CONTINUING EDUCATION: TOWARD A LIFE-COURSE PERSPECTIVE ON SOCIAL LEARNING

Abstract

Sutherland’s differential association theory and the life course perspective have at times been conceptualized as contrasting theories of criminal behavior. I argue instead that our understanding of delinquency, the dynamics underlying criminal persistence and desistance, and intergenerational patterns, will be enhanced by a more explicit integration of these two traditions. I focus on family processes, as these are foundational intimate relationships that remain underappreciated as a source of lifelong learning and influence. While family support and variations in parental supervision have been amply investigated, ‘direct transmission’ takes place within the family as well as within the confines of the more heavily studied world of adolescent peer groups. I identify five dimensions of direct transmission, and illustrate these dynamic processes with qualitative data from two longitudinal studies and results of recent quantitative analyses. The analysis is generally in line with Sutherland’s original formulation, but includes several extensions and modifications. It is important to include a role for human agency, and for ‘non-criminal’ definitions and lifestyle factors, in addition to the directly criminogenic definitions Sutherland and subsequent researchers have emphasized. The focus on social processes is, however, consistent with Sutherland’s goal of highlighting limitations of psychological and biological differences explanations.

Sampson and Laub (1993) advanced the life course perspective (LCP) within criminology, and their work energized research on topics ranging from age-related trends to the study of intergenerational continuity and change. Yet while Laub (2006) suggested in his Sutherland address that the LCP should be considered the ‘soul’ of criminology, it is perhaps most accurate to classify this perspective as a paradigm or lens and set of basic principles. Accordingly, the LCP can be considered a kind of skeleton or scaffolding upon which to build more concrete, fleshed-out theories of criminal behavior. For example, within the context of Sampson and Laub’s research, it is the age-graded theory of informal social control that represents the core of their ideas about the process of ‘desisting’ from crime. A major contribution was to extend key tenets of control theory beyond adolescence and the issue of initial onset, to explore ways in which later transition events influence the likelihood of crime cessation during adulthood.

If we accept the notion that the LCP is a general framework or lens, then it is potentially useful to consider that other theories can similarly benefit from a thoroughgoing life course treatment. I focus here on differential association and other versions of social learning theory to highlight the value of applying a life course lens to this important—and yes—underexplored theory of criminal behavior. However, this endeavor does seem at odds with the idea that differential association and the LCP should be viewed as contrasting approaches. For example, Laub (2006) suggested that it is the LCP and not differential association that best fits known facts about crime (e.g., “the early onset of offending, the decline in offending with age…”) (p. 239). While the process of extending the lens to the full life course has been useful for documenting these overall patterns and trends, the LCP itself does not really tell us much about what is going on in the lives of the ‘early starters,’ suggest why many individuals reduce levels of criminal involvement in adulthood, or identify factors associated with a longer period of persistence. Theories need to add to our understanding of these why questions, and foreground the most salient dynamic processes.

Sampson and Laub’s emphasis on social control mechanisms in general fits these criteria well. Yet a more explicit linking of the life course and social learning traditions and findings of prior research provide the basis for offering several assertions. First, the specific mechanisms identified by control theorists provide an incomplete accounting of human conduct in general and criminal behavior in particular. This is the case whether the focus is on initial onset (Giordano, Cernkovich & Pugh, 1986), the desistance process (Giordano, Cernkovich, & Rudolph, 2002), or the phenomenon of intergenerational transmission (Giordano, 2010; Giordano, Copp, Manning, & Longmore, 2019). Second, the dynamics that have and some that I argue should be included under the umbrella of social learning theory not only help to round out the picture, but should be central to our theorizing. What researchers choose to foreground or ignore has consequences not only for the emphases of future studies, but for policy choices, programmatic efforts, and eventually the lives of offenders and would-be offenders who could potentially benefit from the findings connected to these lines of inquiry.

Frank Cullen (2011), in his own Sutherland address, noted the strong influence of sociology and sociologists, including Sutherland, on the development of the discipline. Yet Cullen argued in part that: a) criminologists need to be more wholehearted in their embrace of other disciplinary areas, including the study of biosocial influences, b) stand to gain little by yet another study examining Hirschi’s four components of “the bond,” c) would benefit from more direct contact with samples of offenders rather than high school students, and d) should find ways to improve rather than simply continue to critique various aspects of the existing criminal justice system. Cullen’s discussion provided a refreshing critique of certain elements of the status quo, but the suggestion that “we can no longer pretend that biology is not intimately implicated in human behavior and thus in criminal behavior,” (p. 311) and focus on the extensive stockpile of existing sociologically oriented studies hints at the need to recalibrate the field’s traditional social emphasis.

My own view is that research on social influence processes, as reflected in the social learning perspective, far from being ‘tapped out,’ not only should remain central to the discipline, but is in some ways still in its infancy. The excitement that surrounds the study of biological and other individual differences (low self-control!) notwithstanding, these general proclivities necessarily unfold within dramatically different social contexts that give shape to thoughts, constraints, daily routines, dramatic happenings, values, emotional responses, possibilities for achievement, salience hierarchies, and things to fear. These dynamics figure heavily into macro-level and more localized life-course variations in criminal behavior, and warrant additional in-depth research attention.

BACK TO THE FAMILY

A basic tenet of the LCP is the ‘linked lives’ notion (Elder & Shanahan, 2007; Settersten, 2015). As Greenfield and Marks (2006, p. 442) described it, this encompasses the idea that “people in salient relationships with each other, such as parents and children, occupy mutually influential interlocking developmental trajectories that extend throughout their lives.” Thus, it is generally accepted that the family is a key site of learning and influence, and as I will emphasize here, lifelong learning and lifelong influence (i.e., the idea of ‘continuing education’). A life course perspective on family effects, then, necessarily includes attention to early as well as later family experiences. At first blush, a focus on early family dynamics does not connect seamlessly to the LCP, however, particularly as criminologists have applied it. In general, proponents of the perspective have emphasized that various aspects of development, including criminal involvement, are not fixed based on early childhood experiences. This has led to an emphasis on later life course transitions such as entry into marriage, as factors associated with variability in criminal behavior. Notably, Sampson and Laub (1993), in their study of desistance among respondents from the Glueck study, found that marriage, employment, and military service distinguished persisters and desisters in this sample of delinquent boys followed into their adult years.

Although Sutherland’s own research did not focus heavily on family processes, his basic idea that associations vary in intensity, frequency, duration and priority (those that occur early on) and his general statements support the idea that family ties are likely to be especially consequential (Sutherland, 1956, pp. 132-133). Nevertheless, as social learning theory has developed, most investigations in this tradition have moved on to study delinquent peer influence. Several considerations likely underlie the tendency to skip over the family as a source of direct influence on delinquency and criminal behavior. First, delinquent careers generally start to take off during adolescence when peers become increasingly central, and indeed most delinquent acts take place in the company of peers (McGloin, Sullivan, Piquero, & Bacon, 2008). Second, as West and Farrington (1977) pointed out early on, most parents do not “want” their children to become delinquent. A third and related theme within the literature is the idea, as articulated initially by Hirschi (1969, p. 94), that parents, even if deviant or antisocial “generally do not advertise this to their children.” And indeed, Hirschi relied on this basic notion to suggest a limited role for direct transmission and to posit a larger impact of attenuated bonds as a critical family-related dynamic.

The idea of direct social influence via the family is certainly contained within the criminological literature, particularly with respect to behaviors such as violence (Dubow, Huesmann, & Boxer, 2003; Widom & Wilson, 2015). Yet a review of the full complement of studies that include attention to the family reveals a continued emphasis on dynamics that resonate most clearly with control theory (Hoeve et al., 2009). It is interesting to note, for example, that while Laub and Sampson’s own work focused heavily on adult crime, Laub’s discussion of the implications of the LCP for prevention and reform nevertheless highlighted the key role of early family experiences:

One of the best-known findings in criminological research s that a small group of delinquents (6 percent) account for a disproportionate share (more than half) of all criminal acts (Wolfgang, Figlio, & Sellin, 1972). A lesser-know n finding, though one of far more importance, stems from research by Donald West and David Farrington (1977). In their study of boys in London, West and Farrington found that fewer than 5 percent of the families accounted for almost half of the criminal convictions in the entire sample (109). Such a finding highlights the role that families play in the causation of crime, especially relating to parent training in monitoring, recognizing, and disciplining the misbehavior of children. Efforts should be devoted to strengthening families to help prevent crime and other problem behaviors.

