Skip to main content
SearchLoginLogin or Signup

Cohort Profile: The Incarcerated Serious and Violent Young Offender Study

Published onApr 20, 2022
Cohort Profile: The Incarcerated Serious and Violent Young Offender Study
·

Abstract

In this profile, we describe how the Incarcerated Serious and Violent Young Offender Study (ISVYOS) leveraged detailed administrative data to create a prospective longitudinal study. We also discuss the research and policy context at the time the ISVYOS was initiated, its methodology, and what has been learned so far. The ISVYOS includes 1,719 participants who were incarcerated in adolescence and followed through adulthood. Between 1998-2002 (Cohort I) and 2005-2011 (Cohort II), male and female youth were recruited for interviews that measured various risk and protective factors, experiences in custody, and attitudes towards the justice system. Follow-up data include measures of justice system involvement, risk assessment data, and social network data. Participants were an average age of 30.57 (SD = 4.65) as of December 2019. Over 90% of the sample was involved in the adult justice system. So far, research using ISVYOS data have addressed the reliability and validity of measures of psychopathy in adolescence, justice system criminal careers, risk factors for chronic and persistent offending, and the adult offending outcomes of youth involved in serious offenses (e.g., sexual offenses, homicide offense). Inquiries regarding data access should be sent to the first author.

Keywords: Administrative data; Incarcerated Serious and Violent Young Offender Study; Longitudinal research; Psychopathy; Youth justice; Youth offending


The Journal of Developmental and Life-Course Criminology uses cohort profile submissions to allow authors to summarize longitudinal studies1. The Incarcerated Serious and Violent Young Offender Study (ISVYOS) is a longitudinal study of 1,719 persons who experienced incarceration in adolescence. A description of the ISVYOS is timely given that the study has transitioned to its next phase, which involves learning about the adult outcomes of youth involved in serious and violent offenses and who come from extremely marginalized circumstances. Developmental and life-course (DLC) criminology longitudinal studies tend to sample from community populations (Farrington, 2015), leaving a gap regarding what is known about youth in the justice system (Mulvey et al., 2004). Canada’s youth justice system is different from most jurisdictions in the United States. For example, Canada’s emphasis that youth experience rehabilitation rather than deterrence while incarcerated is distinct from most juvenile justice systems in the United States (Scott & Steinberg, 2009). ISVYOS participants likely had a different youth justice system experience compared to youth who participated in, for example, the Pathways to Desistance Study (Mulvey et al., 2004). Understanding the youth justice system’s role in preventing or worsening negative outcomes in adulthood is a critical part of life-course criminology that has remained largely unaddressed (see Neil & Sampson, 2021). The ISVYOS provides an opportunity to learn about the adult outcomes of youth in the justice system who perpetrated serious and violent offenses. For example, 186 participants were involved in a sex offense and 145 participants were involved in a homicide offense.

This cohort profile does not present new data addressing a specific research gap. Although key findings from past ISVYOS research are discussed, our main goal was to reflect on themes typically not discussed in empirical articles. Conducting prospective longitudinal research with youth in the justice system poses unique challenges. It is important to communicate the research decisions made to address these challenges so that readers better understand the data. One research decision was to collect prospective longitudinal data from an administrative data source. Historically, issues with administrative data include missingness, inconsistent recording of information, and the failure to measure key constructs beyond demographic characteristics and offense history. We highlight the depth of information available from administrative records and encourage researchers to reach out to local agencies to identify whether their data can be leveraged similarly. The administrative data of 30 years ago are not the administrative data of today2. Changes in technology, data storage capabilities, and improved education and training of justice system practitioners resulted in the ISVYOS having access to detailed records that included repeated measures of key constructs like informal social control. By virtue of their repeated involvement in the justice system, ISVYOS participants were well-known to practitioners (e.g., correctional officers, probation officers, social workers) and were the subject of numerous reports. Thus, practitioners had valuable insight into their attitudes, behavior, experiences, and social environment over the life-course. Information from the administrative data source is timestamped, allowing for the creation of a longitudinal dataset. Below, we describe why the ISVYOS was initiated, its methodology, key findings, strengths and weaknesses, and how to access the data.

The ISVYOS: Why was the Cohort Set Up?

