Turning Points and the Future of Life-Course Criminology

From Criminal Careers to the Life Course

It is within this intellectual space that, in retrospect, we believe our own work must be contextualized. Although in the mid-1980s neither one of us was particularly interested in criminal careers, ¹ a fortuitous discovery led us to consider the NRC report in some detail. The short explanation is that one of us (Laub) stumbled across the original longitudinal studies of the Gluecks in the basement of the Harvard Law School Library. The studies were 500 Criminal Careers (men on parole from the Massachusetts Reformatory followed up for 5 years, 10 years, and then 15 years), 500 Delinquent Women (women on parole from the Women’s Reformatory followed up for 5 years), 1000 Juvenile Delinquents (boys in Boston juvenile court who were referred to the Judge Baker clinic and followed up 15 years later), and the famous Unraveling Juvenile Delinquency (UJD) study and subsequent follow-ups (Glueck and Glueck 1950, 1968). Laub then called his good friend and colleague from graduate school, Rob Sampson, who was then on the faculty at the University of Illinois, Urbana–Champaign. Since our graduate school days, we had talked about writing a paper together. We discussed the idea of using modern methods and current theoretical knowledge on families and crime to see if the Gluecks’ findings held up.

The discovery of the Glueck data archives was propitious. One of the major arguments of the NRC report was that scientific knowledge regarding crime and delinquency had long been hampered by the sheer lack of good data, especially data of a longitudinal nature. The NRC panel called for prospective longitudinal studies that included data on “crime and arrest sequences; on each individual’s early childhood experiences; on his parents, siblings, and peers; on school experiences and work experiences; on deviant behavior of various sorts; and on interactions with the justice system” (Blumstein et al. 1986:209). But according to the panel, “No single data source yet collected contains so rich a set of information on an appropriately broad sample” (Blumstein et al. 1986:209). ² This was a discouraging message in light of the costly and time-consuming reality of collecting new longitudinal data.

By contrast, the existing data from Gluecks’ UJD and follow-up studies (1950, 1968) appeared to have all the properties the NRC panel delineated for a rigorous study of criminal careers. Despite some critical reviews, the matched-sample research design employed in the UJD study and the quality and scope of the longitudinal information collected by the Gluecks’ research team had actually long been recognized (e.g., Hirschi and Selvin 1967; Reiss 1951:175-79). But while these data were some of the best that the field of criminology had to offer, the original authors had coded and analyzed the data using a cross-sectional mind-set. For example, we knew from the Gluecks’ coding scheme that a particular man had seven arrests from age 17 to 25, but we had no idea of the sequencing or ages at these events. Similarly, we knew that the same man had seven criminal justice sanctions between ages 17 and 25, but we did not know which sanction was tied to which arrest. The Gluecks also artificially truncated the upper end of the arrest distribution, negating the possibility of studying “career criminals.” We thus seized the opportunity to design and execute a research strategy to recode, computerize, and analyze what to that point had been an inaccessible data resource.

Luckily, the crucial information needed was available in the raw data, allowing a full-scale assessment of serious and persistent offending. With the help of an intense weekend of study in Pittsburgh with Al Blumstein and Jackie Cohen, we developed a coding scheme to create a longitudinal record of criminal offending for each man that captured the timing and sequencing of each arrest and criminal justice sanction from age 7 through age 32. The Glueck men generated over 6,300 arrest events in this age range, and we coded nearly 60 offense types and more than 20 criminal justice dispositions. This coding scheme, based on the Criminal Careers report, allowed us to capture with precision and rigor the various dimensions of criminal careers. The data were also rich in the number and quality of measures across childhood, adolescent, and adult development. The potential to examine influences on crime, accounting for important background factors and the changing nature of life events—especially during the transition from late adolescence to young adulthood—was enticing.

Based on this potential, the paper records from the UJD study were eventually moved to the Henry A. Murray Research Center for the Study of Lives, then at Harvard’s Radcliffe College. We also wrote a grant proposal to the National Institute of Justice for funds to reconstruct the Glueck data archive and analyze the extensive longitudinal data to study criminal careers. Much to our surprise, we received funding from the National Institute of Justice (NIJ), which ultimately led to a series of papers and our first book analyzing the Glueck data, Crime in the Making: Pathways and Turning Points Through Life (Sampson and Laub 1993). But the pathway from the NRC report and our NIJ grant application to the book was anything but straight.