(Laub, 2006, pp. 244-245).

Although this discussion thus clearly points to the importance of early family processes, Laub nevertheless retained an emphasis on dynamics that generally fall under the domain of control theory and later incarnations (i.e., Gottfredson and Hirschi’s (1990) focus on low self-control). This emphasis is in line with general trends in measurement, as most research studies include indices that tap attachment processes or supervision either as the primary focus or as family-related controls. Yet as we have argued recently, even if we know the attachment and supervision levels of all those within a given sample, we have not begun to capture all aspects of family life that operate to either foster or inhibit the use of drugs, resort to violence, or even those that figure into involvement in acquisitive crimes (Giordano et al., 2019). Thus, family life includes a host of content areas that are largely missing from criminological theorizing and research investigations.

In addition, the family is a social domain that generally keeps on ‘giving’ (i.e., continues to be a source of influence) across all phases of the life course. The idea of interdependence intrinsic to the linked lives concept may have particular resonance with reference to the contemporary U.S. context. Uncertain economic and employment prospects, particularly for young adults without a college education, later age at marriage, rates of marital/relationship instability—and likely related to these uncertainties—higher rates of residing in or returns to the parental home, paint a picture of intimate connections that often continue well past the initial adolescent-to-adult transition period (Britton, 2013; Settersten & Ray, 2010). Some aspects of the parental role clearly fit into the realm of attachment processes, and include the provision of various forms of support and assistance. Yet Sutherland and related learning theories posit another basic set of influences—mechanisms that fall under the umbrella of ‘direct transmission.’

MECHANISMS UNDERLYING DIRECT TRANSMISSION

I will focus here on five dimensions of the transmission process, concentrating primarily on the family due to their outsized and I believe somewhat underappreciated role as a source of direct influence. Yet each of these dimensions can also be viewed as elements of the continuing education that takes place within the peer group, with intimate partners, and at times even within the context of other kinds of less intimate affiliations (Giordano, 1995; Kreager & Haynie, 2011; McGloin, Sullivan, & Thomas, 2014). Key dimensions include: a) direct modeling, b) the transmission of criminogenic attitudes, c) influence via the wider arena of behaviors, attitudes, and lifestyle factors that indirectly increase or decrease the odds of enacting criminal behaviors, d) the role of these intimate ties as channels or conduits to other consequential associations, and e), integral to the life course perspective, the role of agency, as individuals respond to and thus become an integral part of the transmission process.

MODELING

The idea that children later enact behaviors to which they have been exposed early on has been amply explored, particularly in the area of aggressive actions. For example, children who witness intimate partner violence (IPV) or are victims of child abuse more often display aggression, bullying behaviors, and higher levels of self-reported delinquency, and are more likely to be arrested as juveniles and adults (Widom, 2017). Other behaviors such as parental substance use incorporate the idea of direct modeling in that parental use has been shown to be a reliable correlate and predictor of the child’s later use (Henry & Augustyn, 2017; Johnson & Leff, 1999). Yet the idea of modeling strongly implies that the child has directly observed the parent’s behavior, a dynamic that stands in contrast to Hirschi’s ‘parents hide their deviance’ observation. Nevertheless, surprisingly few studies have, for example, described family life from the point of view of children whose parents have serious substance use problems. Yet statistics indicate that a large percentage of individuals in jail and prison report current substance use or problem use histories (Mumola & Karberg, 2006). Further, a significant percentage of children placed in foster care or living with a grandparent have experienced problems related to parental substance use (Lent & Otto, 2018; Mowbray, Victor, Ryan, Moore, & Perron, 2017). Thus it appears that such children are likely to have been directly exposed to and influenced by these behaviors as they have observed them within the home. An important complication considered below is that children who may have experienced the negative effects of these behaviors on their own lives and well-being nevertheless remain at elevated risk of later adapting these or related behaviors themselves.

CRIMINOGENIC ATTITUDES

Sutherland (1947) wanted to clearly distinguish his ideas not only from biological or other individualistic perspectives, but from the work of theorists such as Gabriel Tarde who had previously explored social influence processes. Thus, he emphasized that dynamics associated with differential association included “not only imitation but all other processes of learning” (p. 22). His emphasis on communication processes is particularly important because these activities provide a much larger set of ongoing opportunities to telegraph specific attitudes and more general world views relative to influence that derives from direct observation. For example, in addition to observing a father hitting his spouse or the child, communications within the family might include the need to stand up to bullies or other threats within the neighborhood, or reference other situations in which an aggressive response is needed or at least understandable. This will include not only direct communications to a focal child, but interactions with other family members, siblings, neighbors, and others within the family’s orbit.

Researchers to date have failed to develop nuanced measures of these aspects of the transmission process. For example, in some analyses, direct transmission is considered the variance that cannot be explained once traditional mediators such as parental supervision have been taken into account (Thornberry, Freeman-Gallant, Lizotte, Krohn, & Smith, 2003). Scales developed to measure attitudes often verge on the tautological and provide little in the way of new information (e.g., teens who use marijuana are more likely to affiliate with friends who have positive attitudes toward marijuana relative to their non-using counterparts). Indeed, it is interesting to note that in an early review of the state of differential association, Matsueda offered suggestions for future research, and concluded that “perhaps the most fundamental research problem facing differential association theory involves identifying the content of definitions favorable to crime” (1988, p. 296). Matsueda argued further that these definitions are not oppositional values, but consist of conditions that apply within specific situations and that may be group and historically contingent. In my view, this general critique nicely summarizes the current state of knowledge in this area some twenty years later.

THE WIDER ARENA OF ‘NON-CRIMINAL’ DEFINITIONS

In their discussions of differential association and other social learning theories, Sutherland and subsequent generations of criminologists have tended to hew closely to the world of criminal definitions. This delimited focus seems generally consistent with the field’s raison d’être. Yet even gang members and career criminals spend the majority of their time in non-criminal pursuits (Thrasher, 1927). Thus, a life course perspective on social learning requires attention to ‘lessons learned’ that relate to a range of different content areas that constitute life’s full agenda. This wider world of non-criminal definitions is critically important, as: a) such attitudes and associated behavior patterns have potential to connect to initial forays into delinquent territory, and b) may limit the assumption of traditional prosocial adult roles typically assumed as individuals mature into adulthood. These include areas such as education, work, gender socialization practices, and conduct within romantic relationships (the latter a domain we have focused heavily upon in our own recent research). And while some of these attitudes and associated behavior patterns fall into the realm of preferences, many are influenced daily by an ongoing lack of specific forms of resources and capital that foster more prosocial adaptations with respect to these domains. For example, parents may place a high value on education, but nevertheless ask an older child to stay home from school to care for a younger sibling in the face of a range of family exigencies (Burton, 2007).

FAMILY AS A FOUNDATIONAL WEB OF AFFILIATIONS AND CONDUIT TO OTHER CONSEQUENTIAL RELATIONSHIPS

Parents play a critical role as they position the family within particular neighborhoods with their attendant crime rates and mix of prosocial and antisocial opportunities and definitions. Yet the specific affiliations that develop within the census tracks or even block groups most neighborhood researchers study refine our understanding of the social world a child actually inhabits. In turn, while most studies of intergenerational processes focus on the influence of a particular parent who has participated in a longitudinal study, a wider array of family members are generally part of the child’s world and of the social influence process. In addition, within the contemporary context, this often includes ties formed via increasingly complex relationship, cohabitation, and marital histories (Guzzo, 2014; Lundberg, Pollak & Sterns, 2016; Manning, in press). Further, it is well established that even among samples of adults, individuals involved in criminal behavior tend to have friends and other associates with similar portfolios (Giordano, Cernkovich, & Holland, 2003; Giordano, Copp, Manning, & Longmore, 2020a). Nevertheless, the implications of these affiliations for the social climate children experience has not been systematically investigated. Finally, as young people mature and begin to develop their own social networks, the family continues to matter as a conduit or channel influencing the backgrounds and characteristics of friends and romantic partners (Giordano et al., 2019).

