Peer affiliation and, in return, “peer influence” are regarded as hallmarks of adolescence (Brechwald and Prinstein
2011). Indeed, the finding that delinquency of peers predicts adolescent delinquency has been consistently replicated (Brechwald and Prinstein
2011). However, it has rarely been investigated whether delinquency of peers is also associated with other correlates of delinquency such as short-term mindsets (e.g., impulsivity and lack of future orientation). Furthermore, it is often presumed that delinquent peer affiliation implies that adolescents are also committing delinquency together
with their peers (i.e., co-offending) (Brechwald and Prinstein
2011; Warr
2002). In fact, some theorists posit that engaging in delinquency with peers is a primary mechanism that causes the progression of delinquency during adolescence (Warr
2002). It can be extrapolated from this hypothesis that co-offending moderates the link between peer delinquency and adolescent delinquency. Yet, co-offending is rarely measured directly in the (developmental) psychology literature, and longitudinal studies that directly measure this concept in the criminological literature are also uncommon (cf Goldweber et al.
2011). As a result, longitudinal studies are virtually non-existent on whether co-offending indeed plays a moderating role in the relationship between delinquency on the one hand and delinquency of peers and short-term mindsets on the other hand. To this end, drawing upon the companions in crime hypothesis (Warr
2002), differential association theory (Sutherland
1947) and the psychosocial maturity hypothesis (Steinberg et al.
2009), the current longitudinal study investigates whether the co-development of (non-)best friends’ delinquency with adolescents’ delinquency and short-term mindsets (impulsivity and lack of school future orientation) are stronger when adolescents engage in co-offending with their friends.
Co-development of Friends’ Delinquency and Adolescent Delinquency
The peer context becomes increasingly important for adolescents. In addition to close (smaller in size) friendships with best friends, individuals interactions with peers also occur in cliques, crowds, and larger networks of friends (Brechwald and Prinstein
2011). Research is currently mixed on the influences of best-friends versus such ancillary friends (i.e., non-best friends; Rees and Pogarsky
2011). Adolescents tend to overestimate the similarity between their own behavior and the behavior of their friends’ behavior when perceptual measures of peer delinquency are used (Rees and Greg Pogarsky
2011). Nevertheless, when accounting for measurement issues (e.g., the use of perceptual measures), “peer influence” on delinquent behavior has been shown to be paramount, as it can occur in comparable ways within adolescents’ interactions with delinquent best friends versus delinquent ancillary friends (Rees and Greg Pogarsky
2011).
When (delinquent) peers interact, co-offending can be the result, especially during adolescence (for an overview see: Warr
2002). Research on such delinquent peer influence suggests that it need not exclusively occur within (best) friend relationships, as it is common in broader peer networks as well (Brechwald and Prinstein
2011; Rees and Pogarsky
2011). Moreover, co-offending with delinquent peers has been theorized to link delinquent peer affiliation to adolescent delinquency (Dynes et al.
2015). Scholars have even proposed that “the age distribution of crime, stems from age related changes in peer relations” (p. 99; Warr
2002; but see Stolzenberg and D’Alessio
2008). This assumption has more recently been referred to as the “companions in crime hypothesis” (see Stolzenberg and D’Alessio
2008). This hypothesis suggests that peer influence during co-offending is a mechanism that causes the progression of delinquency during adolescence. For example, adolescents tend to imitate each other’s delinquency (Piquero and Moffitt
2010), and this effect can be even stronger when they are in company of each other while engaging in delinquency.
Although developmental theories and longitudinal research on co-offending is limited (cf. Goldweber et al.
2011), some criminological theories on social learning explicitly hypothesize that co-offending is associated with the onset, persistence and desistance of delinquency (Piquero et al.
2007). For instance, it is presumed that co-offending provides a setting wherein peers can directly influence each other and promote increasing levels of delinquency (Dynes et al.
2015). Individuals who co-offend might also be more susceptible to delinquent peer norms (Dynes et al.
2015). This assertion is in line with one of the propositions of differential association theory, which hypothesizes that that interactions with delinquent peers facilitate the learning of criminal techniques (Dynes et al.