While we were ensconced in the basement of the Murray Center coding the Gluecks’ criminal history data, we had to contend with the swirling debates about age, crime, and criminal careers. Much heat was expended, but in our view this debate had the salutary effect of clarifying the facts about crime that need to be explained. Drawing on the accumulation of criminological research in this regard, our early assessment was that criminology in general and the NRC report in particular had neglected childhood characteristics and thus failed to come to grips with the link between early childhood behavior and later adult outcomes. At the same time, it was clear to us that criminology did not pay much attention to the transition to adulthood and transitions from criminal to noncriminal behavior, which is now referred to as desistance from crime. To be sure, the criminal career debate had ushered in a new era of longitudinal research on factors such as onset, escalation, and persistence within the population of offenders, an era that again we view as a disciplinary turning point. After all, Gottfredson and Hirschi’s (1986) critique would have lacked force absent the “paradigmatic” status that the NRC report promoted.

But despite promising leads and an accumulation of important facts on age, crime, and criminal careers, we felt by the early 1990s that many of the most interesting questions had gone unanswered and were in fact not even asked by the NRC report. Although the aggregate age–crime relationship was central, we began to see the debate as a much larger one: how does the field of criminology conceptualize its dependent variable? Perhaps more importantly, we came to the view that the criminal career debate had done little to advance discoveries about the processes that explain persistent offending and desistance from criminal behavior across various stages of the life course. Thus, while our ambitions were to use the Glueck data to assess criminal careers, we ended up shifting direction to offer a new way of thinking about crime. ³

In 1990, we published the initial results of this shift by introducing a life-course perspective to criminology and developing a theoretical model of age-graded informal social control to account for both persistence in and desistance from crime (Sampson and Laub 1990). The life-course perspective embraces the idea of continuity and change in behavior as individuals age; recognizes the importance of multiple factors, multiple pathways, and multiple contexts in understanding behavior; acknowledges the prominence of co-occurring problem behaviors; and highlights the salience of social ties and social control (see also Sampson and Laub 1992). We later defined this general approach as life-course criminology to distinguish, in our view, the differences from the criminal career paradigm as presented in the NRC report.

The Age-Graded Theory of Informal Social Control

Crime in the Making: Pathways and Turning Points through Life (1993) served as the first step in our theoretical framework that explains childhood antisocial behavior, adolescent delinquency, and crime in adulthood. We began with Hirschi’s (1969) classic theoretical assertion that crime is more likely to occur when an individual’s bond to society is attenuated. But following the similarly classic work of Elder (1974, 1985) and a point that was implicit in the NRC report, we differentiated the life course of individuals on the basis of age and argued that salient institutions of both formal and informal social control vary across the life span. In addition, we accounted for processes of continuity and change in criminal behavior over the life course, an issue that was central to the Criminal Careers report. Thus, despite the effects of cumulative continuity and the “state dependence” of prior antisocial behavior, we contended that salient life events and socialization experiences in adulthood—also known as “turning points”—are capable of modifying trajectories of crime (Sampson and Laub 1990). We assessed the theory by analyzing the longitudinal data from the Gluecks’ UJD study of delinquency from childhood to age 32.

In a second book (Laub and Sampson 2003), we reported on a 35-year follow-up to supplement the existing data with three new sources of data collection: criminal record checks (local and national), death record checks (local and national), and personal interviews with a sample of 52 of the original delinquents. The men were selected on the basis of their trajectories of adult offending in order to maximize variation. This new compilation of data encompassed crime and life experiences of the Glueck men from age 7 to 70, thus extending the longitudinal focus of the Criminal Career report over the full life course. Although not abandoning the importance of informal social controls, the revised version of the theory recognizes the importance of agency and situated choice, routine activities, aging, macro-level historical events, and local culture and community context. ⁴ Thus, Shared Beginnings, Divergent Lives reflects an effort to further unpack the processes of persistence and desistance and illuminates why and how adult social institutions and roles have the capacity to reorder life-course trajectories.

Both books offer findings that bear on the legacy of the NRC report. Through a series of within-individual trajectories of age and crime, we found variability in the age–crime curve among individuals, but also that crime declines even among active offenders. Thus, the aggregate age–crime curve does not necessarily mirror individual age–crime trajectories, as the evidence demonstrates variation in the peak in offending and age at desistance. Yet the underlying process of desistance follows a fairly similar path across all offenders as all offenses decline systemically (Laub and Sampson 2003; Sampson and Laub 2003). Stated differently, even the most serious delinquents in the Glueck sample desisted albeit at different rates and at varying time points across the life course. Moreover, we found that although child prognoses are relatively accurate in terms of predicting criminal behavior among individuals through their 20s, they do not yield distinct groupings that are prospectively valid over the entire life course—regardless of whether offenders are identified prospectively or ex post. This finding gives credence to a middle-ground position in the criminal careers debate in that there is variability among individual age–crime curves and that age has a direct effect on offending, such that “life-course desister” is the more accurate label (Sampson and Laub 2003). In this sense, we argue that general desistance processes are at work across the life course and that these processes can only be understood through the full interplay of childhood, adolescent, and adult experiences.