THE ROLE OF AGENCY

Forging a more explicit connection between social learning theory and the life course perspective also underscores the need for additional attention to the role of human agency, and highlights ways in which it may prove fruitful to conceptualize this important aspect of human behavior. Agency is recognized as a core dimension of life course theorizing, even though definitions and treatments of the concept vary considerably. Shanahan and Elder (2002) offered the following definition:

In life course perspective, agency at the level of the person can be defined as the ability to formulate and pursue life plans. A central assumption of this definition is that agency extends across the phases of life, connecting earlier and later experiences through goal-directed behaviors and strategies. As a sociologist, however, a key issue is reflecting both the active efforts of individuals and the social forces of history

(p. 147).

Within the context of more general life course investigations, the idea of affecting one’s own life trajectory has in practice often been equated with individual capacities such as higher levels of self-efficacy or feelings of mastery (Clausen, 1991). Within criminology, the idea of agency has been discussed most extensively in connection with research on desistance. Theorizing that includes discussions of the role of agency developed as a counterpoint to the focus of Sampson and Laub’s early emphasis on the constraining influence of marital relationships and conventionalizing potential of a stable job (Laub & Sampson, 1993). Although the latter remain key anchors of traditional prosocial pathways out of crime, some desistance researchers have incorporated the idea of agency in connection with the need to explain initial moves toward the prosocial, as well as instances in which individuals fail to benefit from exposure to prosocial influences such as job opportunities or entry into marriage (Farrall & Bowling, 1999; Giordano et al., 2002; Paternoster & Bushway, 2009). Recognizing that some degree of intentionality and motivation are likely involved in making significant and sustained life changes, more recent work by Laub and Sampson (2003, p. 38) includes a role for agency. Defining agency as ‘personal choice,’ the authors conclude that “crime is more than a weakening of social bonds—human agency must be recognized as an important element of understanding crime and deviance over the life course.”

Although it is thus becoming more conventional to incorporate the general idea of agency into discussions of the dynamics associated with desistance, conceptualizations within criminology vary as widely as within the larger area of life course studies (Healy, 2013). If we conceptualize agency as ‘personal choice,’ it is a challenge to figure out how to incorporate it into standard empirical assessments. Accordingly, many investigations that include a role for agency have relied on qualitative methods and the content of in-depth interviews to explore this elusive but important aspect of the desistance process. However, some researchers have developed scales for use in survey research that resonate with the basic notion, such as those that index planfulness or ‘readiness to change’ (Williamson, Day, Howells, Bubner, & Jauncey, 2003).

A ‘life course perspective on social learning’ suggests additional considerations that may be useful as we give more research attention to the role of agency in connection with criminal behavior patterns. First, although agency has been invoked most often within the context of studies that rely on interviews with adult offenders, and in connection with discussions of desistance, there is little reason to confine the study of agentic moves to later phases of the life course (Mulvey & Schubert, 2012), or indeed to investigations of desistance. For example, elements of intentionality and motivation are likely required to explain the adaptations of youths who manage to develop along prosocial lines, even though parents’ and other family members’ backgrounds are characterized by criminal behavior (Giordano, 2010). Further, while it is most intuitive to invoke the idea of agency in connection with movement away from crime and in the direction of the prosocial, the same basic elements of redefinition and acting back on the environment figure into, for example, the choice to abandon a drug treatment program after a period of faithful participation. However, this to a degree challenges the idea of agency as a life plan, or incarnations such as planful competence that tend to evoke more prosocial prospects and outcomes.

The social learning perspective brings to the fore an even more fundamental issue regarding agency and its depiction in the literature. Even though our own work has focused on dimensions of change such as ‘cognitive transformations’ and shifts in identity, we have nevertheless endeavored to advance a social view of these dynamic processes (Emirbayer & Mische, 1998; Giordano et al., 2002). Paternoster and Bachman (2017) recently critiqued our view of agency as entirely too social, favoring a highly individualistic theory of agency and the desistance process. While we agree with the idea that acting back on the environment or making a move to change on one level necessarily is a ‘personal choice,’ this depiction is ultimately unsatisfying as a complete explanation as it does not account for: a) the key role of social influences on initial definition and subsequent redefinition processes and the likelihood/capacity to make the desired agentic moves, b) the role of social support and reinforcement that serves to bolster the new pathway, or c) the significant difficulties experienced by many offenders as they strive to sustain the new direction (see Giordano, 2017; see also Weaver, 2016).

Relapse and parole violation statistics attest to these difficulties, and it is not clear that all such derailments can be traced back to a lack of willingness to change, failure to envision a new identity, or even to construct an initial plan for developing a different lifestyle. These same elements likely figure into the role of agency in relation to other dynamics, such as those associated with intergenerational continuity and discontinuity. For example, young people may start off with a strong desire to develop a life that represents a departure from that of their parents, but at some point, absent sufficient prosocial anchors and resources, may evidence some of the same behaviors they initially positioned strongly against. Conversely, those who ‘manage to make it’ (Furstenberg, Cook, Eccles, Elder, & Sameroff, 2000) are likely to benefit from the availability and impact of prosocial connections that help to catalyze and sustain a different life trajectory.

SOME ILLUSTRATIONS OF DIRECT TRANSMISSION DYNAMICS

Early in my career, I was primarily interested in peer relationships (Giordano et al., 1986), but came to focus on the family dynamics outlined above based largely on my involvement with the Ohio Longitudinal Study (OLS) and later the Toledo Adolescent Relationships Study (TARS). The OLS developed initially as an extension of a community-based study that had centered on girls’ delinquency. Out of a basic concern that typical studies of this type often include relatively few delinquent girls, we eventually interviewed the total population of the state institution for girls, and a comparable sample of boys also incarcerated at the state level in Ohio. Later on, we conducted a follow-up of this sample of serious delinquents when the respondents were approximately 30 years of age, and a third set of interviews when respondents averaged age 39. The latter follow-up included interviews with 158 of the pre-adolescent and adolescent children of these female and male respondents. This investigation has been useful for examining a range of different topic areas (Giordano et al., 2002; Giordano, Schroeder, & Cernkovich, 2007), but the data collected from children of the original respondents, including in-depth interviews as well as structured surveys, provided the opportunity to examine dynamics within the family, including aspects of direct transmission (Giordano, 2010).

The interviews elicited from the children provided a unique window on what it is like to grow up in families in which parents and often other family members have significant criminal histories. In connection with that study, the qualitative data were especially useful in highlighting the need to focus attention on the above described dimensions of direct transmission, as well as more traditional family factors such as low or inconsistent supervision and levels of attachment. Yet due to the somewhat unique nature of the sample, including a central focus on the children of women with very atypical backgrounds, it was necessary to consider insights gleaned from the analyses as quite tentative. Thus, we wondered whether interviews relying on a normative sample would reveal different patterns, and suggest a more central role for mediators such as parental supervision. About this time, we began such a study, the Toledo Adolescent Relationships Study, which was community-based, and followed a diverse sample of young people (1,321) interviewed as adolescents, and five additional times across the transition to adulthood. At the wave six interview, nearing completion, respondents are now in their mid-thirties. The interviews share significant measurement overlap with the original OLS study, and the approach to the in-depth interviews was also similar. Further, while we relied initially on school rosters to draw a stratified, random sample that encompassed 62 schools, school attendance was not a requirement for inclusion in the sample. We believe this strategy has been effective in producing a sample that includes young people with significant delinquent histories, while also reflecting a more complete range of variability in family circumstances and child behavior outcomes.

In connection with this larger data collection effort, at most waves we also conducted in-depth interviews with subsamples of high risks youths. As the interviews were transcribed and later analyzed, we were surprised at how often these respondents incorporated family dynamics and circumstances into their narratives. This was especially striking because the TARS study was designed to focus heavily on extra-familial relationships, particularly the role and impact of romantic partners across a range of developmental outcomes. Further, while the first set of TARS interviews was conducted when the respondents were adolescents (when most respondents still lived with their parents), the later interviews were completed when these women and men were young adults. Nevertheless, respondents frequently incorporated aspects of their family backgrounds as central influences on the direction their lives had taken. Of particular note, the themes these respondents developed paralleled many of the emphases we observed in connection with the earlier elicited OLS interviews. This was of special interest because, while these youths tended to be high risk within the context of the TARS sample, we had not envisioned that their circumstances would approach the level of serious risk that was more typical of the family environments of the OLS children. Below I draw on both sets of interviews to illustrate some of the specific dynamics relating to direct transmission outlined above. In some cases, quantitative analyses have proven a useful adjunct to examine the effects of these family dynamics more systematically relying on the larger TARS sample and its longitudinal framework.