During the 1980s and 1990s, the youth crime rate was on the rise in Canada and the United States (Ouimet, 1993). Discussion regarding how to respond to youth crime mostly occurred within political and policy spheres and was influenced by media coverage of youth involved in atypically serious offenses (Corrado & Markwart, 1994). Police groups publicly denigrated rehabilitative approaches to youth justice and, based on the populist viewpoint that punishment made youth more accountable, claimed “now it is our turn, the punishment people. Punishment works” (Corrado & Marquart, 1994, 345). Canada’s Federal Government decided that a more punitive youth justice system was needed. Similar conclusions were made in the United States (Feld, 1993).

While some policymakers, law enforcement officials, and members of the public were asking why today’s serious and violent youth was tomorrow’s serious and violent adult, paradoxically, DLC criminologists were asking why there was discontinuity in antisocial and criminal behavior between different age-stages (Laub & Sampson, 1993; Loeber & LeBlanc, 1990). Whether this was also true for youth involved in serious and violent offenses received less attention (c.f., Le Blanc & Fréchette, 1989) because popular theories of delinquent behavior assumed that the causes of offending were the same for all persons (Gottfredson & Hirschi, 1990; Hirschi, 1979; Sutherland, 1947). As noted by Cullen (2011), this gave some criminologists the impression that research on incarcerated youth was not necessary. When sensational claims about youth “superpredators” (DiIulio, 1995) entered policy discussions, criminologists were not well-positioned to answer questions about youth involved in serious and violent offenses. The lack of data on youth involved in serious and violent offenses was a catalyst for the ISVYOS.

ISVYOS data collection began in 1998 in the province of British Columbia, Canada. Encouragement and permission to conduct the study was provided by the Ministry of Children and Family Development, which is responsible for the administration of the youth justice system in the province3. Ministry officials were concerned that youth justice legislation changes that increased the punitiveness of the act (Bell, 2014) were made in the absence of data on incarcerated youth. The Ministry, which is responsible for the care of incarcerated youth, consented to the ISVYOS inviting residents of custody centers throughout the province to participate in the study.

ISVYOS data collection was primarily funded by Canada’s Social Sciences and Humanities Research Council, which supports criminological research designed to be completed within 1-4 years. A requirement for funding is that researchers demonstrate how their new project is distinct from their past research. This is a barrier to collecting prospective longitudinal data where an essential condition is the repeated measurement of the same constructs. This partially explains the lack of longitudinal research on crime and delinquency in Canada (c.f., Le Blanc & Fréchette, 1989). It was necessary for ISVYOS grant proposals to balance the collection of new data while justifying continued collection of core measures (e.g., offending, informal social controls) across different age-stages. So far, there have been three main phases to the ISVYOS.

Phase 1 (1998-2002) occurred during Canada’s Young Offenders Act. Participants from this phase are referred to as Cohort I. Boys and girls in various custody centers throughout the province of British Columbia participated in a structured intake interview providing a snapshot of the risk/need profiles of incarcerated youth (Corrado et al., 2000). Participants who completed an intake interview were eligible for an additional interview used to score the Psychopathy Checklist: Youth Version (PCL:YV). This was a starting point for addressing the lack of reliable and valid measures available to guide practitioners’ judgment regarding psychopathology and associated intervention strategies for incarcerated youth (Corrado et al., 2004).

Phase 2 (2005-2011) emerged after the Young Offenders Act was replaced by the Youth Criminal Justice Act. A main goal of this new act was to reduce Canada’s reliance on custody while also increasing the amount of time spent incarcerated for those who received a custody sentence (Bala et al., 2009). The change in legislation allowed the ISVYOS to examine whether risk/need profiles and adult offending outcomes varied depending on the youth justice system under which a participant was adjudicated (see McCuish et al., 2021). This helps with examining the social and scientific developments that provide the backcloth to individual development. Having two cohorts that experienced different youth justice systems allows for the examination of questions regarding whether it is “who a person is” or “when a person is” that most strongly impacts justice system involvement (Neil & Sampson, 2021). Participants recruited during this second phase are referred to as Cohort II and received an expanded intake interview that included new measures of school and family life, motivations for offending, crime sequences, gang activity, and mental health.