2015; Sutherland
1947). Hence, it is assumed that this transmission of criminal techniques will be stronger if individuals are in the presence of each other during the engagement in delinquency (i.e., co-offending), versus whether individuals merely affiliate with delinquent peers (e.g., Dynes et al.
2015). Such co-offending could also be a result of peer group conformity (Asch
1951), and it may even influence (subsequent) delinquent behavior in the long-term. Taken together, it stands to reason that the influence of peer delinquency could be stronger for adolescents who co-offend compared to solo-offenders who merely have delinquent peers but do not co-offend with them. However, studies that could address this question are presently lacking (cf., Dynes et al.
2015), particularly because studies that assess delinquent peer affiliation do not specifically assess co-offending (cf. Dynes et al.
2015; McGloin and Stickle
2011).
Nevertheless, at least two cross-sectional studies based on court-involved youth have investigated related questions. One of those studies found that the link between friends’ delinquency and adolescent delinquency only existed for adolescents who actually co-offend with their peers (Dynes et al.
2015). Additionally, the second study reported that compared to other offenders, chronic offenders were less likely to mention “peer influence” as a reason for engaging in delinquency (McGloin and Stickle
2011). However, these two offender groups were equally likely to engage in co-offending with peers (McGloin and Stickle
2011). It was thus concluded that although chronic offenders are less likely to engage in delinquency because of their peers, they are still just as likely to engage in co-offending (McGloin and Stickle
2011). These results suggest that delinquency of peers does not necessarily have to imply that adolescents are engaging in delinquency with their peers, and that delinquency with peers can moderate the link between peer delinquency and adolescent delinquency. Building on these two cross-sectional studies on court-involved youth, an important aim of the current study is to use a longitudinal design to establish whether there is co-development among adolescents’ delinquency and best friends’ delinquency in a community sample. Furthermore, the current study uniquely examines whether this hypothesized co-development is stronger when adolescents co-offend with their friends. These hypotheses are in line with the companions in crime hypothesis (Warr
2002).
Co-development of Friends’ Delinquency and Short-term Mindsets
The psychosocial maturity hypothesis suggests that influence of peers on (deviant) behaviors, and indicators of short-term mindsets (e.g., impulsivity, lack future orientation), all show significant and similar non-linear development particularly during adolescence, with these behaviors peaking during mid-adolescence (Steinberg,
2008; Steinberg et al.
2009; but see e.g., Chen
2009; Duell et al.
2016). Extrapolating from this notion, the current study pioneers a test of whether the mere delinquency of best friends is developmentally interrelated with impulsivity and lack of school future orientation during adolescence, and whether this is particularly the case when adolescents engage in delinquency together. This hypothesis also aligns with differential association theory on transmission of attitudes and beliefs (Sutherland
1947; but see Hochstetler et al.
2002). Namely, exposure to values and beliefs about violation of the law via delinquent peers could promote short-term mindsets such as impulsivity and lack of future orientation. This hypothesis is important to investigate, because if peer delinquency and short-term mindsets show such co-development, this could imply that peer delinquency is related to adolescent delinquency
because it contributes to the development of short-term mindsets in youth. As such, short-term mindsets could be the conduit through which the association between peer delinquency and adolescent delinquency develops over time. Essentially, adolescents may adapt their own attitudes from observing the impulsive and risky behavior of their delinquent peers, for example by hearing them talk about taking risks and disregarding the future (Meldrum et al.
2012).
Co-offending (i.e., being present and thus observing others engaging in delinquency) could likely also make the association between delinquency of peers and short-term mindsets stronger. Specifically, co-offending could lead adolescents to conclude that they must be impulsive individuals who do not care about the future—and this could thus encourage the development of short-term mindsets. As such, it is further conceivable that co-offending would exacerbate the link between delinquency of peers and adolescent delinquency, which the current study explicitly investigates. These assertions overlap with the three hypotheses mentioned earlier, namely, companions in crime hypothesis, psychosocial maturity hypothesis, and differential association theory. Then again, of note is that other scholars (Hochstetler et al.
2002) have been more critical of such assertions related to differential association which posits that crime-condoning tendencies are transmitted during co-offending. It has for example been argued that although interactions with delinquent peers during co-offending have shown to exacerbate delinquency, this is not because of increases in crime-condoning attitudes (Hochstetler et al.