DIRECT MODELING

The interviews we conducted with these young respondents clearly reveal direct exposure to violence that occurred within the family context. Thirteen year old, Jason, who participated in the OLS study, described his concerns about his mother’s use of violence:

Every time she comes she always gets mad at my grandma… like she choked my grandma before she went to jail. I’m always like, I don’t want to hit her and stuff but like I have a hammer and I sit it right next to my bed ‘cause I know if my mom comes in messed up on drugs she gets real violent… and she always comes up and rips the phone jack out of the wall and stuff… I just have to be ready.

[Jason, age 13; OLS] (Giordano, 2010, p. 1)

Brandon, a TARS respondent quoted below, includes a reference to violence in connection with a more general discussion of his early home life, while Leigh, another TARS respondent, developed a more pointed description of her mother’s aggressive actions:

My parents are kind of, um, not some people you want to live with. Well, my mom’s kind of crazy. She was married to a guy that smoked crack, so we really didn’t have a lot of birthdays and Christmases and you know it wasn’t a really nice house to live in. There was always fighting and stuff.

[Brandon, age 20; TARS]

I was brought up in a physical family. My mom was very very very very very very physical. She always took it there so I believe that that’s probably rubbed off on me quite a bit so I would always be ready to take it there.

[Leigh, age 24; TARS]

These discussions were incorporated into the much longer narratives in a variety of ways. Jason, who was only 13 at the time of this interview, was simply describing some of the stresses he was currently experiencing. He used the example to make more concrete for the interviewer why he often felt sad and yet somewhat relieved when his mother stayed away from home. Brandon’s matter-of-fact description emerged as part of a discussion about all the difficulties he had experienced as a young adult in securing stable housing (“after the divorce money came through she took off…I haven’t seen her in months”). Leigh’s narrative is of particular interest, as she offered what we categorized as a linking statement, directly connecting her mother’s early use of violence to her own tendency to ‘take it there’ with romantic partners and others. Thus, statements such as those provided by Leigh not only highlight respondents’ awareness of their parents’ use of violence, but go on to describe how this had influenced their own behavior. Such accounts suggest the play of cognitive processes (including the element of identification) in connection with the direct modeling dynamic, rather than simply ‘mindless’ imitation (Sutherland 1947). Jack, an OLS father interviewed in prison, similarly underscores the importance of the identification aspect of the transmission process:

It frightens the hell out of me… I see where he’s going and he’s right here. He wants to run the streets and carry guns and drink and all that. It’s yea he’s going to end up like I am… and I try to tell him but he ain’t trying to hear it. He thinks it’s a joke---he’s been coming to see me all his life. In and out. To him it ain’t nothing. You know he don’t care.

You don’t want your kids to, you know every parent wants their kids to have better, to do better than what you are… but if I’m leading and I’m getting in trouble and then he thinks that’s slick and cool what he’s doing… because I did it. I wouldn’t want him to turn out the way I did. But right now he’s in a lot of trouble.

[Jack, age 36; OLS]

Other narratives reflect on an even more active engagement of parents as part of direct modeling dynamics. Amanda provides an example: “My dad… he taught me to fight like a guy, fight like he does…no, he’s taught me how to fight my whole life” [Amanda, age 19; OLS] (Giordano, 2010, p. 144). This idea of more active training can occasionally figure into acquisitive crimes, although this appeared less common relative to ‘lessons’ related to fighting and other instances of aggression. Below are two examples, one referencing a parent who sold marijuana, and the second a respondent who got caught up in his father’s burglary scheme:

My husband… he would go to Texas, he’d get pounds. And a lot of times he would make my son clean them. And clean the bits out, extend them and everything.

[Viv, age 38; OLS] (Giordano, 2010, p. 145)

Anthony described how his father had asked if he wanted to ‘work a job’ with him… he believed that this was legitimate plumbing work, but instead the job involved stealing copper pipe from a construction site.

[Anthony, age 25; TARS] (Copp, Giordano, Longmore, & Manning, 2019a, p. 25)

Although such examples can be found in both the OLS and TARS interviews, references to parental substance use were considerably more common within the narratives of the high-risk youths who participated in both studies. Based on our work with the narrative accounts elicited across these two very different sample groups, it is exposure to parents’ and other family members’ substance use that most clearly challenges Hirschi’s ‘hidden deviance’ notion.

As noted above, various strands of research have documented the connections between substance use and involvement in the criminal justice system. Consistent with Wakefield and Apel’s (2018) observations, these young people were generally aware of parental absence due to incarceration. However, the data we analyzed make clear that substance use issues often connected to criminal justice contacts and/or incarceration also frequently figured into the nature of the child’s family experiences (Giordano et al., 2019). Further, prior research has shown that substance use problems inhibit the odds of desistance directly (Huebner, DeJong, & Cobbina, 2010; Schroeder, Giordano, & Cernkovich, 2007) and affect crime patterns indirectly via a negative impact on the stability of marriage and maintenance of other prosocial ties (Homish, Leonard, & Cornelius, 2008). These collateral consequences of substance use thus contribute, along with the direct negative effects of parental incarceration, to the length of child exposure and the character of the child’s home life (e.g., parental IPV, breakups, changes in residence). Other research findings telegraph the potential impact on child well-being in even more obvious ways, as for example, research indicating that parental substance use places children at increased risk of abuse and neglect, and is often a factor associated with referrals to child protective services and foster care. References to parents’ substance use vary from a general awareness of the parent’s problems to recollections of using drugs together:

Yeah, I’ve seen it, pulled it out. Because I had snuck in the kitchen. Because…so I could spot him. So he would quit lying about it…

[Robert, age 11; OLS] (Giordano, 2010, p. 142)

I really don’t want to picture my mom using because she had this side where she would look, like, so good when she was clean. But then when she got, when she used, she just looked so tore up.

[Josie, age 16; OLS] (Giordano, 2010, p. 141)

The OLS respondents were all adolescents when interviewed, and thus their descriptions reflect the immediacy of their experiences. As the TARS interviews occurred during adolescence as well as later on when respondents were young adults, older respondents’ interviews necessarily became more reflective about a longer span of time. Danielle, for example, pointed out that eventually she and her mother started drinking and getting high together: “She always just stayed away so she could party, you know, and I’m like, well I can party too…if you’re going to do it, why not do it together. So that’s just how it was” [Danielle, age 24; TARS]. Another TARS respondent Emily not only referenced early childhood problems related to her mother’s substance use, but offered a longer window on these and related circumstances that had continued up to the present day:

We were driving down the street, she had just got out of prison, so she was clean, she didn’t have no drugs in her system, no alcohol no nothing… [Emily’s mother told her that she was proud of her for getting her GED and a job]…It’s like you know she will do drugs, or you know she has no respect for herself, but yet, she sets my standards high, like I should be doing that…. And well that’s what you’re supposed to be doing but yet she don’t do it. And it’s just like. She’ll cry and just the drugs and the fast life and the men, beating her up. If she gets beat up by her boyfriend, the first person she’s calling is me. Crying, he just beat me up….and its my mother. I love her, and I still care for her, so I’m gonna run to her and try to comfort her. That’s my mommy. You know what I mean, but I’m getting to the point, where I can’t deal with it no more. It’s starting to be a burden on me. I’m losing sleep over my mom.

[Emily, age 24; TARS]

The above description incorporates emotional reactions as well as facts about her life circumstances (for a more detailed examination of the role of emotions in the transmission process, see Giordano, 2010; Giordano et al., 2007). These examples are particularly useful in illustrating the complexity of children’s responses as well as the longer life course lens. On one hand, many of the young people describe negative emotions and include rather harsh judgments. At the same time, Emily makes clear that she loves her mother (that’s my mommy) and worries about her, comments that challenge the idea that ‘lack of attachment’ arguments fully capture the dynamics that have and at times continue to unfold within these disadvantaged families.