Phase 3 began in 2011 and was implemented to address the heart of the DLC paradigm: longitudinal research. Rarely are youth justice system practitioners aware of which youth continue to offend in adulthood. For multi-need youth, it is not always clear which risk factor should be intervened on to prevent offending in adulthood. Phase 3 leveraged administrative data to link the cross-sectional data from Phase 1 and Phase 2 with follow-up data consisting of repeated measures of adult outcomes, including offending, risk assessment data that included measures of informal social controls (e.g., employment, intimate partner relationships) and criminogenic attitudes (e.g., moral disengagement), and social networks.

Who is in the Cohort?

The ISVYOS includes 596 participants from Cohort I and 1,123 from Cohort II (n = 1,719). Among those with available data, there are 1,339 male participants (78.3% of the sample) and 372 female participants (21.7%). The data, therefore, are useful for addressing research gaps with respect to the longitudinal trajectories of justice system involvement among girls and women (Ahonen et al., 2016). Most participants are White (54.8%), but Indigenous youth are overrepresented in the sample (29.6%) relative to the general population (Statistics Canada, 2013). The ISVYOS allows for the use of a DLC perspective to examine the unique challenges and structural barriers faced by Indigenous Persons, who have been neglected by criminology research (Piquero, 2015). The remaining sample (15.6%) self-identified as being part of a non-Indigenous ethnic minority group (e.g., Asian, Indian/South-East Asian, Black). Participants were an average age of 16.06 (SD = 1.28; Range = 12-19) at the time of their interview. Participants lived throughout the province of British Columbia, which covers 587,031 square miles (approximately 1.5 times the size of Texas). At the time of data collection, there were approximately 250,000 youth living in the province. About 1 in 1,500 of these youth experienced incarceration (Statistics Canada, 2020), meaning that such youth rarely appeared in community samples.

Procedures

Most participants were recruited from the province’s two largest custody centers, located in Burnaby (a suburb of Vancouver) and Victoria (the province’s capital city). In Canada, youth are never placed in adult custody, meaning that the ISVYOS had access to all youth who experienced incarceration. To recruit participants, research assistants (RAs) approached incarcerated youth while they were residing in their living unit at the custody center. Interviews were conducted in a closed-door, private room away from other residents and custody staff. To obtain assent, participants were read and given a copy of an information sheet explaining the purpose of the study, how information would be collected, and that confidentiality would be maintained unless a participant made a direct threat to hurt themselves or someone else. Of the 1,719 participants, 1,548 (90.1%) had data from the intake interview available. The remaining 9.9% included youth who declined to be interviewed (n = 149; 8.7%) or whose intake interview was unavailable due to administrative issues (e.g., a lost interview instrument or an unreliable interview; n = 22; 1.3%). Missing data occurred when participants failed to finish the interview (n = 158). Further, because research funding required addressing new themes, certain questions from the intake interview were only asked of Cohort I and not Cohort II, or vice-versa.

Follow-Up Data. Phase 3 did not include interviews administered by ISVYOS RAs. Access to incarcerated youth required the permission of the Ministry of Children and Family Development, which is responsible for the care of all incarcerated youth. Youth can only assent to participate in research. Therefore, the Ministry’s consent was required for the project to proceed. The Ministry could not provide consent beyond adolescence and so the research agreement specified that participants were not to be contacted following their involvement in the youth justice system. Another barrier to directly collecting self-report data in adulthood was that participants were from various cities throughout the province. This distance made it impractical to conduct follow-up interviews. Ensuring the safety of RAs was another barrier to completing follow-up interviews. Participants were the targets of gang-related homicide and police duty-to-warn notifications which made conducting interviews in public settings dangerous. As an alternative, administrative data were coded to create a longitudinal dataset.

