Skip to main content
SearchLoginLogin or Signup

Incarceration as a turning point? The impact of custody experiences and identity change on community reentry

Published onApr 29, 2018
Incarceration as a turning point? The impact of custody experiences and identity change on community reentry


Despite the asserted importance of community reentry as part of the pathway to desistance, there is relatively little empirical research examining the role of custody experiences and a young person’s personal transformation while incarcerated. With the increasing emphasis on service delivery within Canadian facilities, it is possible that some aspects of custody help facilitate desistance, but research has yet to clarify which specific aspects of custody promote desistance. Guided by rational choice, life course, and cognitive transformation perspectives on desistance, the current study attempted to address this question by examining the impact of various custody experiences and transformations impact on short and longer-term offending outcomes for a sample of incarcerated male and female youth (n = 217). The reliable change index was used to examine changes in self-identity between admission to custody and the weeks prior to release. Short versus longer-term offending outcomes were influenced by different custody experiences and the importance of identity change varied across these different outcomes. Incarceration as a transformative experience during a key transition period in the life course is discussed.

Keywords: Community reentry, desistance, incarceration, juvenile delinquency, longitudinal study, reliable change index


Despite the importance of community reentry for rehabilitation and desistance research (Petersilia, 2003), little is known about whether an individual’s experiences and transformations while incarcerated impact the likelihood of future offending. This neglect is perhaps because the incarceration experience and the process of institutionalization, often referred to as prisonization (e.g., Thomas & Foster, 1972), are assumed to adversely impact all individuals (e.g., Sykes, 1985) by immersing them in a prison culture that acts as a barrier to effective socialization (e.g., Zingraff, 1975) and prosocial thinking styles (e.g., Walters, 2003). From a social learning theory perspective, prisons provide individuals with opportunities to increase their network of criminal accomplices (Shover, 1996). Labeling theorists argued that formal mechanisms of social control, such as incarceration, have negative consequences on an individual’s self-image that favor the persistence of crime (e.g., Becker, 1963; Lemert, 1967). Similarly, developmental and life course researchers suggested that formal interventions weaken ties to family and other sources of informal social control that are important for desistance (Sampson & Laub, 1997). Formal interventions can impede desistance by delaying a person’s timely entry into important social roles (e.g., employment) that can act as turning points in the life course (Loeber & Le Blanc, 1990). Part of this delay may be because the stigmatizing effect of incarceration includes decreasing the likelihood of entering the workforce, working specific jobs, and accessing housing (e.g., Lopes et al., 2012). In support of these assertions, Gilman, Hill, and Hawkins (2015) illustrated that incarceration decreased the likelihood of desistance for youth (also see Bernburg & Krohn, 2002).

Although the above perspectives may characterize the pains of imprisonment experiences of American youth, such experiences may not be entirely generalizable elsewhere (e.g., Bonta & Gendreau, 1990; Killias, Gilliéron, Villard, Poglia, 2010). For example, the Canadian model of youth justice emphasizes service delivery to promote the rehabilitation of incarcerated youth (Youth Criminal Justice Act, 2002). Although specific services are not investigated in the current study, this more rehabilitation-focused custody context supports the premise that serious and violent young offenders in Canada may experience positive events and transformations while incarcerated that influence their pathway to desistance. This rehabilitation context is shaped by specific legislation (Youth Criminal Justice Act, 2002) that is consistent with principles of the risk-need-responsivity (RNR) model (Andrews, Bonta, & Hoge, 1990). The RNR model involves offering a level of service that meets a person’s needs and minimizes the risk of future offending (e.g., Andrews, Bonta, & Wormith, 2006). In accordance with RNR principles, the Canadian model aims to facilitate successful community reentry (i.e., prevent reoffending) by offering transitional services as part of the youth’s custody sentence (Youth Criminal Justice Act, 2002). Evaluation studies have shown that correctional programs implemented and delivered in accordance with RNR principles are associated with a reduction in recidivism (e.g., Koehler, Lösel, Akoensi & Humpreys, 2013). Importantly, needs may vary depending on the youth’s prior criminal history. Thus, assessment tools were developed to understand the specific needs of, for example, violent youth (e.g., Borum, Bartel, and Forth, 2002).

Especially considering the extensive risk factor profile of incarcerated youth (e.g., Baglivio & Epps, 2016), the structured nature of custody may provide a suitable environment for the implementation of intensive intervention and treatment strategies that promote a process of slowing down in the level of offending (i.e., desistance; see Bushway, Piquero, Broidy, Cauffman, & Mazerolle, 2001; Lussier, McCuish, & Corrado, 2015). Custody may help individuals reconnect with family, sever ties with criminal associates in the community, and make plans to return to school or acquire a job upon community reentry. Dmitrieva, Monahan, Steinberg, and Cauffman (2012) indicated that treatment and rehabilitation-focused custody facilities were conducive to improving maturation, which is associated with desistance (Rocque, 2017). As well, Rocque, Bierie, and MacKenzie (2011) found that some individuals’ social bonding mechanisms improved while incarcerated and Paternoster and Bushway (2009) noted that restriction on freedoms and dissatisfaction with current life circumstances, both of which are likely in custody, might promote positive behavioral change.

Custody should not be used as a means of promoting desistance when less punitive options provide a suitable response (Gilman et al., 2015). However, in instances where incarceration is deemed a necessary response to serious and violent offending, what becomes a critical question is whether there are specific custody experiences or personal transformations that occur while incarcerated that impact the desistance process. The purpose of the current study was to address this question using data from the Incarcerated Serious and Violent Young Offender Study (ISVYOS). Just prior to their release, adjudicated male and female youth (n = 217) were interviewed regarding their experiences in custody. Following community reentry, participants were followed for an average of 72.26 months (SD = 21.11) to examine the impact of custody experiences on short-term and longer-term reoffending. Although there is a dearth of research addressing the impact of custody on successful community reentry, there are several major perspectives on desistance that can be used to guide the measurement of custody experiences and transformations.

Models of Desistance from Crime

A person’s conscious decision to cease their involvement in crime is a key component of desistance. Per the rational choice perspective, formal and informal consequences of criminal activity, when weighed against benefits of crime involvement, initiate the decision to desist (e.g., Clarke & Cornish, 1985; Cusson & Pinsonneault, 1986). A reassessment of the costs versus benefits of crime can be initiated via formal criminal justice sanctions such as being arrested, interrogated by the police, the courtroom experience, and the experience of incarceration (Cusson & Pinsonneault, 1986). Collateral consequences of formal sanctions on life domains such as employment and marriage (e.g., Lopes et al., 2012) are also part of this reassessment. Formal sanctions may also increase an individual’s perception of the likelihood that future crimes will be detected by police (e.g., Horney & Marshall, 1992). Moreover, the consequences of detection increase with a more extensive prior criminal record (i.e., an aggravating factor at sentencing). An offender’s awareness of this may contribute to the perception that the costs of offending outweigh the benefits. On the other hand, informal consequences refer to ‘shocks’ that are initiated outside the realm of the formal justice system (Cusson & Pinsonneault, 1986). These shocks include actual and perceived likelihood of injury resulting from offending (e.g., victim resistance, violence from co-offenders), the death of co-offenders during the commission of an offense, and confrontations with police (Cusson & Pinsonneault, 1986). Although incarceration is typically framed in terms of its influence on offending persistence (e.g., Thomas & Foster, 1972), from a rational choice perspective, the restriction on freedom and other negative custody experiences would factor into an individual’s decision to stop offending.