THE WIDE-ANGLE LENS

As noted above, most longitudinal studies focused on parental effects logically zero in on the parent who has participated in the study since its inception, and/or the individual who is the primary caregiver. However, this strategy does not provide a comprehensive window on family climate and thus the dynamics associated with direct transmission. Prior studies have shown, for example, that individuals in prison often report that they have additional relatives in prison or with previous criminal justice experience (Mumola & Karberg, 2006). Nevertheless, these social network patterns have not always been considered from a social learning standpoint. To illustrate, Wildeman and Wakefield (2014) noted this tendency, but focused primarily on the additional stress experienced by families who have multiple family members with incarceration histories. Thus, the qualitative interviews are particularly helpful in widening the lens to consider the frequent opportunities for influence presented by other members of the family. For example, Brandon, cited above, mentioned a former stepdad who ‘was on crack’ and constantly fought with his mother. Rebekah, another TARS respondent quoted below, not only focused on her stepfather, but an array of other family members, as she offered a clear linking statement connecting this family exposure to her own later aggressive behavior. In turn, Marcus, another TARS participant, provided specifics about social support from an uncle, but also lessons that helped him to comfortably navigate the time he spent in the county jail:

It, just my family’s just really aggressive and outspoken and just mean and they’re always fighting. Literally at every function that we have, somebody’s fighting…and then with my [step]dad… I kinda blame him because maybe if she woulda picked somebody better I woulda turned out better. I think watching them fight all the time made me violent.

[Rebekah, age 26; TARS] (Giordano et al., 2019, p. 413)

Ended up there with him [uncle] on the sixth floor. The sixth floor was like the harder criminals but it was cooler because most of them were white. My uncle has been in prison his whole life. He’s a real old guy and he’s just known around jails and stuff…. everybody knew him and I got right in then. He asked if I needed another blanket and food [uncle took the items from another inmate]. So after I got with him it was pretty easy the rest of it.

[Marcus, age 22; TARS]

Sibling concordance has been repeatedly documented in criminological studies, and it is unlikely that these similarities all stem from the shared influence of inconsistent parenting practices or even genetic predispositions. To illustrate, Jeff, quoted below, conveys the impact of an older sister and her boyfriends’ actions on his own later treatment of romantic partners:

I grew up with, you know, my sister and her ex a lot and they were abusive, you know, they talk to each other in certain ways, yea like beating each other or whatever, and I guess I was around them a lot so I must’ve…picked it up in the back of my mind and stored it there and one day just used it I guess… And maybe consciously, picking up that’s how I need to treat a girl, ‘cause he obviously had a girl, and he was keeping her, and that’s what I wanted.

[Jeff, age 26; TARS] (Giordano et al., 2019, p. 413)

THE LONG-FOCUS LENS

As the above examples illustrate, the in-depth interviews are valuable as they can increase our awareness of certain dynamics, including direct transmission processes that trace to the wider circle of family members. However, the life course perspective inevitably highlights the need for a longitudinal perspective. In this regard, most qualitative studies based on interviews or observations focus on a particular slice of time, or retrospective accounts also elicited within a given time frame. Thus, for example, some individuals who had desisted, may have focused on identity shifts and other ‘cognitive transformations,’ but in later follow-ups—in spite of apparently sincere intentions to change their lives—they go on to reoffend (Halsey, Armstrong, & Wright, 2017). Similarly, while the OLS study did include a longitudinal component for the respondents who were the parents, we were only able to conduct one set of interviews with the teenage children of these original respondents. Thus, we have often speculated about how particular children had fared over the long haul. For example, Jason, cited above (see the first illustration quoted) was a young man of 13 at the time of the interview, actively grappling with his mother’s substance use and tendency toward violence. Yet he figured into the ‘success stories’ chapter of the monograph about these OLS children nevertheless:

I’ve seen her pass out and stuff fall out of her pockets. I flush it down the toilet and I throw it as far as I can. And, like, I spit on it, and if its weed or stuff, I flush it down the toilet. If its pills I put them in hot water…then I put it down the sink ‘cause I know she’ll drink it… She’s so messed up on drugs… she like takes pipes off the sink and starts sucking on them to get drugs and stuff…and she always looks for wrenches and stuff. [later] Everybody expects me to be just like my mom…[but] I just ain’t like my mom.

[Jason, age 13; OLS] (Giordano, 2010, p. 144)

In analyzing his lengthy narrative, we focused on Jason’s aversion to drugs, strong statement of a desire for a different kind of life, and the development of a contrasting identity (everybody expects me to be just like my mom…but I ain’t just like my mom). However, we did not have the resources that would have been required to launch a systematic, statewide search of each of these teen respondents as adults. Nevertheless, to update us on Jason’s adult life circumstances, we relied on a simple internet search and later conducted a search of official records. Even based on a rudimentary google search, we located the following entry in Jason’s local newspaper under the heading “indictments returned during a 2019 session of the X County Grand Jury”: “Jason _____ indicted on two counts of possession of cocaine, and one count each of illegal use or possession of drug paraphernalia, aggravated trafficking and aggravated possession.” A subsequent records search accorded with this unfortunate portrait of intergenerational continuity: as a 28 year old, Jason had already been incarcerated at the state level, and was currently in jail awaiting the resolution of the above charges.

The lesson for this researcher (and for building realistic theories of criminal continuity and change) is that social psychological factors are only one piece of the puzzle for understanding long-term pathways in and out of crime. Clearly, an array of other disadvantages, including features of the individual’s intimate network, may at times swamp the impact of positive personal perspectives and resources. Within the context of prior analyses, we have tried to emphasize some of the limits of human agency, and the role of these broader social and structural forces (Giordano et al., 2002). Yet a theory that focuses on cognitive transformations inevitably seems to turn the spotlight on individual-level factors. Nevertheless, the example of Jason’s life history accords well with the thrust of the ‘continuity’ quotes also included above, and the difficulty I faced finding OLS children who could reasonably fit into the monograph’s ‘success stories’ chapter (Giordano, 2010, pp. 163-204). Collectively, such research experiences support the need to focus additional research on family and other disadvantage factors that favor intergenerational continuity, as well as dynamics associated with resilience. The TARS study, as it relies on a community-based sample, provides a better mix in terms of intergenerational continuity and discontinuity (see e.g., Johnson, Manning, Giordano, & Longmore, 2015). Yet it is likely that the experiences of OLS study participants would line up very well with those of individuals whose parents have had significant exposure to the criminal justice system.

THE TRANSMISSION OF ATTITUDES: THE EXPANSIVE WORLD OF INTERACTION AND COMMUNICATION

It is intuitive to focus primarily on the behavior-to-behavior linkages that are most readily evoked by the idea of direct modeling. Yet as Sutherland emphasized, the opportunities presented by the recurrent interaction that occurs within intimate primary groups greatly expand the child’s understanding of definitions favorable or unfavorable to the enactment of these behaviors. As noted above, this wider arena of attitudes has proved elusive to capture and measure systematically on standard surveys. One problem with studying crime is that most people do not exactly endorse—at least in a blanket fashion--these types of actions (but for a nice example of an exception see Katz’ (1988) discussion of gang youth reliving their exploits). This has been an issue for researchers who study a range of behaviors that fit under the crime umbrella, but the general point is particularly well illustrated by research on intimate partner violence.

Scales measuring the acceptability of using violence against an intimate partner generally reveal low rates of endorsement, even though IPV remains a widespread problem, as reflected in both self-report studies as well as official statistics (Copp, Giordano, Longmore, & Manning, 2019b; Mumola & Karberg, 2006). The issue of transmission of “definitions favorable” is even more complex when the focus is on parents, who undoubtedly do not want their children to become delinquent, come in contact with the criminal justice system, take drugs, or hit their partners. Yet communications may telegraph “conditions under which” and nuances that are not covered by general wishes or admonitions. To illustrate, at one point during the middle of both the OLS and TARS structured surveys completed by parents, we included a single question (How do you feel about your child dating?) that provided for open-ended responses. Unlike the qualitative interviews, then, we received over 1,200 of these TARS responses and 126 from OLS parents. These answers to what was essentially a throw-away question turned out to be extremely illuminating—reflecting widely different attitudes and providing a small window on this wider world of definitions. A surprising number of parents even answered the question as if they were communicating directly to their child:

I think dating is ok and part of the teen experience.