In Phase 3, the ISVYOS gathered follow-up information by accessing administrative data through the British Columbia Corrections Network (CORNET). CORNET is an electronic software program for client management and supervision. It includes data on all individuals involved in the provincial justice system in British Columbia. Each participant has a searchable electronic file with a series of tabs corresponding to different sources of information, including criminal history and risk assessment data, as well as time stamps for when this information was recorded. CORNET files were retrieved for 1,620 participants (94.2% of the sample). CORNET files were unavailable in instances where a participant’s name or correctional service number was inaccurately recorded (n = 29; 1.7%) or because their record was sealed (n = 70; 4.1%). CORNET only covers justice system involvement in British Columbia. As of 2020, 7.4% of the sample (n = 127) moved outside the province (Mean age = 23.60; SD = 5.46); however, 34.7% of this group (n = 44) moved back to British Columbia and so data were missing only temporarily. CORNET includes information from a participant’s time in both the youth and adult justice system. The use of administrative data to facilitate multi-variable longitudinal research is not new, but this work is primarily in fields like epidemiology (Drefahl et al., 2020; Waldenlind et al., 2014; Wall-Wieler et al., 2017; Zylbersztejn et al., 2018). The level of detail in justice system administrative data is advancing and criminologists may be able to use these data to create cost-effective longitudinal studies. Table 1 summarizes the types of information available on CORNET, the source of the information, and how the information was used to create a longitudinal dataset.

--Insert Table 1 about Here--

CORNET captures the moment a participant becomes involved in the justice system and is updated by practitioners while the participant is serving a community sentence or while incarcerated. New justice system involvement was not necessary for new information to be added to CORNET. For example, a participant who received a three-year probation sentence would continue to be regularly assessed by their probation officer and this information is stored on CORNET. Most research ethics agreements do not allow researchers to contact persons familiar with a participant. A benefit of relying on justice system practitioners for follow-up data was that their job involved regularly contacting, interviewing, and incorporating information from collateral contacts. For example, probation officers directly interview their client, but they can also contact practitioners who worked with their client in the past, incorporate information provided in psychopathology assessments, speak with their client’s family members, intimate partners, and so on. This information is then incorporated into, for example, the completion of risk assessment tools stored on CORNET and coded by ISVYOS RAs.

How Often Have They Been Followed Up?

Research ethics agreements, the need to protect participants’ identities, funding strategies, and safety concerns prevented RAs from conducting new follow-up interviews in the community. However, it was possible to collect follow-up information from adolescence through adulthood because CORNET data include timestamps (see Table 1). Criminal history, risk assessment, and social network data are the main types of follow-up data collected so far and data collection remains ongoing. These data are coded to facilitate the examination of outcomes at yearly intervals (e.g., number of convictions at age 12, at age 13, and so on). The length of the follow-up period depends on the type of data collected because it was necessary to submit different research grants at different timepoints for each type of data. RAs began collecting criminal history data in 2011, risk assessment data in 2016, and social network data in 2019. CORNET includes historical records and so data prior to these dates were also available and coded. At the most recent wave of data collection (December 2019), for criminal history data, the average age of participants was 30.57 (SD = 4.65). The average age of participants at the time of their last risk assessment is 24.15 (SD = 3.64). For social network data, which is the newest endeavor for the ISVYOS, for the participants coded so far (n = 99), the average age at follow-up is 26.65 (SD = 1.97).

What has been Measured?

Interview Data

Figure 1 summarizes the number of participants who participated in each interview.

--Insert Figure 1 about Here--

Intake Interview. The intake interview included measures of demographic characteristics, family structure, substance use, self-reported offending, family dysfunction, mental health, self-identity, and fear and safety in prison. The intake interview for Cohort II added an English translation of LeBlanc et al.’s (1996) Mesures de l’Adaptation Sociale et Personnelle pour les Adolescents Québécois, which included measures of antisocial behavior from a variety of domains, including at home, at school, and in the community. Descriptive statistics for some of these measures are shown in Table 2. Part of the aim in describing these data is to demonstrate the severity of the sample in terms of risk factors and patterns of justice system involvement. A codebook on the ISVYOS ResearchGate site (https://www.researchgate.net/project/The-Incarcerated-Serious-and-Violent-Young-Offender-Study) includes descriptive statistics for all variables from the intake interview.

--Insert Table 2 about Here--

School questions included whether participants attended school prior to being incarcerated, whether they skipped school (including the age they first did so), the number of times they changed schools for reasons other than grade changes (e.g., moving, expulsion), and whether they engaged in 21 different school behavioral problems (e.g., drinking, drug use, fighting, talking back to teachers). Participants frequently changed schools, most were not attending school prior to incarceration, and a quarter reported skipping school before age 12. The regularity with which participants were either not attending school, enrolled in an alternative school, or repeatedly skipping school reiterates that sampling from schools likely means excluding youth who experience incarceration.