From a life course perspective, the decision to stop offending is bound to the nature of an individual’s social environment and life circumstances (Sampson & Laub, 1997). Individuals may view crime involvement as having serious consequences on their existing social relations and life circumstances and thus become unwilling to engage in crime for fear of jeopardizing their current position (Laub & Sampson, 2001). The importance that an individual ascribes to sources of informal social control is critical for desistance. For example, if employment, marriage, or parenthood occurred prior to a certain level of maturity, or was not valued by the individual, then desistance is less likely (e.g., Kazemian & Maruna, 2009; Uggen, 2000). Thus, when examining the initiation of desistance earlier in the life course (e.g., adolescence) it may be necessary to focus on social roles (e.g., student) and relations (e.g., ties to family, friends) that are more in line with the individual’s developmental stage, interests, and level of maturity. On the other hand, some suggest that social roles at any stage will not influence desistance if the individual does not first undergo a positive identity change (e.g., Paternoster & Bushway, 2009)1.

According to cognitive transformation perspectives, within-individual identity change is critical to the desistance process (e.g., Giordano, Cernkovich, & Rudolph, 2002; Paternoster & Bushway, 2009). In contrast to rational choice perspectives, it is not a reassessment of the costs and benefits of crime that influences desistance but rather a reassessment of oneself (e.g., Giordano et al., 2002; Paternoster & Bushway, 2009). Giordano et al. (2002) described this reassessment as swapping an antisocial identity for a prosocial one (i.e., ‘the replacement self’). An individual’s emotional health has implications for whether a positive sense of self initiates this motivation to avoid criminal behavior (e.g., Giordano, Schroeder, & Cernkovich, 2007). In contrast, Paternoster and Bushway (2009) described this reassessment as an individual’s initial concern with who they are and their current life path (i.e., ‘the feared self’). The order in which this reassessment occurs also varies across these two perspectives. Giordano et al. (2002) suggested that hooks for change and social environment were necessary precursors to identity change whereas Paternoster and Bushway (2009) suggested that identity change was a necessary precursor to the acquisition of informal social controls.

Prospective research on identity change and desistance is in its infancy. Rocque, Posick, and Paternoster (2016) used growth curve modeling to examine the impact of prosocial identity change on desistance among male and female participants (n = 1,380) from the Rutgers Health and Human Development Project (HHDP). The HHDP is a community-based study of participants recruited at age 12, 15, or 18 and followed until at least their late 20s. Controlling for relevant covariates such as social bonds and connections to delinquent peers, both between-individual differences and within-individual change in identity were informative of desistance. In another study using prospective data on adult serious drug users, Na, Paternoster, and Bachman (2015) found that self-identity did not directly influence reductions in re-arrest after controlling for relevant covariates (e.g., substance use, relationship quality). Na et al. (2015) suggested that having a more positive self-identity may have initiated a more prosocial and healthy lifestyle, including motivation to seek treatment for substance use, and this change in drug use habits directly influenced desistance. This finding is in line with Paternoster and Bushway’s (2009) suggestion that identity change precedes the acquisition of informal social controls. Whereas the samples in Rocque et al.’s (2016) study and Na et al.’s (2015) were more representative of community members and serious drug users, respectively, less has been learned about identity change among serious young offenders, especially identity change that occurs while incarcerated. Labeling perspectives would suggest that incarceration negatively impacts an individual’s self-identity (e.g., Becker, 1963). However, given the recent movement towards emphasizing the strengths of young offenders (e.g., Tweed et al., 2011; Viljoen et al., 2012), an individual’s time in custody may also be a period in which their positive characteristics are reinforced by practitioners, resulting in more positive self-perceptions.

Study Aims

In the Canadian context, where rehabilitation and service delivery are legislated aspects of the youth justice system (Youth Criminal Justice Act, 2002), the custody experience is potentially a turning point for serious and violent offenders. The current study measured an individual’s experiences and changes that took place while incarcerated with the aim of identifying whether such factors influenced recidivism in the short-term (i.e., the first year of release) and repeated offending over the longer-term (i.e., several years after community reentry). To address calls for specific guidelines for responding to violent offenders (e.g., Polaschek, 2012), and in line with the need principle of the RNR model, both non-administrative convictions (i.e., all offenses other than violations of court orders) and violent convictions (i.e., hands-on offenses) were examined. Investigating the experiences and transformations of adolescent offenders helped address questions about factors that might help initiate the desistance process early in the life course. Indeed, existing research on turning points primarily focused on factors such as marriage and stable employment (Nguyen & Loughran, 2018), which may not be applicable or helpful to the desistance process of young offenders. Focusing on experiences and transformations while incarcerated ensured that the factors linked to desistance did not emerge well-after the desistance process was initiated. The measurement of custody experiences and transformations was guided by rational choice, life course, and cognitive transformation perspectives on desistance.

Rational choice perspectives on desistance were evaluated by assessing whether formal consequences (e.g., length of time incarcerated) and informal consequences (e.g., victimization while incarcerated) influenced lower levels of offending. Guided by the life course perspective, we examined whether support received from custody staff and family and expectations about experiencing different sources of informal social control influenced offending outcomes upon community reentry. Cognitive transformation perspectives on desistance were examined by assessing whether an individual-level change in self-identity during incarceration was associated with lower levels of offending. This line of analysis examined whether identities can change for the better against the backdrop of a prison environment and whether such change was associated with lower levels of offending. Participants experiencing a positive shift in identity were expected to forgo crime opportunities upon community reentry, either because of a desire to make behavior congruent with their new sense of self (e.g., Giordano et al., 2002) or because opportunities for crime were no longer regarded as such because of changes in moral attitudes (Schneider & Ervin, 1990). Guided by these different perspectives, the current study sought to identify specific aspects of incarceration that promote or hinder successful community reentry.


Sample and Procedures

The current study used data from Cohort Two of the Incarcerated Serious and Violent Young Offender Study (ISVYOS), conducted in British Columbia (BC), Canada. As part of this study, adjudicated youth between ages 12 and 19 were interviewed in open and secure custody facilities in Burnaby and Victoria between 2005 and 2011. The BC Ministry of Child and Family Development is the legal guardian to all incarcerated youth and consented to their recruitment. Individuals were recruited if they were English-speaking, capable of understanding interview questions (e.g., lower functioning youth [per official file information] were excluded), and were willing to provide accurate information. Regarding the latter criterion, two youth were permanently removed from the interview schedule because they continued to purposefully provide research assistants (RAs) with false information about themselves. Before beginning the interview, individuals approached for recruitment were informed that their involvement in the study would not affect their stay at the custody centre. Participants were informed that information would be kept confidential, except in circumstances where they made a direct threat to harm themselves or others. Approximately five percent of youth declined to participate in the study when approached by RAs. Youth assented to participate in the study after a RA read aloud an information sheet that described details of the study. To help ensure confidentiality, participants were interviewed in a private room away from other residents and staff.

RAs conducted an intake interview, typically within the first week of a participant’s arrival to custody, that included information on demographic characteristics, school behavior, family history, substance use, aggressive behaviors, and perceptions of self-identity. After completing this interview, if participants were incarcerated for at least 30 days, RAs conducted a second interview that was initiated just prior to community reentry. Referred to as the ‘Exit’ interview, questions were asked about participants’ experiences while incarcerated, such as missing their life in the community and remaining attached to their parents. A second assessment of self-identity was also performed and was used to evaluate potential transformations during their time incarcerated. The Exit interview was not conducted with participants that spent few days in custody because interview questions would not apply to them (e.g., “in the last few weeks, how often have you felt…”). A total of 231 participants completed both interviews; however, criminal histories were unavailable for ten participants and four participants were not released during the study’s most recent wave of data collection (January 2017). Thus, the final sample consists of 217 participants, the majority of whom were male (n = 188; 86.8%) and White (n = 112; 51.6%). As well, although approximately six percent of the population of British Columbia self-identifies as Indigenous (Statistics Canada, 2013), 31.3% of offenders in the full sample self-identified as such.