[Maureen, age 46; TARS] (Giordano, 2010, p. 159)

Your boyfriend is ugly—and you should watch how he treats you. If a man puts his hands on you, pick something up and knock him out. Take nothing from him.

[April, age 26; OLS] (Giordano, 2010, p. 160)

Based on such responses and the content of the longer in-depth interviews, additional research is needed that relies on nuanced items and scales that tap “conditions under which” x or y may be necessary or at least ‘understandable.’ Such conditional questions are likely to capture evident variability more effectively than assessments regarding a behavior’s general acceptability. Specific behavioral referents may also add to an important line of research that has focused on global dimensions of criminogenic attitudes, beliefs, knowledge, and general assessments (Agnew & Messner, 2015; Simons & Barr, 2014).

ATTITUDES IMPARTED VIA REACTIONS ‘AFTER THE ACT’

Although my own research has focused heavily on precursors of criminal behavior, Akers’ (1998) version of social learning theory is important in highlighting that reinforcement patterns (essentially what happens after an act has been committed) are integral to the learning process. The following example illustrates multiple levels of potential influence, including such ‘after the act’ social dynamics. Kevin, a TARS respondent, described an incident in which he had hit his girlfriend Trish, and given her a black eye. However, his discussion also encompassed an account of the aftermath of his actions. Trish immediately called her parents, who came over that same night:

Her mom thought that her dad was gonna come down and just wanna fight. But the day before, her dad beat the shit out of her mom and just got out of jail for it. So he came to the house and he said, what were you drinking? And I pulled the Jose [Cuervo] out and he said that shit will do it to you. And then I explained it to him… and he said, Trish don’t punch on this man like he’s a punching bag, you can only punch on a man so much before he hits you, and she was like “I know dad.” And he was like Kevin, I believe everything you’re saying but next time it happens we will have a altercation. And I was like I understand that but she was hitting me and you know that Jose…

[Kevin, age 26; TARS]

This example is useful in simultaneously illustrating the role of the wider circle (her parents got involved, not his), direct modeling (Trish’s father had just gotten out of jail for hitting Trish’s mother), and these ‘after the fact’ definitions. Although Trish’s father expressed general disapproval, Kevin was surprised at how understanding he was when Kevin recounted the circumstances that had led to an escalation of the disagreement. Indeed, Trish’s father’s line of questioning went immediately to and thus reinforced well-worn justifications for IPV, including the disinhibiting effects of alcohol (“that shit will do it to you”), and the role of his partner’s actions (she started it). At the end of the discussion, Trish’s father did suggest that if Kevin were to do something like this again “there will be a altercation.” On the one hand this provides a general admonishment that suggests negative reinforcement of such behaviors. On the other, this warning ultimately telegraphs that when all else fails, violence may be the only viable strategy for handling disagreements.

BEYOND THE ORIGINAL FRAMEWORK OF DIFFERENTIAL ASSOCIATION THEORY

Learning like any other?

Sutherland’s propositions included the idea that the basic dynamics involved in learning crime are identical to those involved in learning any other form of behavior (Sutherland, 1947). While this has been a very useful underpinning of his theorizing, there is a need to offer some caveats to this general tenet. As all parties to the learning process, including children, generally come to recognize the negative aspects of their parents’ problem behaviors, they often focus on the normal-but-not-entirely-normal aspects of such family experiences:

You know other people with their moms would never do that…

[Kaley, age 19; OLS] (Giordano, 2010, p. 146)

Why steal from your own son?

[Jason, age 13; OLS] (Giordano, 2010, p. 145)

She (stepmom) should have been protecting me, but she just said put it [crack] in the pipe and here you go.

[Brittani, age 18; OLS] (Giordano, 2010, p. 153)

Some things that I grew up with ruined my life because I saw what my dad do…

[Todd, age 22; TARS]

These reflections highlight feelings of anger and demoralization that would not as often accompany the process of learning more neutral or prosocial behaviors. Such expressions accord well with Agnew’s (1992) focus on strains and the role of anger as a potential mediator influencing the odds of adopting a delinquent repertoire. However, social learning theory adds a significant component in highlighting how it is that individuals often come to cope with these and other negative circumstances in the particular ways that they do:

When she passed, I started using more to block out the pain, you know, not think about it. It really tore me up and I’m lucky to be alive because that’s when I started injecting heroin… It ain’t that I thought it would be different [from what happened to his mother]. But that was the best way of blocking out the pain for me…. Once I lost her, I felt like I lost everything. It’s not that I thought it would be different—I just didn’t care.

[Chad, age 24; TARS] (Giordano et al., 2019, p. 412)

And when I had the abortion, the first thing she gave me was a fat bud. She was like smoke this you’ll feel better.

[Kaley, age 19; OLS] (Giordano, 2010, p. 146)

Non-criminal definitions and pathways in and out of crime

Focusing on non-criminal definitions may appear to be outside the purview of a field that is concerned primarily with the precursors and consequences of criminal behavior. And indeed Sutherland’s own emphasis on the importance of an excess of criminal definitions is consistent with the field’s delimited focus. Yet if we agree that the business of “life” encompasses much more than crime, definitions and behaviors that connect to these other life course experiences are nevertheless likely to be directly and indirectly implicated in criminal pathways. Recall the example above in which a parent offered the advice that if a man hits you, it is acceptable to pick up something and “knock him out.” Such communications make a clear connection to violence itself. However, the OLS and TARS interviews reveal attitudes that cover a range of aspects of life, including the world of romantic relationships, which do not include explicitly criminogenic definitions:

Boys will use you then leave you.

[Ava, age 36; OLS] (Giordano, 2010, p. 159)

You can’t believe a word that a man say.

[Karen, age 37; OLS] (Giordano, 2010, p. 160)

My dad has played women his whole life.

[Mackenzie, age 28; TARS] (Giordano et al., 2019, p. 414)

While these statements do not contain any references to criminal behavior, the attitudes and potential ‘lessons’ contained within them may nevertheless foster world views such as cynicism and distrust of the opposite gender that affect conduct within the romantic realm. In turn, this broader set of attitudes and associated behavior patterns may set the stage for behavior that is criminal (destructive conflict within romantic relationships) or more generally limit the conventionalizing potential of close romantic ties (i.e., consistent with the finding that duration of romantic relationships is generally associated with desistance from crime see e.g., Laub, Nagin, & Sampson, 1998). In an attempt to tap this broader realm of attitudes, we developed a structured scale indexing “parents’ negative/cynical attitudes toward dating and sexuality” that was included in the parent questionnaires. Subsequently, results of a longitudinal analysis relying on the TARS data indicated that in models controlling for parental IPV and other traditional predictors, higher scores on this parental attitudes index were associated with increased odds of infidelity and IPV as reported by their young adult children some ten years later (Giordano, Johnson, Manning, & Longmore, 2016).¹

Certainly infidelity is not in itself a crime, and IPV relates directly to the dating world. Yet recent analyses demonstrate how these broader relationship-specific behaviors may also be implicated in more general patterns of criminal behavior. In another analysis that relied on five waves of TARS data, we focused on variability in self-reports of criminal behavior across the full study period. Results show that net of traditional social learning constructs (partner’s level of criminal behavior; parents’ antisocial behavior as a between subjects factor), and variables indexing control-based dynamics (e.g., attachment to partner), IPV and infidelity were significantly related to within-individual variability in crime as reported across the five waves of assessments (Giordano, Copp, Manning, & Longmore, 2020b). Such analyses thus provide evidence that these other kinds of relationship-specific behaviors ‘matter’ for understanding life course patterns of criminal continuity and change.

