V. The Age–Crime Relationship in Traditional Criminological Theory
Despite the critique leveled by Hirschi and Gottfredson (1983), the major theoretical traditions in criminology (i.e., strain, social control, and social learning theories) have all been used to provide explanations for variation in criminal behavior over the life span. For example, strain theory argues that adolescents and young adults experience more status frustration and strain, which eases with entry into adulthood and legitimate employment. The relative deprivation experienced by youth declines with entry into the legitimate adult labor market (Greenberg, 1985). Theorists have also incorporated some elements of strain theory when considering Easterlin’s (1978) perspective on relative cohort size. Easterlin argued that larger cohorts (e.g., the Baby Boomers) face certain disadvantages, such as competition for scarce resources, that result in higher levels of economic deprivation for that cohort. Although Easterlin highlighted the negative economic conditions consistent with a strain perspective, he also suggested that large cohorts may overwhelm social institutions, subjecting cohort members to additional criminogenic social conditions, such as reduced supervision, weakened socialization, and lower levels of social control. These conditions prove to be most detrimental for adolescents and young adults and may account for increasing crime rates when these large cohorts enter the most crime-prone years (i.e., late adolescence and early adulthood).
As Easterlin (1978) and Greenberg (1985) suggested, social control theory may also provide an argument for the changing of crime rates by age. Sampson and Laub (1995) pointed out that the impact of both formal and informal social controls varies by age. This theory argues that social bonds are weakened during adolescence, freeing an individual to violate social norms. Thus, adolescence represents a time when attachments to conventional others, especially parents, and commitment to conventional institutions are reduced. Social bonds may be re-formed in adulthood as individuals accumulate conventional ties to jobs and begin to build their own families through marriage and parenthood. In addition, the consequences of crime become more serious with age and function as more of a control on behavior as individuals amass a greater stake in conformity (Steffensmeier et al., 1989).
Differential association would anticipate that increasing involvement in crime during the adolescent years is due to variation in experiences with delinquent peers (Warr, 1993). In support of this perspective, Warr (1993) used data from the National Youth Survey to demonstrate age-related changes in exposure to delinquent peers, including the percentage of friends who are delinquent, time spent with peers, and the self-reported importance of peers, that correspond to the age–crime curve. During later periods of adolescence individuals in the National Youth Survey reported a larger number of delinquent friends, more time spent with those friends, and more importance of peers in their lives. In multivariate models, the relationship between age and crime was attenuated when peer variables were included. Thus, Warr’s research suggests that the age– crime relationship may be at least partially explained by changes in peer associations.
Stolzenberg and D’Alessio (2008) more recently examined the implications of peer association in a different way, addressing whether changes in co-offending account for age-related variations in criminal involvement. Researchers have consistently noted that criminal behavior during adolescence is largely a group phenomenon. This pattern of co-offending may explain the increasing prevalence rates during adolescence that are apparent in the aggregate age– crime curve. Stolzenberg and D’Alessio put forth the following argument:
The greater prevalence of co-offending among juveniles, engendered to a large extent from the influence of criminally inclined peers, in turn explains why crime levels peak during adolescence and then begin to decline in early adulthood following graduation from high school. (p. 69)
If this perspective is true, the age–crime relationship should be most apparent when one is considering crimes involving co-offending, but it should disappear when co-offending is taken into account. In other words, the age– crime curve for solo offending should be flat, whereas the curve for co-offending should demonstrate the typical curvilinear pattern. Using National Incident Based Reporting Systemdata, however, Stolzenberg and D’Alessio (2008) found two interesting results. First, contrary to much of the discussion surrounding adolescent offending, co-offenses are not the most common pattern; instead, solo offending is more common for all age groups, including juveniles. Also, the age–crime curve emerges for both solo and co-offending, suggesting that accounting for co-offending does not attenuate the typical age–crime relationship. Thus, patterns of co-offending do not appear to account for the age–crime curve.
Marvell and Moody (1991) argued that although there is no shortage of speculation about the causes of the age– crime curve, there is little empirical support for any of these explanations, concluding that there is no firm theoretical foundation for the age–crime curve. Tittle and Grasmick (1998) examined a variety of age-varying criminogenic factors but found that including these factors did not seem to account for the age–crime relationship. They reported an inability to discount Hirschi and Gottfredson’s (1983) theoretical arguments and concluded that it is not easy to explain away the age–crime curve.
A. Propensity Versus Developmental Theories
More recent criminological theories have attempted to explain the curve itself as well as to understand changes in levels of crime over the course of the age–crime curve. Two strategies of accounting for variation across the life course are apparent in criminological theory: (a) propensity theories and (b) developmental or life course theories. Propensity theories point to a single underlying stable trait that causes crime at all ages. The most well-known and most frequently tested propensity theory in criminology is Gottfredson and Hirschi’s (1990) general theory of crime, which suggests that crime and other risky behaviors at all ages are the result of an individual’s low level of self-control. Using a theoretically stable trait to account for obvious age-related variation in criminal offending patterns may seem counterintuitive. Gottfredson and Hirschi contended that low self-control produces criminal behavior in the presence of criminogenic opportunities. Thus, opportunities for crime may vary across the life course even though levels of self-control are relatively stable. Hirschi and Gottfredson (1995) suggested that variation in the opportunity for crime by age accounts for a great deal of the variation in actual criminal activity observed.