Risk and protective factors and demographic characteristics (for comparisons across gender, see Table 1) were measured using a series of structured interviews. Gender, ethnicity (White, Indigenous, and non-Indigenous minority groups), and age at community reentry were controlled for in all negative binomial analyses. On average, participants returned to the community when they were 16.79 years of age (SD = 1.48; age range of 13-24). Risk and protective factors were measured in accordance with key constructs specified by different theories on desistance (see Appendix A for more detailed information concerning the instruments used). All desistance mechanisms were measured prior to the offending outcomes under examination, which helped avoid concerns that, for example, an identity change was a consequence of, rather than a precursor to, lower levels of offending.

--Insert Table 1 about Here--

Informal Consequences of Crime Involvement. To capture informal consequences of involvement in crime as described by rational choice perspectives on desistance, the number of times that participants reported being physically attacked while incarcerated was used to capture the types of ‘shocks’ that lead to desistance (Cusson & Pinsonneault, 1986). Such victimization experiences were considered informal consequences of crime involvement because, although they occurred within the context of a formal justice system sanction (i.e., custody), they were not part of the formal intent of the justice system. To capture the qualitative aspect of formal sanctions, challenges with the custody experience were coded using a four-point Likert scale, with response options ranging from not at all difficult to extremely difficult. Example items included whether the participant had difficulty with: getting along with staff, getting along with other residents, missing family in the community, and missing the freedom that they had in the community. The amount of time an individual spent incarcerated prior to community reentry was used to capture the more quantitative aspect of formal consequences of criminal justice system involvement. There was no correlation between time incarcerated and number of physical victimizations in custody (r = -.01; p = .877), suggesting that victimization experiences were more than just a matter of exposure to the custody environment.

Informal Social Controls and Support. To capture the types of informal social control described by life course theories of desistance, Turner and Marino’s (1994) Measure of Perceived Social Support was adapted for youth to examine support that participants received from family and custody staff while incarcerated. Participants’ intentions to return to school and find a job were also measured as anticipated turning points. It was not possible to measure whether these anticipated events occurred because interviews in the community were not performed. That said, measuring the impact of intentions while incarcerated on offending outcomes upon release helped establish temporal order between predictor and outcome.

Identity Change. Prior research often captured self-identity using measures of self-esteem (e.g., Rocque et al., 2016), which might reflect the emotional health aspect that helps influence desistance for those that experience an identity change (Giordano et al., 2007). It may be helpful to expand on this research by looking at a broader range of identity traits (Crank, 2016; Maruna, Immarigeon, & LeBel, 2004; Rocque et al., 2016). In line with symbolic interactionist perspectives (e.g., Matsueda, 1992), Schneider’s (1990) 15-item Good Citizen Scale captures a wider range of self-reported identity traits that are aligned with desistance scripts emerging from the narratives of ex-convicts (e.g., Maruna, 2001). The scale was designed to differentiate “good citizens” from “lawbreakers”, was expected to be correlated with delinquent behavior, and was developed using a sample of adjudicated youth (see Schneider & Ervin, 1990). The Good Citizen Scale was administered during the intake interview (Time 1) and again during the Exit interview (Time 2). This allowed for an assessment of whether participants’ self-identity changed while incarcerated. To evaluate different aspects of identity, a principal components analysis was performed. In line with prior research using data from the ISVYOS (McCuish, Lussier, & Corrado, 2016), three aspects of identity emerged: obedience (e.g., cooperative, polite, obeys rules), prosociality (e.g., good, helpful, honest), and hypermasculinity (e.g., tough, attractive, rich). As the latter aspect of identity was not characterized by personality traits and role-based perspectives of self-identity that would be helpful for desistance (Rise, Sheeran, & Hukkelberg, 2010; Rocque et al., 2016), only obedience and prosociality were examined in the current study. To control for emotional experiences that may impact the likelihood of change (Giordano et al., 2007) and involvement in offending, we used a measure of whether participants viewed their life as a success or failure during the last couple of weeks prior to their Exit interview.

Offending Measures. The British Columbia Corrections Network (CORNET) is a secure software program that includes information regarding the official criminal history of all individuals in contact with the provincial court system. This criminal history includes the specific criminal code violation for which an individual received a conviction, as well as the date of every entry and release from custody. For the current study, CORNET was used to measure non-administrative and violent convictions incurred both before and after community reentry (see Table 1). Non-administrative offenses included all convictions that were not related to administrative offenses (e.g., failure to comply with a probation order). Administrative offenses were excluded to focus on criminogenic behavior as opposed to rule violations. Violent convictions were restricted to sexual-based offenses, different forms of assault, and homicide. Uttering threats and weapons possession were excluded from the definition of violence to focus solely on behaviors that resulted in physical harm (i.e., to avoid saturating the measure of violence with less serious behaviors likely to occur more frequently). These two measures were used as the outcomes of interest in analyses of the short-term effects of custody on recidivism (i.e., conviction after one year of release) and of the longer-term effects of custody on reoffending (i.e., number of total convictions during the follow-up period). The former was examined via logistic regression analyses and the latter was examined via negative binomial analyses. The combination of short-term and longer-term measure of offending minimizes the risk of false negatives in the examination of desistance (Lussier et al., 2015).

Data Analysis

The first step in the data analysis involved using the reliable change index (RCI) to evaluate participants’ change or stability in levels of obedience and prosociality between the start of their custody term and just prior to community reentry. Christensen and Mendoza’s (1986) formula for calculating reliable change was used:

  1. RCI=yijyij1SEMj2+SEMj12RCI = \frac{y_{\text{ij}} - y_{ij - 1}}{\sqrt{\text{SE}M_{j}^{2} + S}EM_{j - 1}^{2}}

In this formula, yijy_{\text{ij}} is the person’s score for the Exit interview measurement period (i.e., Time 2) and yij1y_{ij - 1} is the score for the intake interview measurement period (i.e., Time 1). SEM \text{\ SEM\ }is the standard error of measurement2. This formula produces a range of scores and accounts for measurement error by determining whether an individual’s level of change was more than expected based on chance alone (e.g., measurement error, regression to the mean). Whereas correlations, t-tests, and rank-order comparisons evaluate stability at the group-level, the RCI is a person-oriented measure and useful for evaluating within-individual change when just two measurement periods are available RCI values of -1.96 and below indicates a reliably lower score at T2 compared to T1 (e.g., a more negative self-identity in the context of the current study). RCI values of 1.96 and above indicates a reliably higher score at T2 compared to T1 (e.g., a more positive self-identity).

Demographic characteristics and the different theory-informed constructs believed to promote desistance were then included in logistic and negative binomial analyses to examine the short-term and longer-term effects of these measures on offending. All analyses accounted for exposure time. The logistic regression analysis accounted for proportion of the one-year follow-up that individuals were not incarcerated. For the negative binomial analyses, the length of the follow-up period varied across participants (M = 72.26 months; range of 17.07 months to 126.5 months), as did the amount of time spent incarcerated (M = 15.39 months; range of zero months to 89 months. The varying range in follow-up was because some participants were interviewed as early as 2005, others as late as 2011. To account for this variation in both the short-term and longer-term analyses, CORNET data were used to create a measure of exposure time, which was defined as the number of days that an individual was free in the community during their individual follow-up period, divided by 126.5 months (the longest individual follow-up period in the study). There were no issues with multicollinearity (see Appendix A) among all variables included in the analyses (r < .600; variance inflation factor < 2); the strongest correlation was an expected association between age at release and length of incarceration (r = .596).