Agency: learning as we go

The idea that the individual is a passive ‘vessel’ who behaves according to the balance of definitions favorable and unfavorable to the violation of law has been amply criticized (Box, 1971; Taylor, Walton, & Young, 2013). Clearly, individual differences and perspectives come into play. For example, an intellectually gifted child may latch onto school as an important pro-social anchor, in spite of an early criminogenic family environment. Yet beyond these personal resources and linked experiences, agency reflects the human capacity to ‘put it all together’ in unique ways, and even to make a range of course corrections. This potential to push back on the environment is related to the cognitive and emotional underpinnings of human conduct, including the ability to reflect on the past, imagine the future, or fear repercussions. The longitudinal findings described above, then, provide a small indication of the idea that the life course consists of a continuous process of learning and adjustment. While parents’ criminal behavior and IPV were significant risk factors for the related young adult behaviors assessed, the analyses also revealed considerable variability in the children’s relationship-based actions across the adolescent to adult transition period, as well as in their self-reports of criminal behavior and IPV.

Most researchers accept in general terms the idea that individuals possess the capacity to make agentic moves. Yet we lack a complete understanding of why, when, and how they choose to effect small and major life adjustments. We have previously argued that cognitive and emotional processes are integral to an understanding of such personal transformations (Giordano et al., 2002; Giordano, Copp, Longmore, & Manning, 2015; Giordano et al., 2007). Yet if new thoughts and emotions come to the fore as impetus and guide to action, is this best understood as an individualistic process and an individual accomplishment? (see Paternoster & Bushway, 2009). As a contrast to this theoretical position, Mead (1938) and other symbolic interactionists contend that thoughts and emotional experiences are inherently social. This general point would seem to apply to the redefinition process as well as the process of acquiring one’s initial complement of definitions. The following quote from Mike’s narrative nicely illustrates that individuals must ultimately ‘own’ the decision to move away from an earlier pattern of criminal behavior; yet, as his account underscores, interactions with intimate others have served both as catalysts and support for the redefinitions associated with these changes in perspective:

I mean I can’t blame nobody else but myself…but I’m trying to get my life together… just the stuff that I went through, just making all types of crazy mistakes and learning from them, knowing now that that stuff wasn’t the right thing to do. Yet it was a lot of people. Um like people might say you’re a good person, you know…like you don’t need to be doing that type of stuff, you know, it don’t fit your description, like you should be in college somewhere. It’s just like little stuff that they would say to me and I would be like you know what you’re right. [Who?] …like my family, a couple of my friends, the ones that were going down the right path. They would say like you need to get a job somewhere or work with me, just anything to help, to turn me into a positive direction. I guess I heard so much I really started telling myself that they were right, you know. I’m kind of doing the wrong thing. I need to change. I can’t keep doing this.

[Mike, age 23; TARS]

IMPLICATIONS FOR THE LIFE COURSE PERSPECTIVE AND SUGGESTIONS FOR FUTURE RESEARCH

Differential association and later versions of social learning theory highlight the centrality of definitions as underpinnings of human behavior, including criminal conduct. This includes the process of learning initial definitions within the family context as well as new ‘definitions of the situation’ acquired via later social experiences. This focus provides a counterpoint to control theory, a perspective that has been important to research on family influences on delinquency, as well as to the study of the dynamics associated with desistance. The latter research tradition is particularly salient, as to some degree the life course perspective itself has become intertwined with this perspective. This owes largely to the seminal contribution of Sampson and Laub (1993) to the development of the life course perspective within criminology, and their focus on the age-graded theory of informal social control in connection with it. My own views have undoubtedly been influenced by disciplinary background and training (sociology; symbolic interaction), but also by the ‘continuing education’ afforded by my research on intergenerational transmission (Giordano, 2010), investigations of desistance processes (Copp et al., 2019a; Giordano et al., 2002; Giordano et al., 2007), and a broader agenda focused on intimate social ties (Giordano et al., 2003; Giordano, Longmore, & Manning, 2006). All of these research experiences reinforce the basic notion that control and even attachment processes, while generally protective, are not nearly enough—either as explanations of pro-social conduct, or of the dynamics associated with criminal onset, continuity, and change.

Research on the family dynamics associated with intergenerational transmission in particular highlights the need for additional study of ‘ongoing conditions’ as well as on the major events that have tended to dominate the action within the life course perspective. We recently argued, for example, that while parental incarceration is correctly viewed as a corrosive and often traumatic event for children, the broader family context, including parents’ and other family members’ involvement in criminal behavior, are also significant influences on children’s behavior and well-being (Giordano et al., 2019). Dynamics such as those outlined and illustrated above are important formative influences, and to an unknown extent may affect reports of delinquency and other negative child outcomes that are often attributed specifically to the parental incarceration period itself. Research is thus needed not only on the specific features of incarceration events (e.g., whether jail or prison, duration, timing in relation to the child’s age), but on the seriousness, timing, and duration of parents’ criminal involvement. This requires further integration of the typically distinct research traditions focused on incarceration effects and more basic research on life course patterns of crime, criminal career trajectories, and desistance (Uggen, 2013). There is thus a need to focus on parents’ behaviors as more than a control to be ‘netted out,’ in order to investigate the complex links between criminal justice exposure, life course patterns of criminal behavior, and child outcomes. For example, researchers studying effects of parental incarceration have often centered on mediators such as economic costs and marital instability, but the impact of parents’ incarceration experiences on later substance use problems or crime patterns has not often been incorporated into these analyses.

Taking a life course perspective on social influence processes also underscores the utility of investigating intergenerational ties and their effects across multiple phases of the life course. Criminologists have begun this work, as for example, Siennick (2011), who examined economic assistance to young adult children, and how the child’s delinquency involvement influenced levels of assistance. Yet focusing on parents’ own criminal histories is likely to develop an increasingly complex portrait. For example, Anthony, a respondent cited above who had committed a burglary with his father (and both were subsequently incarcerated for this offense), went on to reside at his father’s house after his release. On the one hand, this indicates the presence of needed social support. On the other, records indicate that this may not have been the most favorable environment for making a life change, as Anthony was later arrested for another burglary carried out with his father’s next-door neighbor.

Other illustrations show that when parents garner long criminal histories, children may need to step in with housing and other forms of assistance for them, which as Emily indicated, was a source of considerable spillover stress within her own life. The in-depth interviews also suggest that long-term consequences of the parent’s lifestyle and disadvantaged circumstances may result in parents’ declining health or even early mortality (see Chad’s example above; see also Stenbacka, Moberg & Jokinen, 2019). These are close-in experiences that may compound stresses that relate more directly to young adults’ own social and economic circumstances. Examples adolescent and young adult children provided of using drugs with their parents or dating their father’s drug dealer represent extreme, but even more straightforward applications of the differential association perspective. Conversely, it would be useful to determine, relying on multiple measures of parents’ behavior, whether children of parents who go on to exhibit a pattern of sustained desistance, experience reduced odds of criminal involvement and other problem outcomes, relative to those whose parents continue a longer term pattern of criminal activity or substance use.

Consistent with our focus on the ‘wide-angle’ lens as well as the vantage point afforded by longitudinal investigations, research is needed that captures variability in family climates as represented by more complete assessments of family ties. To date, much of the work on social networks has focused on peer affiliations (Haynie, 2002; Haynie & Soller, 2014; Kreager, Rulison, & Moody, 2011), and to some extent romantic and sexual networks (Bearman, Moody, & Stovel, 2004). Although it would be impractical to interview a full roster of each individual’s list of family members, self-reports and official records searches would add to studies that focus primarily on parental behavior measures (see e.g., Farrington, Barnes, & Lambert, 1996). Data sets that include information on siblings’ behavior profiles and that include such measures in analyses would help to round out our understanding of family climate. Perhaps the biggest challenge for capturing the reality of the child’s family context is that instability and changes in family circumstances are hallmarks of the populations in which criminologists have an interest. Thus, for example, stepparents, whether official or based on cohabitation, half siblings, and parents’ former dating partners may move in and out of the family scene, and some of these contacts may qualify as influential, intimate others. This focus on influence dynamics adds to the idea that the instability itself is often associated with a range of negative developmental outcomes (Cavanagh & Fomby, 2019).