2002). Namely, empirical research demonstrated that crime-condoning attitudes do not appear to be the mechanism that links delinquent peer influences into solo-offending or co-offending, which questions whether group influences such as co-offending is the mechanism of differential association (Hochstetler et al.
2002). However, a search of the current literature did not result in any studies that have explicitly investigated this implied interaction between delinquency of peers and co-offending in the prediction of crime-condoning tendencies such as impulsivity and lack of future orientation. Nevertheless, three longitudinal studies were located that investigated whether an indicator of peer delinquency is longitudinally associated with levels of impulsivity, or self-control, more broadly.
The first study (Meldrum et al.
2012) used a social network-design and showed that changes in delinquency of classmates were related to subsequent changes in adolescent self-control (see also: Huijsmans et al.
2019; Mcgloin and Shermer
2009). Hence, this study showed support for differential association theory (Akers
2008; Meldrum et al.
2012; Sutherland
1947). A fixed-effects model in combination with a sequential latent growth model (LGM) was used for the analyses (Meldrum et al.
2012), which could account for unmeasured time-stable correlates and identify a specific temporal ordering, respectively. The results showed that delinquent classmates encourage lower self-control in adolescents (Meldrum et al.
2012). However, since a sequential LGM was used, it is unclear whether
reverse temporal-ordering between constructs might also exist. That is, individuals with lower levels of self-control might also be the ones who gravitate more towards delinquent peers. In order to facilitate the testing of such bi-directional links, a recent study based on the current study sample used cross-lagged panel modeling and showed that a broad measure of self-control (including impulsivity, self-centeredness, risk seeking, short temperedness, and preference for physical activities) was bi-directionally related to friends’ delinquency during different stages of adolescence (Huijsmans et al.
2019). However, none of the two abovementioned studies (i.e., Huijsmans et al.
2019; Meldrum et al.
2012) investigated whether peer delinquency and self-control “travel together” and show co-development over time. To address this possibility, the current study uses a
parallel LGM approach. Additionally, unlike the current study, both of the aforementioned studies (Huijsmans et al.
2019; Meldrum et al.
2012) only assessed delinquency of peers, and not whether adolescents were actually engaging in co-offending. The current study adds to literature by additionally investigating co-offending, and whether it moderates the hypothesized co-development between delinquency of best friends and short-term mindsets—as extrapolated from a combination of the abovementioned three theories that guide this study.
The final relevant study did assess co-offending, although it did not explore it as a moderator (Goldweber et al.
2011; see also Ashton et al.
2020). Via trajectory group modeling among a sample of serious male adolescent offenders, it was demonstrated that the group which increasingly engaged in co-offending showed less psychosocial maturity (i.e.,
more short-term mindsets) versus the occasional/mixed solo-offenders and the exclusively solo offenders from age 14 to 17. Those results show support for psychosocial maturity hypothesis and differential association theory. The current study extends the aforementioned study (Goldweber et al.
2011) in various ways. First, in addition to including a measurement of co-offending, the present study simultaneously investigates whether mere delinquency of best friends also predicts short-term mindsets and whether this is dependent on whether adolescents are actually engaging in co-offending with their friends. That is, whether co-offending (versus non co-offending) serves as a moderator in the co-development between delinquency of best friends and short-term mindsets is investigated. Secondly, instead of using one measurement wave of short-term mindsets and group trajectory modeling, the present study uses three measurement waves of both short-term-mindsets and friends’ delinquency. This is achieved via parallel LGMs in order to investigate whether the development of these constructs show correlated change (i.e., co-development) over three waves. Thirdly, the current research questions are examined using a community sample as opposed to a clinical sample (see Goldweber et al.
2011).
Finally, unlike the similar abovementioned study (Goldweber et al.
2011), the current study takes well-established correlated risk factors of delinquency into account. For example, it has consistently been demonstrated that demographic factors such as gender and ethnicity are strong correlates of crime, namely males and ethnic minorities are over-represented in the juvenile justice system (for a review see: Piquero et al.
2015). Additionally, an extensive review of longitudinal studies found that socioeconomic factors, such as low family income (which is intertwined with parental occupation) as well as youth’s own school achievement are among the strongest predictors of juvenile delinquency (Murray and Farrington
2010). These important correlates will be controlled for in the current study.