Reliability of Change in Self-Identity

For individual items comprising the obedience and prosociality subscales, a neutral score would be indicated by an of ‘4’. At Time 1 (T1), only one identity trait (Breaks Rules/Obeys Rules) had an average below four. The same observation was made at Time 2 (T2). The average score on the obedience subscale was 17.95 (SD = 3.85) at T1 and 19.16 (SD = 4.08) at T2. The average score on the prosociality subscale was 23.80 (SD = 4.56) at T1 and 24.66 (SD = 4.81) at T2. Paired samples t-tests showed that average scores on both subscales significantly improved between T1 and T2 (p < .05). However, such analyses mask any potential patterns in within-individual change and instead simply reflect an ‘average’ change.

Christensen and Mendoza’s (1986) formula for the RCI was used to examine whether within-individual change on the obedience and prosociality subscales was reliable. Beginning with the obedience subscale, 16 participants evidenced a reliable increase in test score (7.4% of the sample) and 11 participants experienced a reliable decrease in test score (5.1% of the sample). For the prosociality subscale, 25 participants experienced a reliable increase in test score (11.5% of the sample) and 13 participants (6.0% of the sample) experienced a reliable decrease in test score. Due to concerns about low cell counts, those experiencing a negative change and those that did not reliably change were collapsed into the same group and compared against individuals that experienced positive change in obedience and prosociality.

A chi-square analysis showed that there was significant overlap between the group experiencing a positive change in obedience and the group experiencing a positive change in prosociality (χ 2 [1] = 6.60, p < .05, ϕ = .17). Specifically, five of the 16 participants that experienced a positive change in obedience also experienced a positive change in prosociality. It was not simply the case that individuals with positive identities at T1 became more positive in their perception of self and individuals with negative identities at T1 became more negative. Indeed, independent samples t-tests showed that participants experiencing a reliable increase in obedience had significantly (p < .05) lower scores at T1 compared to the reference group. The same finding was observed for those that experienced a positive change in prosociality.

Short-Term Influence of Custody on Offending

The analyses in Table 2 show the relationship between demographic characteristics, self-identity changes, and experiences while incarcerated on short-term recidivism (i.e., the first year of release from custody). Separate analyses were conducted for non-administrative recidivism and violent recidivism. Regarding the covariates of non-administrative convictions (Model 1), empirical support was observed for each of the three desistance perspectives examined. In line with cognitive transformation theories of desistance, individuals that viewed themselves as more obedient were less likely to reoffend during the one-year follow-up. When baseline levels of obedience were included in the model (not shown), the effect of positive change in obedience remained. In other words, the lower likelihood of recidivism for the group that experienced a positive identity could not be explained by differences in baseline scores between this group and the reference category. In line with life course perspectives, expectations about retuning to school decreased the odds of reconviction. Finally, formal criminal justice sanctions in the form of number of months incarcerated significantly reduced the odds of reconviction during the first year of release, which was in line with rational choice perspectives. Experiences of victimization in custody, which may be interpreted as informal consequences of justice system involvement, significantly increased the odds of conviction for a non-administrative offense. Number of victimizations in custody was unrelated to factors such as age at release or number of months spent incarcerated (see Appendix A), suggesting that the number of times a participant was victimized could not be explained simply by differences in the amount of time exposed to the custody environment. A greater number of convictions prior to community reentry significantly increased the odds of reconviction whereas an older age at community reentry decreased the odds of reconviction. Gender and ethnicity had no bearing on the likelihood of reconviction.

For the model with short-term violent reconviction as the outcome of interest (Model 2), only covariates reflecting the life course perspective on desistance had their expected effect. Specifically, increases in the level of support participants received from family while incarcerated significantly decreased the likelihood of incurring a violent conviction following community reentry. Similarly, increases in the level of support participants received from custody staff while incarcerated significantly decreased the likelihood of incurring a violent conviction in the first year of release. Contrary to expectation, those that reported they were going to find a job upon release were significantly more likely to incur a violent conviction. Like Model 1, physical victimization in custody increased the likelihood of incurring a violent conviction. Also like Model 1, neither gender nor ethnicity were related to the outcome of interest, but increases in the number of convictions prior to community reentry significantly resulted in a significant increase in the odds of incurring a violent conviction.

--Insert Table 2 about Here--

Longer-Term Influence of Custody on Offending

Analyses of the longer-term influences of different custody experiences and transformations on non-administrative convictions (Model 3) and violent convictions (Model 4) are shown in Table 3. As shown in Model 3, experiences and transformations important for offending in the short-term did not always have prolonged effects on reoffending. Reliable increases in obedience while incarcerated, number of months incarcerated prior to release, and the intention to return to school, although important for short-term recidivism, were not significantly related to the incident rate of non-administrative convictions over the longer-term. Consistent with Model 1, Model 3 indicates that victimization in custody may act as a barrier to desistance over the longer-term, as individuals that reported a greater number of physical assaults while incarcerated incurred a significantly greater incidence of non-administrative convictions. A key protective factor against the incidence of non-administrative reconvictions was participants’ belief that their life was not a failure. Somewhat unexpectedly, participants’ reliable increase in levels of prosociality that occurred while they were incarcerated resulted in an increased likelihood of incurring additional convictions during the follow-up period. When baseline levels of prosociality were included in the model (not shown), the effect of change in prosociality remained. Consistent with previous models, a greater number of convictions prior to community reentry was associated with a significant increase in the incidence rate of non-administrative convictions. Whereas being male had no impact on short-term offending, over the longer-term, being male significantly increased the incident rate of non-administrative convictions. As well, an earlier age at community reentry was associated with a significant decrease in the incidence rate of non-administrative convictions.

--Insert Table 3 about Here--

Turning to the examination of violent reconvictions over the longer-term follow-up (Model 4), the short-term effect of support from custody staff remained. Specifically, individuals that reported a more positive relationship with staff were more likely to have a lower incidence rate of violent convictions following release from custody. In support of rational choice theory descriptions of the role of informal criminal justice system consequences on desistance, participants that reported a greater number of difficulties with their incarceration experience incurred a significantly lower incidence rate of violent convictions. These difficulties typically referred to challenges associated with missing friends and family in the community. Like Model 3, being male, returning to the community at an earlier age, and having a greater number of prior non-administrative convictions were all significantly associated with an increased incidence rate of violent convictions.


Using data on male and female offenders from the ISVYOS (n = 217), the current study examined custody experiences and identity change while incarcerated as factors that promoted or hindered desistance following community reentry. Whereas key turning points like marriage and employment are typically the focus of desistance research (Nguyen & Loughran, 2018), the current study examined factors that were more relevant to the adolescent and emerging adulthood developmental stages. Measuring experiences and transformations that emerged during a time at which youth were in the ‘deep-end’ of the justice system (Mulvey et al., 2004) implied that such factors preceded the desistance process and were not consequences of desistance3. Findings from the current study challenge the idea of systematic pains of imprisonment (e.g., Thomas & Foster, 1972; Sykes, 1985). Participants went through positive experiences that reduced the likelihood of offending in both the short-term and longer-term. That said, some custody experiences also acted as barriers to successful community reentry. As well, some experiences were important for the reduced likelihood of non-administrative offending (i.e., all crimes other than violations of court orders) whereas other experiences were specifically important for the reduced likelihood of violent offending. Such observations reiterate the need for guidance regarding the specific needs factors of violent offenders (Polaschek, 2012). Recommendations for improving the custody experiences of adolescent offenders to help facilitate desistance are outlined below. These experiences are related back to the theoretical perspectives on desistance that guided their measurement.