In concert with their earlier assertion that the relationship between age-varying social factors and offending is spurious because of age (Hirschi & Gottfredson, 1983), propensity theory argues that any relationship between social factors (e.g., deviant peer associations and weak social bonds) and offending is spurious because of self-control. In other words, deviant peer associations and weakened social bonds are related to offending because they are all caused by the underlying factor of self-control. Individuals with low levels of self-control are also more likely to associate with deviant peers and have difficulty forging and maintaining the conventional connections that foster strong social bonds. Thus, they continue to argue that criminological theories attributing changes in offending over time to changes in social factors are inadequate.
In contrast to propensity theories, developmental or life course theories of offending point to age-related variations in criminogenic factors to explicitly account for the age– crime relationship. Both developmental and life course theories look to the full life course in their explanations of offending. Moffitt’s (1993) developmental theory, for example, starts with the conclusion, based on empirical research, that the age–crime curve represents differences in prevalence by age, with a larger proportion of the adolescent population engaging in delinquent or criminal activity. She also argued that the aggregate age–crime curve masks group differences in the relationship between age and crime. In other words, she noted that individual variation in the frequency of offending by age is hidden within the aggregate age–crime curve.
Moffitt (1993) proposed a typological perspective that identifies two separate groups of offenders, each with a different age–crime curve. Thus, the aggregate age–crime curve is a mix of a small group of long-term offenders (referred to as the life-course-persistent offenders), which has a relatively flat and stable age–crime curve, and a larger group of individuals with a short-term period of delinquent involvement occurring during adolescence (referred to as adolescent-limited offenders), which demonstrates the typical age–crime curve with a large peak during late adolescence. With two different offending patterns, these two groups require different etiological explanations. According to Moffitt, life-course-persistent offenders become involved in criminal behavior early in life and persist in their criminal activity because of the combination of neuropsychological deficits, inadequate parenting, and cumulative disadvantage associated with the negative consequences of early criminal involvement. Adolescent-limited offenders, on the other hand, engage in offending for a relatively short duration. Entry into offending is explained by a maturity gap, in which youth may be biologically mature but remain dependent on and under the control of their families. Minor offending occurs in an attempt to gain some independence and as a result of the imitation of antisocial models. Desistence in this group occurs in early adulthood as social bonds increase and the consequences of criminal activity become more punitive.
Life course theories of offending similarly point to long-term patterns of offending and social forces that operate over the full life course. Sampson and Laub’s (1993) age-graded theory of social control highlights the processes of both continuity and change in behavior over the life course, looking at both differences between individuals and differences within individuals over time. Entry into delinquent behavior is accounted for by a variety of social factors, including weak social bonds to family and school in childhood and adolescence. Desistance from delinquency occurs with the accumulation of social bonds, namely, strong marriages, stable employment, and other stabilizing influences, in the transition to adulthood. Persistence (i.e., continuity), on the other hand, occurs as a result of the cumulative disadvantage of early criminal involvement. Sampson and Laub (1995) contended that criminal behavior further attenuates already-weakened social bonds by limiting opportunities within conventional society. Some scholars have argued that the focus on social bonding is too narrow and that many life events may function to alter peer association more in line with a learning perspective instead of a social control perspective (Stolzenberg & D’Alessio, 2008). Sampson and Laub (1995) contended that this does not directly contradict their theory, and the most recent version of the theory (Laub & Sampson, 2003) has expanded to accommodate social bonding, peer, and routine activity influences that change over the life course. For example, marriage may strengthen social bonds as well as attenuate preexisting deviant peer associations and restructure routine activities and criminal opportunities. These life events, then, account for the peak of offending during late adolescence and the dramatic decline in offending that occurs shortly after the transition to adulthood.
B. Variation in the Causes of Crime by Age
One question remaining from traditional and life course/developmental theories is whether the causes of crime are the same regardless of age. As might be expected, Hirschi and Gottfredson (1983; Gottfredson & Hirschi, 1990) have argued that the causes of crime are the same at all ages; in other words, social factors do not interact with age to produce criminal behavior. Other theorists suggest that the causes of offending may vary by age. For example, Moffitt (1993) pointed to a variety of theoretical factors that may influence offending at different ages, including early neuropsychological deficits and parenting influences, negative peer associations in adolescence, and social control mechanisms in later adolescence and early adulthood. Tittle and Grasmick (1998) examined the interaction thesis and found no evidence that age interacts with criminogenic forces to produce criminal behavior. Again, they had difficulty discounting Hirschi and Gottfredson’s (1983) assertions; however, this issue remains open for debate and empirical investigation.