduct and above serve as an important research target to identify the factors that best differentiate the truly dangerous from the mostly chronic. Morgan et al. (2025) identified several covariates associated with extreme offender status, but these were offending career, externalizing psychopathology and demographic features we already know are stable correlates of sustained criminal activity. Further inquiry and granular analysis of the offending outliers are needed.
There is also a need for measurement refinement. According to the US Bureau of Justice Statistics, more than 77% of prisoners admitted to custody across 34 states in 2014 had five or more arrests on their record, the criterion – harkening back to Wolfgang – most often used to identify career criminality. That is a disconcerting datum to digest: three in four American prisoners is a career criminal, likely one reason why more than 80% recidivate within a few years of prison release. In fact, the 10% of prisoners with the greatest arrest histories account for nearly one in four total arrests across American prisons (Durose and Antenangeli, 2023).
I have personally interviewed offenders with a few hundred arrests and have seen news reports about individuals with between 500 and 700 arrests (not arrest charges, discrete arrests). These highly prolific offenders are usually transient and typically perpetrate municipal, public-order violations (e.g. public intoxication, disorderly conduct, trespassing, verbal harassment/menacing), which routinely carry very brief jail sentences. Prodigiously chronic, one could argue they are not necessarily dangerous.
As a result, measurement refinement of career criminality is needed to best differentiate subtypes of habitual criminality. The 5+ arrest criterion works very well in the community evidenced by birth cohort and other study groups worldwide (see, Edelstein, 2016; Liggins et al., 2019; Martinez et al., 2017), but in the prison catchment area, the criterion seems too low.
Recently, I created the concept of a superhomicide offender, one with at least five murder convictions, to better locate multiple homicide offenders in the criminal career nosology and place them at the 95th percentile of the offending distribution. Compared to less prolific murderers, superhomicide offenders have more extensive criminal careers (e.g. more arrests, convictions and prison sentences and earlier onset), are more psychopathic overall and specifically regarding interpersonal and affective deficits, and have much worse homicidal ideation (DeLisi, 2024a, 2024b). The superhomicide offender is a research innovation which shows the compatibility between homicide studies and criminal career research, which is imperative given that career criminals are most responsible for homicide offenses.
Basic research is essential to science and I will always have a place in my heart for scholarly exploration and discovery even when research is conducted for its own sake. But there also comes a point where an academic discipline has to stand confidently on its knowledge base and translate that knowledge into practice and meaningful application. Career criminals research is precisely an area where this is true.
We know that a small cadre of offenders are recalcitrant and dangerous, thus they serve as essential targets to reduce crime and enhance public safety. Below are some ideas criminologists will hopefully explore:
Three-strikes sentencing policies in the US are routinely criticized in academia but unfortunately very rarely used in practice. Serious offenders routinely have exponentially more felony convictions, even for the most strikable violent felonies, than three, which would theoretically trigger sentencing enhancements. Concerted research on criminal careers, sentencing practices and the invocation of habitual offender statutes need updating.
Serious offenders serve very little time in prison. The median time served for a violent conviction in the US is 2.4 years. Murderers serve less than 18 years in prison, rapists serve about seven years (Kaeble, 2021). Those convicted of the most violent crimes are most likely to be career offenders, yet they are recirculated into custody with a speed that belies their original sentence. Examination of the feedback loop between limited time served in prison and subsequent offending careers is needed.
Led by the pioneering work of Joan Petersilia, an early emphasis of career criminal research in the 1970s was producing estimates of the actual amount of crime they inflict by using self-reports and arrests, and thus estimating the dark figure of crime. In this journal, we recently found one-third of condemned murderers has additional homicides for which they are likely responsible (Minkler et al., 2024). The most active murderers have dark figures upwards of nearly 100 additional victims, and they too are usually career criminals. More data are needed from diverse data sources (e.g. community, forensic inpatient, correctional) to provide better estimates of the dark figure of crime especially since crimes usually remain unsolved.
Pathological criminals are an invaluable data source, but for several logistical and bureaucratic reasons, they are effectively inaccessible to criminological researchers. As the RAND researchers did decades ago, more interviews with prolific criminals are needed to provide first-person insight about the factors, situations, dynamics, processes and decision-making that drive recurrent criminal activity. Career criminals can illustrate the variables most responsible for their conduct, an approach that can offer interesting synergy between narrative criminology, qualitative methodology and criminal careers research.
Greater research of severe institutionalized offenders, such as those serving death or multiple life sentences, could help to identify those with information about unsolved and cold cases. These individuals might be useful for exceptional case clearance of homicide as I have indicated elsewhere (DeLisi, 2024a, 2024b). If pathological offenders are willing to divulge accurate information, which is more likely then they are in their 70s or 80s, they should be listened to to provide legal closure for victim families.
Finally, the etiology of career criminality highlights the importance and interplay between individual-level risk factors and criminogenic environments. Nowhere is this more the case than adverse childhood experiences, the array of childhood abuse, neglect and household dysfunction, which commonly set antisocial trajectories in motion. We know that greater trauma history is associated with more antisocial outcomes (Baglivio and Wolff, 2023; Malvaso et al., 2022), in-depth study of the severity, frequency and range of those experiences along with the specific relation to offending patters are needed.