Given the heterogeneity in adult criminal trajectories that could not be predicted from childhood, we further argue that institutions play an important role in understanding crime over the life course. Consistent with the evidence found in Crime in the Making, turning points in the process of desistance from crime include marriage/spouses, military service, reform school, work, and residential change. We exploited the longitudinal nature of the data to examine within-individual change, holding stable characteristics of the person constant while evaluating the impact of changes in social location (e.g., marriage) on criminal trajectories. When employed, serving in the military, and/or married, the Glueck men were less likely to engage in criminal behavior. These models provide strong evidence of the probabilistic enhancement of desistance associated with turning points such as marriage, military, and employment.

The qualitative narratives in Shared Beginnings, Divergent Lives further illuminate the mechanisms underlying the desistance process. In particular, institutional or structural turning points involve, to varying degrees: (1) new situations that “knife off” the past from the present, (2) new situations that provide both supervision and monitoring as well as new opportunities for social support and growth, (3) new situations that change and structure routine activities, and (4) new situations that provide the opportunity for identity transformation. In this view, institutional turning points tend to reorder short-term situational inducements to crime and, over time, redirect long-term commitments to conformity. Consistent with this view, we also found that while not all of the Glueck men explicitly described choosing to desist from crime, many of them revealed examples of what Becker (1960) called side bets that ultimately promoted desistance. That is, before they knew it, the men had invested resources and time in a marriage or a job such that risking this investment became nonnegotiable. Although the Glueck men were active participants in their own desistance, much of the explanatory work was therefore below the surface of active consciousness and did not involve purposeful identity change, leading to what we conceptualized as “desistance by default” (Laub and Sampson 2003:278-79).

An empirical assessment of the age-graded theory and life-course framework is beyond the scope of this essay. We refer the reader to our most recent effort to make sense of the data (Laub et al. in press). We believe that the evidence is, on balance, favorable to the theory. But we have an obvious conflict of interest and thus encourage readers to ponder our original work and subsequent assessments to make their own, independent determination.

Reflections on the Future of Life-Course Criminology

Before concluding, we wish to make three brief points about the legacy of the NRC report, with an eye toward future research directions. First, potential turning points in a criminal trajectory are not static but themselves undergo change and must be theoretically reconceived in light of social change (Sampson 2015). While there will be continued attention to employment and marriage, research is needed on how broad societal changes such as the precariousness of labor markets, increasing inequality, the technological revolution, declines in violence, and the decline in marriage may be changing the context of development, a key point of emphasis in the life-course perspective. For example, technological changes have made reinventing oneself more and more difficult—nothing disappears from the Internet, computerized record checks are widely available as employers try to determine the criminal past of job applicants, and virtually every state now has a sex offender registry. Subsequent collateral consequences that foster disenfranchisement may further contribute to weakened social ties to conventional institutions. Perhaps the biggest unstudied societal change for individual development is the great American crime decline. Turning points still exist, then, but they are changing from the manifestations we found in the Glueck data.

Secondly, what has not been appropriately recognized in the debates about criminal careers is the fundamental way that life-course trajectories can be altered by government policy. A notable example is the five-fold increase in the rate of incarceration in the United States over the last 40 years (Travis, Western, and Redburn 2014). The 1986 NRC report saw sanctions mainly as an individual-level phenomenon to be attached to the explanation of criminal careers in the form of incapacitation and deterrence. In contrast, its main competitor, Gottfredson and Hirschi (1990), saw sanctions as largely irrelevant because self-control explains variation in crime and subsequent punishment, also at the individual level. In our view, neither viewpoint is conceptually equipped to account for the profound shift in punishment since the 1980s. As shown in Pettit and Western (2004), for example, macro-level changes in incarceration have reshaped the life course of recent cohorts and made going to prison a normative transition for Black men with low education. This is a classic life-course issue: the principle of cohort differences in aging (Ryder 1965). Besides the collateral consequences facing men and women when exiting prison in the “mass incarceration” era with respect to employment, education, housing, and civic life (see Travis et al. 2014 for a review), the reach of the criminal justice system extends to the courts and police as well. Like high rates of incarceration, the effects of zero-tolerance policing strategies are also not uniform and have largely fallen on young Black men, as protests in Ferguson, Baltimore, and other cities have clearly demonstrated. There is a sense that children, especially minorities, are growing up in a less forgiving environment of criminal justice response. Life-course criminology thus needs to focus more intently on social structural and institutional (including policy) contexts (Sampson 2013, 2015).