The measures used in the current study did not allow for a complete test of the rational choice perspective on desistance because participants were not asked to weigh the benefits of offending versus the consequences of getting caught. Nevertheless, such measures captured important information about informal and formal consequences of crime involvement that may ‘shock’ (Cusson & Pinsonneault, 1986) an offender into reassessing the costs of crime. Regarding informal consequences, increases in the number of instances of physical victimization significantly increased the likelihood of (a) non-administrative offending over both short-term and longer-term follow-up periods and (b) the likelihood of violent offending in the short-term. A Canadian study found that at least some young offenders have difficulties adjusting to life in custody (Cesaroni & Peterson-Badali, 2005). These difficulties are particularly arduous for those who have experienced many life adversities, who have a limited number of friends in custody, and who tend to internalize problems and issues (e.g., socially withdrawn, being anxious and depressed; Cesaroni & Peterson-Badali, 2005). Victimization experiences may exacerbate internalizing tendencies, especially for those who were not socially integrated into the custody environment in the first place. This in turn may complicate community reentry and reintegration possibilities such as returning to school. In other words, these young persons have trouble fitting in, even in the custody subculture4. These findings are contrary to hypotheses derived from the narratives of ex-convicts (Cusson & Pinsonneault, 1986). Instead of acting as a shock that added to the perceived costs of offending, physical victimization while in custody contributed to further marginalization. The relationship between victimization and continued offending likely better reflects the general strain theory perspective that abuse experiences are a risk factor for continued offending (e.g., Agnew, 1992) rather than a shock that might promote desistance. Unfortunately, the current study lacked data on participants’ ability to cope with strains that would allow for a fuller test of this perspective.

The quantitative aspect of formal consequences of justice system involvement, measured via time incarcerated prior to community reentry, had a deterrent effect on the likelihood of non-administrative reconviction in the short-term. This effect dissipated over the longer-term follow-up period. It is noteworthy that an older age at release was associated with (a) a lower incidence rate of non-administrative and violent convictions over the full study period and (b) months spent incarcerated (r = .596; p < .001). Thus, those spending a greater amount of time in custody were released at an older age, implying that they were closer to aging out of crime (Hirschi & Gottfredson, 1983) and may have had higher levels of maturity that mitigated the longer-term influence of incarceration length on desistance. For the qualitative aspect of formal consequences of justice system involvement, participants that reported difficulties dealing with the nature of incarceration had a significantly lower incidence rate of violent convictions over the longer-term. These negative custody experiences, such as a difficulty getting along with staff and other residents, may have deterred participants from committing the types of violent crimes that are likely to be responded to with a custody sentence. On the other hand, the measure of negative custody experiences may also be supportive of life course perspectives on desistance, as this measure also captured whether participants missed family and friends. This experience may have lead to the decision to stop offending violently to avoid further jeopardizing these relationships (e.g., Laub & Sampson, 2003). Future research is needed to unpack the types of negative custody experiences most strongly related to desistance.

Rocque et al. (2011) noted that social bonds can improve during periods of incarceration, and we observed that strong bonds while incarcerated helped reduce violent offending. In the one-year period following community reentry, participants that reported receiving more support from custody staff and from family were significantly less likely to violently reoffend. Support from custody staff was also associated with a lower incidence rate of violence over the longer-term. Having positive relationships with custody staff echo the value of a rehabilitation-focused custody environment (Levenson, Willis, & Prescott, 2016). With respect to anticipated turning points, the intention to return to school was helpful in reducing the likelihood of incurring a non-administrative conviction in the first year of community reentry. However, this impact disappeared with a longer follow-up period, which may be related to the tendency for youth to leave school (e.g., graduation, dropping out, expulsion) over the longer-term. Intentions to enter the work force was unrelated to reoffending over the longer-term, which was consistent with the perspective that entering the workforce at too young an age will not improve the likelihood of desistance (e.g., Uggen, 2000).

Drawing from cognitive/identity transformation theories of desistance (e.g., Giordano et al., 2002; Paternoster & Bushway, 2009), the RCI was used to evaluate change in participants’ levels of obedience and prosociality during their time in custody. Most participants entered custody with a positive self-perception, which was in line with assertions from Andrews and Bonta (2010) and Le Blanc (1997) that youthful offenders are often egotistical, self-centered, narcissistic, arrogant, and not afflicted with self-esteem issues. Negative changes in obedience and prosociality were especially unlikely, which was counterintuitive to claims of a prisonization effect (e.g., Thomas & Foster, 1972). Although more common than negative change, only a small proportion of the sample exhibited a positive identity change. This was in line with Topalli’s (2005) observation that persistent offenders made efforts to maintain their criminogenic identity. Another plausible explanation is that time incarcerated, generally 6-month periods, might not have been long enough for these effects to occur. Another possible explanation is that participants lacked the types of experiences and levels of maturity that can help drive identity change (Giordano et al., 2002). The type of identity change studied here was more consistent with Giordano et al.’s (2002) notion of the ‘replacement self’ than Paternoster and Bushway’s (2009) ‘feared self’ given that participants’ attitude toward their identity was not measured.

Findings provided mixed support for assertions about the role of identity change on desistance. In the short-term, positive changes in obedience significantly reduced the likelihood of non-administrative reconviction. In the longer-term, a reliable and positive increase in prosociality while incarcerated was associated with an increased incidence rate of non-administrative convictions. This contrasted with Maruna’s (2001) emphasis on “making good” that emerged from the desistance script narratives of adult ex-convicts. It is possible that this longer follow-up period increased participants’ exposure to negative experiences in the community. Such negative experiences may have undone positive change that occurred while incarcerated5. Another possibility is that young persons who show a reliable change in identity while incarcerated might have difficulties with consistently putting this new identity into practice within various life settings (e.g., school, workplace, friendship and routine activities). Maintaining these changes in the community may also be difficult if the young person is immersed within a criminogenic environment. Unfortunately, the current study lacked measures of negative experiences and sense of identity following community reentry and thus these speculations must be addressed in future research. Finally, an individual’s positive attitude about their current life circumstances significantly predicted a lower incidence rate of non-administrative convictions over the longer-term. Custody staff and practitioners should be attuned to signs of despair and dissatisfaction with life circumstances, as a youth’s emotional state in custody may act as a barrier to positive identity change (Giordano et al., 2007).

Limitations and Future Research

The current study examined identity at just two periods in time; all other factors related to desistance were measured just once. The ideal research design would have repeated measures of both identity and informal social controls and would establish temporal order between the two. That is, does an openness to change that is created by a more prosocial self-identity precede the acquisition of informal social controls (e.g., Bushway & Paternoster, 2009), or does the structurally-induced nature of informal social controls (e.g., Laub & Sampson, 2003) provide individuals with a social context that facilitates identity change? Future research should also examine factors that might inhibit identity change, such as immaturity or depression, and whether certain intervention strategies, such as trauma-informed practices (Levenson et al., 2016), facilitate positive change. Trauma-informed practices may be especially important for facilitating positive identity change given that depression is both a consequence of trauma (Collin-Vezina, Coleman, Milne, Sell, & Daigneault, 2011) and a barrier to identity change (Giordano et al., 2007).