Additional research is also needed that would allow us to develop more complex portraits of the gendered aspects of these social influence processes. As Sutherland outlined a very general conceptualization, researchers have relied on this perspective and a variety of other traditional and feminist frameworks as background for research on girls’ and women’s offending, gender differences in risk factors and consequences, the dynamics associated with desistance, and unique issues associated with successful reentry (Chesney-Lind & Pasko, 2013). Some social dynamics (e.g., involvement with delinquent peers) are sufficiently basic that they have generally been shown to be significant predictors of delinquency involvement regardless of gender (Alarid, Burton, & Cullen, 2000; Giordano, 1978). Nevertheless, researchers have shown that effects of cognitive processes such as moral beliefs and other social network differences continue to complicate the picture (Mears, Ploeger, & Warr, 1998; Piquero, Gover, MacDonald, & Piquero, 2005).

Such differences come into particular focus when we consider the family as a source of criminogenic definitions, since daughters are less likely than sons to report delinquency involvement, even when they have been exposed to the types of family circumstances we have described above and in earlier analyses (Giordano, 2010). Certainly the same influencers can impart different definitions, depending on the receiver, as much research on gender socialization practices attests. Differences in supervision and other forms of informal social control also come into play (Heimer & DeCoster, 1999). Nevertheless, antisocial family background has been shown to be a significant predictor distinguishing the smaller subset of young women who do become involved in delinquency/crime from their more conforming counterparts. (Belknap, Holsinger, & Dunn, 1997; Giordano & Copp, 2019). Some of the suggestions for developing the life course perspective on social learning could help us to further illuminate gendered and more generic aspects of social influence processes. For example, following up on the idea of assessing antisocial behavior across the wider circle of family members may reveal that young women involved in delinquency and crime are more often, relative to their male counterparts, fully ‘encapsulated’ in criminogenic social contexts. The life course perspective also requires further examination of the family’s role as channels or conduits that influence later ties with peers and romantic partners. Another hypothesis is that, following principles of identification, young women may be more directly influenced by maternal than paternal substance use or other forms of criminal involvement.

A significant theme within the literature on gender and social influence is the idea that male partners are more pivotal to understanding girls’ and women’s offending patterns. This set of relationships deserves additional research attention as well. The idea of negative influence of male partners accords with analyses based on qualitative investigations of the lives of delinquent girls and women offenders (Barlow, 2016; Giordano, 2010; Richie, 1996), but results of quantitative analyses are somewhat inconsistent about the impact of gender on the nature of partner influence (Cauffman, Farruggia, & Goldweber, 2008; Haynie, Giordano, Manning, & Longmore, 2005). The disparate portraits that emerge from qualitative and some quantitative assessments thus present a puzzle that warrants further research attention. And while social influences on crime typically focus on increased risk, a life course perspective points up the need for research on the role of intimate others as catalysts for making positive life changes. In line with a symbolic interactionist perspective, research is needed that focuses directly on women’s and men’s ‘agentic moves’ to distance from ‘bad companions,’ partners, or even family members who are seen as negative influences on their chances for affecting a pattern of sustained desistance.

Finally, as noted at the outset, a central distinguishing feature of differential association theory is its focus on ‘definitions,’ and yet our conceptualization and measurement of them remains surprisingly primitive. Additional qualitative research will likely be most helpful to the process of identifying specific cognitive underpinnings of criminal behavior (including attitudes parents impart), recognizing that many of the behaviors we study do not routinely receive blanket endorsements. In-depth interviews would provide a basis for developing scales for use in structured surveys administered to parents as influencers as well as to respondents whose behaviors we are scrutinizing. Although criminologists will likely be reluctant to explore too far afield from directly criminogenic attitudes, our analyses have shown that attitudes and behaviors relating to other aspects of life figure into life course patterns of criminal behavior nevertheless. We focused on the world of romantic relationships, given TARS’ depth of measurement of this aspect of respondents’ lives, and the importance of these relationships as individuals mature into adulthood. Yet our understanding of specific attitudes related to other key domains (e.g., engagement with/success in academics) could also be further developed, again recognizing the value of conditional statements (i.e., most parents want their children to do well in school). And, following from a life course perspective, research is needed on attitudes and sequences that influence derailments after a period of forward progress, as well as changing definitions that support a long-term pattern of desistance. In short, it is one thing to argue that cognitive transformations are central to the desistance process, but research is needed on the specific shifts in perspective that undergird long-term successful behavior change.

POLICY IMPLICATIONS

Those who work with offenders generally have only the individual at hand. Accordingly, program content (consistent with the entire thrust of the criminal justice system) often stresses the need for individuals with delinquent or criminal histories to begin to take responsibility for their own actions. This fits well with programs that focus on individual-level deficits that have impeded the process (e.g., the idea of ‘cognitive distortions’). The current flurry of research on biological factors reinforces the individualistic emphasis, although to be sure, this line of inquiry provides a bit more of an off-ramp in terms of personal responsibility. Sutherland’s original focus on social influences on criminal behavior was developed as a counterpoint to the individualistic emphasis of psychological and biological approaches of the day. Recognizing that contemporary researchers have developed more nuanced approaches to our understanding of these individual-level factors, theory building necessarily involves foregrounding the pathways considered most pivotal to an understanding of criminal behavior.

A large literature has now accumulated that highlights the role of disadvantage and neighborhood factors in the observed patterns of crime and violence. Within these disadvantaged contexts, however, close-in social influence processes further distinguish from their similarly situated counterparts those individuals more likely to evidence criminal behavior. Recall West and Farrington’s (1977) finding that a small percentage of families in their study accounted for almost half of the criminal convictions in that sample. This basic finding and more recent studies underscore the need to increase support for family based approaches. Our own studies of intergenerational transmission processes suggest that in addition to providing parent training in, for example, the value of consistent supervision, attention should be directed to specific communications and how they may directly or indirectly foster delinquent adaptations, violence, or substance use. Children of incarcerated parents should be a high priority for the provision of a full range of services, as they represent a nexus of poverty, antisocial influences, as well as the considerable burdens posed by the incarceration periods themselves. Although we have argued that the incarceration effects literature has not given sufficient attention to the broader family context that may accompany the experience of a parent’s incarceration, the policy implications tend to converge. There remains a critical need for services for children whose parents have criminal justice exposure, as well as for their parents and other family members. Ultimately, it will be necessary to move away from the heavy reliance on incarceration and toward the wider use of alternatives such as high quality drug treatment and housing services that target underlying stresses and problems that confinement is ill-equipped to address.

While early family intervention represents a programmatic ideal, those who work daily with delinquent youth and adult offenders more typically will only have access to the adjudicated individuals themselves. Nevertheless, approaches that recognize some of the social influences on criminal behavior and specific adaptations (“hostile attributional bias”) may increase receptivity to change efforts, relative to positioning these attitudes and world views as individual psychological states or personal deficits. A strong social emphasis would also serve to bolster the other end of intervention efforts (i.e., chances for affecting a pattern of long-term desistance). A personal commitment is considered the sine qua non of behavior change, but, as we have argued in prior work, social factors are intimately related to the success of the individual’s own desistance efforts. For example, in a recent study of young adult desistance, affiliation with delinquent friends was independently linked to within-individual variability in self-reported crime across the study period, even after identity changes had been taken into account (Copp et al., 2019a). And, while intimate partner violence is an area where taking ownership and responsibility for one’s actions is especially important, practitioners will need to incorporate that research has nevertheless shown that partner characteristics and couple-level dynamics may also figure into conflict and conflict escalation (Giordano et al., 2015; Moffitt, Robins, & Caspi, 2001).

Parents and other family members represent an especially complex set of social relationships that must be navigated across every stage of the life course. Traditional health promotion messages—consistent with theories of crime that focus on attachment processes—emphasize the positive effects of close ties and social support on physical and emotional health. Yet the social learning perspective and research on intergenerational risk patterns suggests that where the referent is criminal behavior, at times the ability to distance from specific intimate others may well be the most successful option for accomplishing successful behavior change. This will accordingly require the provision of sustained, concrete resources and support that many times family members are the only ones willing to provide.

Biography

Peggy C. Giordano is an emeritus distinguished research professor of sociology at Bowling Green State University. Her research is centered on basic social network processes and the ways in which close relationships influence delinquency and criminal behavior over the life course. She is especially interested in family dynamics associated with the intergenerational transmission of crime and other problem outcomes, as well as in the impact of extrafamilial influences such as peers and romantic partners.