Third, there has been considerable attention in criminology of late on desistance processes, a move that can be traced back to the NRC report. One strand of this research has focused on using narratives to try and get inside the minds of offenders in order to theorize human “agency” and “identity.” Although we have gone down this analytic road in considerable depth (e.g., Laub and Sampson 2003: 141-49, 162-64, 280-82), as have several others, we now worry that this move may have been a distraction rather than a theoretical advance. Watershed changes in cognitive science and behavioral economics have made it clear that articulated motives and self-reported reasons for action are rife with ambiguity. Humans are notoriously prone to bias in understanding or explaining the causes of their own actions, in part because so much of our thinking operates automatically, beyond our deliberative awareness (Kahneman 2011).

Methodologically, the implications of cognitive distortion are profound for much of the recent narrative- and interview-based research employed in criminology, including our own. ⁵ As Jerolmack and Khan (2014) crisply put it, and as Goffman (1956) would no doubt endorse, “talk is cheap” when it comes to inferring causal mechanisms. Similarly, Watts (2014) argues that while theories of what he calls “rationalizable action” (including but not limited to rational choice) may be satisfying on empathetic grounds, they fail consistently on scientific explanation grounds. In short, our main point is that while interesting and often fruitful in the inductive discovery of possible mechanisms, offender-given reasons for committing crime (or stopping crime) are not in themselves dispositive. If they were, we would not need criminologists.

We therefore think that returning to the idea of desistance by default is an important insight that is consistent with a by-now large research program in behavioral economics and cognitive science. In their book, Nudge, Thaler and Sunstein (2008) synthesize research on decision-making to make a strong case that people can be nudged to make better decisions that will improve their lives. They argue that small interventions can have large effects, a direct contradiction to the idea that people first need to decide to change their conception of self or identity before changing behavior (e.g., Paternoster et al. 2015). The reality for behavioral economists and much recent research is more nearly the opposite: behavior changes identity.

The good news for policy is that we can act on this insight by changing the choice architecture and “nudging” offenders to make better decisions. As states look to reduce both the rate and length of incarceration, for example, attention can be directed to making probation and parole practice more “desistance focused” through supervision of offenders in the community. At the same time, probation and parole officers can serve as resources for offenders regarding employment, housing, substance abuse treatment, and other human needs. In this manner, probation and parole officers can act as monitors and mentors to provide both supervision and social support. As a result, behavioral change away from crime may well occur even in the absence of articulated identity change (Laub 2016). In short, a turn to behavioral strategies for engineering turning points seems to us to be better supported by research than continued attempts to divine inner mental states or change people’s minds.

Conclusion

The NRC report was clearly generative of debate and inspired a virtual crush of empirical studies, but it lacked an orientating theoretical explanation of crime. Moreover, the report was limited to a focus on the dependent variable—criminal behavior—with little attention paid to key independent variables beyond basic demographic characteristics of offenders. Likely as a result of this empirical stance, much of the research embraced by the criminal career paradigm has been atheoretical or empiricist in nature. In particular, a major research question stemming from the NRC report is the ex post identification of persistent high-rate offenders that constitute a distinctive group. This question goes to the heart of typological thinking about crime that has come to dominate recent criminological literature. ⁶ In our view, the typological view of criminal groups is, on balance, not supported in terms of prospective or ex ante evidence (Gottfredson and Hirschi 1990; Sampson and Laub 2003), especially in terms of distinct causal processes.

Furthermore, the policy implications stemming from the Criminal Careers report were focused primarily on criminal justice sanctions at the individual level and neglected the role of informal social controls in preventing and controlling crime. In addition, a sole focus on criminal offending trajectories misses the possible connections with other life-course trajectories such as family and marriage, education, work, and residential mobility. The key insight is that life-course trajectories can shape future criminal offending trajectories, but at the same time, life-course trajectories can be altered by criminal behavior and its consequences. Moreover, the life course is shaped by macro-level institutional changes in criminal justice policy, such as the growth in imprisonment and the emergence of “broken windows” policing.

It was into this theoretical breach of the NRC report that we stepped in the 1980s. Our journey evolved in directions that we did not anticipate; there is little doubt in our mind that a random event—the discovery of the Glueck data—was instrumental in our discoveries. Although the direction of our research did not always follow the path laid out in the NRC report, certainly we were influenced by its strong arguments. Our life-course theory of crime may therefore be seen as an unintended consequence of the Criminal Careers report and its critics—one that we hope continues to productively shape research and theory.