The current study used official data, which are often associated with the underestimation of offending. However, incarcerated adolescent offenders are disproportionately responsible for more serious offenses, which tend to be underreported in self-report studies (e.g., Stouthamer-Loeber, Loeber, Stallings, & Lacourse, 2008). An ideal research design would include both self-report and official measures of offending to examine whether, for example, a change in identity results in a greater effort to avoid detection (i.e., larger discrepancies between self-report and official data) rather than a greater effort to stop offending. Finally, although the longer-term analyses allowed for the examination of offending outcomes in adulthood, the follow-up period was relatively short by the standard of desistance research (Kazemian, 2007). Nevertheless, prior research using data from Cohort I of the ISVYOS, which relied on a longer follow-up period, found that patterns of desistance, if they did emerge, did so by the end of adolescence (Corrado, McCuish, Hart, & DeLisi, 2015). Thus, the start of the desistance process was likely captured here.


Guided by rational choice, life course, and cognitive transformation perspectives on desistance, the current study found that the custody experiences of youth and transformations that occur while in custody have implications for offending outcomes upon community reentry. Contrary to expectations from rational choice perspectives, increases in physical victimization while incarcerated did not ‘shock’ an offender into avoiding criminal behavior upon community reentry. Instead, physical victimization significantly increased the likelihood of offending over the short-term and longer-term. This finding reiterates calls for trauma-informed practice (Branson, Baetz, Horwitz, & Hoagwood, 2017) and a rehabilitative rather than punitive custody environment (e.g., Dmitrieva et al., 2012). Creating a safe environment for adolescents may help improve treatment outcomes and reduce in-custody victimization that, at least in the current study, was associated with continued offending. This rehabilitation-centered environment may also reinforce the types of positive relations with custody staff that influenced lower levels of violent offending in the current study. Moving forward, as noted by Crewe (2011), it is important that staff-resident relationships are not built through superficial kindness. Crewe (2011) illustrated that prisoners appreciate relationships that are neither disingenuous nor distant. That is, prisoners benefit from positive interactions with staff that do not exist for the sole purpose of maintaining order. Regarding cognitive transformation perspectives on desistance, positive changes in obedience decreased the likelihood of short-term offending. Future research should examine whether positive changes in identity facilitate how individuals get along with staff and family, or whether positive interactions with staff and family help change a person’s sense of self.


Agnew, R. (1992). Foundation for a general strain theory of crime and delinquency. Criminology, 30, 47-88.

Andrews, D. A., & Bonta, J. (2010). The psychology of criminal conduct (5th ed.). New Providence, NJ: LexisNexis.

Andrews, D. A., Bonta, J., & Hoge, R. D. (1990). Classification for effective rehabilitation: Rediscovering psychology. Criminal Justice and Behavior, 17, 19-52.

Andrews, D. A., Bonta, J., & Wormith, J. S. (2006). The recent past and near future of risk and/or need assessment. Crime & Delinquency, 52, 7-27.

Baglivio, M. T., & Epps, N. (2016). The interrelatedness of adverse childhood experiences among high-risk juvenile offenders. Youth Violence and Juvenile Justice, 14, 179-198.

Becker, H. (1963). Outsiders: Studies in the sociology of deviance. New York: Free Press.

Bernburg, J. G., & Krohn, M. D. (2003). Labeling, life chances, and adult crime: The direct and indirect effects of official intervention in adolescence on crime in early adulthood. Criminology, 41, 1287–1318.

Borum, R., Bartel, P., & Forth, A. (2006). Manual for the Structured Assessment for Violence Risk in Youth (SAVRY). Odessa, FL: Psychological Assessment Resources.

Branson, C. E., Baetz, C. L., Horwitz, S. M., & Hoagwood, K. E. (2017). Trauma-informed juvenile justice systems: A systematic review of definitions and core components. Psychological Trauma: Theory, Research, Practice, and Policy, 9, 635-646.

Bushway, S. D., Piquero, A. R., Broidy, L. M., Cauffman, E., & Mazerolle, P. (2001). An empirical framework for studying desistance as a process. Criminology, 39, 491-515.

Cesaroni, C., & Peterson-Badali, M. (2005). Young offenders in custody: Risk and adjustment. Criminal Justice and Behavior, 32, 251-277.

Christensen, L., & Mendoza, J. L. (1986). A method of assessing change in a single subject: An alteration of the RC index. Behavior Therapy, 17, 305-308.

Clarke, R. & Cornish, D. (1985). Modeling offenders’ decisions: A framework for research and policy. In M. Tonry & N. Morris (Eds.), Crime and justice: An annual review of research (v. VI) (pp. 147-185). Chicago, IL: University of Chicago Press.

Collin-Vézina, D., Coleman, K., Milne, L., Sell, J., & Daigneault, I. (2011). Trauma experiences, maltreatment-related impairments, and resilience among child welfare youth in residential care. International Journal of Mental Health and Addiction9, 577-589.

Corrado, R. R., McCuish, E. C., Hart, S. D., & DeLisi, M. (2015). The role of psychopathic traits and developmental risk factors on offending trajectories from early adolescence to adulthood: A prospective study of incarcerated youth. Journal of Criminal Justice, 43, 357-368.

Crank, B. R. (2016). Accepting deviant identities: the impact of self-labeling on intentions to desist from crime. Journal of Crime and Justice, (online first), 1-18.

Cusson, M., & Pinsonneault, P. (1986). The decision to give up crime. Cornish, D.B and R.V. Clarke (Eds.) The reasoning criminal: rational choice perspectives on offending (pp.72–82). New York, NY: Springer.

Dmitrieva, J., Monahan, K. C., Cauffman, E., & Steinberg, L. (2012). Arrested development: The effects of incarceration on the development of psychosocial maturity. Development and Psychopathology24, 1073-1090.

Elder, G. H. (1998). The life course as developmental theory. Child Development, 69, 1-12.

Gilman, A. B., Hill, K. G., & Hawkins, J. D. (2015). When is a youth’s debt to society paid? Examining the long-term consequences of juvenile incarceration for adult functioning. Journal of Developmental and Life-course Criminology1, 33-47.

Giordano, P. C., Cernkovich, S. A., & Rudolph, J. L. (2002). Gender, crime, and desistance: Toward a theory of cognitive transformation. American Journal of Sociology, 107, 990-1064.

Giordano, P. C., Schroeder, R. D., & Cernkovich, S. A. (2007). Emotions and crime over the life course: A neo-Meadian perspective on criminal continuity and change. American Journal of Sociology112, 1603-1661.

Hirschi, T., & Gottfredson, M. (1983). Age and the explanation of crime. American Journal of Sociology, 89, 552-584.

Horney, J., & Marshall, I. H. (1992). Risk perceptions among serious offenders: The role of crime and punishment. Criminology, 30, 575-594.

Kazemian, L. (2007). Desistance from crime: Theoretical, empirical, methodological, and policy considerations. Journal of Contemporary Criminal Justice, 23, 5-27.

Kazemian, L., & Maruna, S. (2009). Desistance from crime. In M.D. Krohn, A.J. Lizotte, & G.P. Hall (Eds.), Handbook on crime and deviance (pp. 277-295). New York, NY: Springer.

Koehler, J. A., Lösel, F., Akoensi, T. D., & Humphreys, D. K. (2013). A systematic review and meta-analysis on the effects of young offender treatment programs in Europe. Journal of Experimental Criminology, 9, 19-43.

Laub, J. H., & Sampson, R. J. (2001). Understanding desistance from crime. In M.H. Tonry (Ed.), Crime and justice: A review of research (pp. 1-69). Chicago, IL: Chicago University Press.

Laub, J. H., & Sampson, R. J. (2003). Shared beginnings, divergent dives: Delinquent boys to age 70. Massachusetts: Harvard University Press.

Le Blanc, M. (1997). A generic control theory of the criminal phenomenon: The structural and the dynamical statements of an integrative multilayered control theory. In T. Thornberry (Ed.), Developmental theories of crime and delinquency: Advances in criminological theory (v. 7). New Brunswick, NJ: Transaction Press.

Lemert, E. (1967). Deviance, social problems and social control. Englewood Cliffs, N.J.: Prentice-Hall.

Levenson, J.S., Willis, G.M., & Prescott, D.S. (2016). Adverse childhood experiences in the lives of male sex offenders: Implications for trauma-informed care. Sexual Abuse28, 340-359.

Loeber, R., & Le Blanc, M. (1990). Toward a developmental criminology. In M. Tonry & N. Morris (Eds.), Crime and justice (pp. 375-473). Chicago, IL: University of Chicago Press.

Lopes, G., Krohn, M.D., Lizotte, A.J., Schmidt, N.M., Vasquez, B.E., & Bernburg, J.G. (2012). Labeling and cumulative disadvantage: The impact of formal police intervention on life chances and crime during emerging adulthood. Crime & Delinquency, 58, 456–488.

Lussier, P., McCuish, E., & Corrado, R. R. (2015). The adolescence–adulthood transition and desistance from crime: Examining the underlying structure of desistance. Journal of Developmental and Life-Course Criminology, 1, 87-117.

Maruna, S. (2001). Making good: How ex-convicts reform and rebuild their lives. Washington, DC: American Psychological Association.

Maruna, S., Immarigeon, R., & LeBel, T. P. (2004). Ex-offender reintegration: Theory and practice. In S. Maruna & R. Immarigeon (Eds.), After crime and punishment: Pathways to offender reintegration, (pp. 3-26). Cullompton: Willan.

Matsueda, R. L. (1992). Reflected appraisals, parental labeling, and delinquency: Specifying a symbolic interactionist theory. American Journal of Sociology, 97, 1577-1611.

Mulvey, E. P., Steinberg, L., Fagan, J., Cauffman, E., Piquero, A. R., Chassin, L., & Losoya, S. H. (2004). Theory and research on desistance from antisocial activity among serious adolescent offenders. Youth Violence and Juvenile Justice, 2, 213–236.

Na, C., Paternoster, R., & Bachman, R. (2015). Within-individual change in arrests in a sample of serious offenders: The role of identity. Journal of Developmental and Life-Course Criminology1, 385-410.

Nguyen H, Loughran TA. 2018. On the measurement and identification of turning points in criminology. Annual Review of Criminology, 1, 335–358

Paternoster, R., Bachman, R., Kerrison, E., O’Connell, D., & Smith, L. (2016). Desistance from crime and identity: An empirical test with survival time. Criminal Justice and Behavior, 43, 1204-1224.

Paternoster, R., & Bushway, S. (2009). Desistance and the" feared self": Toward an identity theory of criminal desistance. The Journal of Criminal Law and Criminology, 99,1103-1156.

Polaschek, D. L. (2012). An appraisal of the risk–need–responsivity (RNR) model of offender rehabilitation and its application in correctional treatment. Legal and Criminological Psychology, 17, 1-17.

Rise, J., Sheeran, P., & Hukkelberg, S. (2010). The role of self-identity in the theory of planned behavior: A meta-analysis. Journal of Applied Social Psychology, 40, 1085-1105.

Rocque, M. (2017). Desistance from crime. New York, NY: Palgrave Macmillan.

Rocque, M., Bierie, D. M., & MacKenzie, D. L. (2011). Social bonds and change during incarceration: Testing a missing link in the reentry research. International Journal of Offender Therapy and Comparative Criminology, 55, 816-838.

Rocque, M., Posick, C., & Paternoster, R. (2016). Identities through time: An exploration of identity change as a cause of desistance. Justice Quarterly33, 45-72.

Sampson, R. J., & Laub, J. H. (1997). A life-course theory of cumulative disadvantage and the stability of delinquency. In T.P. Thornberry (Ed.), Development theories of crime and delinquency (pp. 133-161). New Brunswick, NJ: Transaction Publishers.

Schneider, A. L. (1990). Deterrence and juvenile crime. New York, NY: Springer.

Schneider, A. L., & Ervin, L. (1990). Specific deterrence, rational choice, and decision heuristics: Applications in juvenile justice. Social Science Quarterly71, 585-601.

Shover, N. (1996). Great pretenders: pursuits and careers of persistent thieves. Boulder: Westview Press.

Statistics Canada. (2013). British Columbia (Code 59) (table). National Household Survey Aboriginal Population Profile. Statistics Canada: Ottawa. (accessed July 2017).

Stouthamer-Loeber, M., Loeber, R., Stallings, R., & Lacourse, E. (2008). Desistance from and persistence in offending. In R. Loeber, D. P. Farrington, M. Stouthamer-Loeber, & H. R. White (Eds.), Violence and serious theft: Development and prediction from childhood to adulthood (pp. 269–308). New York, NY: Routledge.

Thomas, C. W., & Foster, S. C. (1972). Prisonization in the inmate contraculture. Social Problems, 20(2), 229-239.

Topalli, V. (2005). When being good is bad: An expansion of neutralization theory. Criminology43, 797-836.

Turner, R. J., & Marino, F. (1994). Social support and social structure: A descriptive epidemiology. Journal of Health and Social Behavior, 35, 193-212.

Tweed, R. G., Bhatt, G., Dooley, S., Spindler, A., Douglas, K. S., & Viljoen, J. L. (2011). Youth violence and positive psychology: Research potential through integration. Canadian Psychology/Psychologie Canadienne, 52, 111-121.

Uggen, C. (2000). Work as a turning point in the life course of criminals: A duration model of age, employment, and recidivism. American Sociological Review, 529-546.

Viljoen, J. L., Beneteau, J. L., Gulbransen, E., Brodersen, E., Desmarais, S. L., Nicholls, T. L., & Cruise, K. R. (2012). Assessment of multiple risk outcomes, strengths, and change with the START: AV: A short-term prospective study with adolescent offenders. International Journal of Forensic Mental Health, 11, 165-180.

Walters, G. D. (2003). Changes in criminal thinking and identity in novice and experienced inmates: Prisonization revisited. Criminal Justice and Behavior, 30, 399-421.

Youth Criminal Justice Act, SC 2002, c 1.

Zingraff, M. T. (1975). Prisonization as an inhibitor of effective resocialization. Criminology, 13, 366-388.

Tables and Figures

Table 1. Descriptive characteristics of adjudicated female and male youth


(n = 29)

% (n)/ M (SD)


(n = 188)

% (n)/ M (SD)

χ2/t, p, Φ/d

Demographic Characteristics



62.1% (18)

50.0% (94)

χ2(2) = 4.45, n.s., Φ = .14


34.5% (10)

30.9% (58)

Non-Indigenous Minority

3.4% (1)

19.1% (36)

Age at Community Reentry

16.62 (1.78)

16.80 (1.44)

t (215) = -0.67, n.s., d = 0.11

Prior Criminal Behavior

Non-Administrative Convictions at Community Reentry

5.00 (3.15)

5.02 (4.17)

t (215) = -0.02, n.s., d = 0.01

Custody Influence

Obedience (T1)

16.93 (4.42)

18.11 (3.74)

t (215) = -1.54, n.s., d = 0.29

Obedience (T2)

19.41 (4.67)

19.12 (4.00)

t (215) = 0.37, n.s., d = 0.07

Prosociality (T1)

24.52 (4.71)

23.69 (4.54)

t (215) = 0.91, n.s., d = 0.18

Prosociality (T2)

25.55 (5.18)

24.52 (4.75)

t (215) = 1.07, n.s., d = 0.21

Views Life as a Success

55.2% (16)

63.8% (120)

χ2(1) = 0.81, n.s., Φ = .06

Family Support in Custody

15.93 (3.39)

16.31 (3.79)

t (215) = -0.51, n.s., d = 0.11

Staff Support in Custody

14.52 (2.31)

13.38 (3.51)

t (50.6) = 2.28, p < .05, d = 0.38

Returning to School upon Reentry

96.4% (27)

80.4% (148)

χ2(1) = 4.32, p < .05., Φ = .14

Finding a Job upon Reentry

92.9% (26)

87.0% (160)

χ2(1) = 0.79, n.s., Φ = .06

Negative Experiences in Custody

34.90 (7.10)

37.60 (7.67)

t (215) = -1.79, n.s., d = 0.37

# of Victimizations in Custody†

0.64 (1.06)

2.32 (3.89)

t (154.4) = -4.80, p < .001, d = 0.59

Months in Custody before Release

5.02 (11.55)

6.70 (9.57)

t (215) = -0.86, n.s., d = 0.16

Short-Term Recidivism

Non-Administrative Recidivism

55.2% (16)

63.8% (120)

χ2(1) = 0.81, n.s., Φ = .06

Violence Recidivism

21.4% (6)

26.6% (50)

χ2(1) = 0.34, n.s., Φ = .04

Longer-Term Reconvictions

Non-Administrative Convictions†

3.79 (4.07)

5.49 (5.57)

t (46.0) = -1.98, n.s., d = 0.35

Violent Convictions†

0.76 (1.15)

1.32 (1.85)

t (53.2) = -2.22, p < .05, d = 0.36

Note. Φ = phi; d = Cohen’s d; n.s.= not significant (p ≥ .05)

† Levene’s test violated

Table 2. Logistic regression models with non-administrative reconviction and violent reconviction as the outcomes of interest

Non-Administrative Reconviction

Model 1

Violent Reconviction

Model 2


(95% C.I.)


(95% C.I.)

Demographic Characteristics


1.74 (0.64-4.70)

1.25 (0.38-4.10)

Ethnic Group1


0.87 (0.40-1.91)

1.18 (0.51-2.75)

Non-Indigenous Minority

0.60 (0.24-1.49)

0.89 (0.30-2.61)

Age at Community Reentry

0.69 (0.51-0.94)*

0.91 (0.66-1.26)

Prior Criminal Behavior

Non-Administrative Convictions at Community Reentry

1.13 (1.03-1.24)**

1.21 (1.09-1.34)***

Custody Influence on Desistance

Positive Change in Obedience

0.21 (0.06-0.77)*

0.57 (0.10-3.20)

Positive Change in Prosociality

2.14 (0.61-7.47)

0.41 (0.09-1.83)

Views Life as a Success

0.91 (0.44-1.87)

1.19 (0.51-2.78)

Family Support Scale

0.96 (0.88-1.06)

0.90 (0.82-1.00)*

Custody Staff Support Scale

0.95 (0.85-1.06)

0.86 (0.76-0.96)*

Returning to School upon Reentry

0.28 (0.09-0.82)*

0.58 (0.23-1.44)

Finding a Job upon Reentry

2.51 (0.86-7.31)

4.34 (1.20-15.65)*

Negative Experiences in Custody

0.97 (0.93-1.02)

0.98 (0.94-1.03)

# of Victimizations in Custody

1.19 (1.03-1.37)*

1.18 (1.06-1.32)**

Months in Custody before Release

0.93 (0.86-1.00)*

0.97 (0.90-1.04)

Model fit







Note. * p < .05 ** p < .01 *** p < .001; OR = odds ratio. Odds ratios that are significant despite containing “1” are due to rounding

1 ‘White’ is reference category

Table 3. Negative binomial regression models with frequency of non-administrative reconvictions and violent reconvictions as the outcomes of interest

Non-Administrative Reconviction

Model 1

Violent Reconviction

Model 2


(95% C.I.)


(95% C.I.)

Demographic Characteristics


2.40 (1.36-4.25)**

3.38 (1.55-7.38)**

Ethnic Group1


0.82 (0.53-1.25)

1.12 (0.65-1.94)

Non-Indigenous Minority

0.81 (0.49-1.34)

0.81 (0.41-1.60)

Age at Community Reentry

0.79 (0.68-0.92)**

0.77 (0.63-0.94)**

Prior Criminal Behavior

1.08 (1.03-1.13)**

1.10 (1.03-1.17)**

Custody Influence on Desistance

Positive Change in Obedience

0.90 (0.45-1.81)

1.67 (0.66-4.21)

Positive Change in Prosociality

1.92 (1.05-3.55)*

0.94 (0.41-2.14)

Views Life as a Success

0.64 (1.00-1.12)*

0.71 (0.43-1.19)

Family Support Scale

0.97 (0.92-1.02)

1.02 (0.95-1.08)

Custody Staff Support Scale

0.95 (0.90-1.01)

0.92 (0.85-0.99)*

Returning to School upon Reentry

0.99 (0.61-1.62)

0.84 (0.45-1.56)

Finding a Job upon Reentry

1.21 (0.68-1.62)

1.37 (0.64-2.97)

Negative Experiences in Custody

0.99 (0.96-1.01)

0.96 (0.94-0.99)*

# of Victimizations in Custody

1.06 (0.43-0.95)*

1.06 (0.99-1.12)

Months in Custody before Release

0.99 (0.89-1.09)

0.99 (0.88-1.13)

Model fit







Note. * p < .05 ** p < .01 *** p < .001; IRR = incident rate ratio

1 ‘White’ is reference category

Appendix A: Description of scales capturing desistance that were included in the current study

Name of Instrument

# of items

Item Scale

Cronbach’s Alpha

Example Items

Obedience (Good Citizen Scale)



T1 = 0.64

T2 = 0.69

I would like you to tell me which word you think best describes you:

Breaks Rules (1)/Obeys Rules(7)



Prosociality (Good Citizen Scale)



T1 = 0.68

T2 = 0.76

I would like you to tell me which word you think best describes you:

Bad (1)/Good(7)



Family Support Scale




Members of my family (including my parents, brothers and sister, grandparents) …

…let me know that they cared for me

…let me know that I’m a good and worthwhile person

…let me know that things were going to get better

…let me know that I could count on them if I needed help

Custody Staff Support Scale




Staff in here…

…let me know that they cared for me

…let me know that I’m a good and worthwhile person

…let me know that things were going to get better

…let me know that I could count on them if I needed help

Negative Experiences in Custody




Can you tell me how difficult it has been for you to deal with the following situations since coming to this Youth Custody Centre?

Being away from your parents

Being away from other family members (e.g., siblings, grandparents)

Being away from your close friends including boy/girl friend

Not having all your personal belongings like clothes or stereo

Correlation Matrix for all Measures in the Study













Non-Administrative Convictions (1)













Positive Change in Obedience (2)












Positive Change in Prosociality (3)











Views Life as a Success (4)










Family Support Scale (5)









Custody Staff Support Scale (6)








Returning to School upon Reentry (7)







Finding a Job upon Reentry (8)






Negative Experiences in Custody (9)





# of Victimizations in Custody (10)




Months in Custody before Release (11)



Age at release (12)


For complete scales developed by the principal investigator, please contact the first author

Notes. All correlations are two-tailed Pearson or point-biserial depending on level of measurement. * p < .05 ** p < .01 *** p < .001

No comments here
Why not start the discussion?