Measures of substance use included participants’ history of using nine different substances: alcohol, marijuana, hallucinogens (e.g., mushrooms, acid), ecstasy, cocaine, crack cocaine, heroin, crystal methamphetamine, and illegal use of prescription pills. On average, participants used about five different drugs. Hard drug use (at least one of crack cocaine, heroin, or crystal methamphetamine), was reported by half the sample and nearly 75% of girls.

Family questions included whether participants were ever kicked out of their home for at least 24 hours (and at what age) and whether they ever left their home by their own choice to live somewhere else for at least 24 hours (and at what age). Girls were significantly more likely to have left their family home to live somewhere else. Participants were asked whether a family member (and which one) had trouble with or experienced alcohol abuse, drug abuse, physical abuse, sexual abuse, a criminal record, and a mental health problem. These indicators were summed to create a family dysfunction scale, which girls scored significantly higher on compared to boys.

Regarding sexual development and abuse, participants were asked about whether they had engaged in sexual activity, whether they experienced physical abuse, and whether they experienced sexual abuse. For each, participants were asked about age of onset. About half the sample reported physical abuse and 20% reported sexual abuse. Girls were significantly more likely than boys to have experienced physical and sexual abuse.

Identity was measured using the Good Citizen Scale (Schneider, 1990). Participants were asked to rate themselves from 1-7 on 15 items reflecting two opposing traits (e.g., Good/Bad, Rude/Polite). Using the same items, youth were asked to rate how they think other people perceive them. Girls reported a significantly less positive self-identity compared to boys.

Supplemental Interviews. Both cohorts received an interview used to score the PCL:YV. Cohort II also received an interview used to score the CAPP-IRS. The inclusion of reliable and valid multi-item measures of criminal propensity helped address concerns that DLC criminology is sacrificing measurement quality in favor of longitudinal data (Cullen et al., 2019). The Exit interview measured participants’ experiences in custody, including who visited them, the treatment they participated in, victimization, identity, and fear and safety in prison. The Millon Adolescent Clinical Inventory (Millon & Davis, 1993) measured broad indicators of psychopathology like anxiety. Youth gang and foster care interviews explored the details of participants’ experiences as a gang member and child in care.

Administrative Data from CORNET

Official Offending. CORNET contains information pertaining to each person’s movement in and out of custody, the custody center a person resided in, the type of offense committed, date of conviction, court location, and sentence type and length. The number of convictions incurred were recorded for each year of age, beginning at age 12, which is the age of criminal responsibility in Canada. Convictions were coded into one of seven crime-types: violent, property, violation, weapon, drug, miscellaneous, and sexual. Incarceration facilities included sheriff’s cells, remand/pre-trial centers, and centers for those sentenced to custody. The date of each admission to and release from custody was recorded to measure the number of days spent incarcerated at each year of age. It was not a matter of if, but rather how often, members of the ISVYOS re-entered the justice system (Table 2). Boys averaged significantly more convictions and days incarcerated compared to girls. Girls averaged a significantly earlier age at first conviction. This may relate to judges’ tendencies under the Young Offenders Act to respond more punitively to girls as part of protectionist sentencing philosophies (Corrado et al., 2000). In adolescence, 60% of the sample was convicted of a hands-on violent offense and 18.5% were convicted of weapons possession.

Risk Assessments. Trained justice system practitioners interviewed participants for the purpose of scoring various risk assessment tools specific to adolescence and adulthood age-stages that were completed as part of court-mandated reports and case management planning. This information was stored and timestamped on CORNET. This allowed the ISVYOS to link developmental data in adolescence with risk assessment data in adulthood. For brevity, we focus on the adult community risk/need assessment (ACRNA), which is regularly completed as part of case management practices (Gress, 2010). Practitioners rate their clients on nine dynamic factors. Five of these items measure informal social controls (family relationships, living arrangements, intimate relationships, vocational skills, employment patterns. The remaining items measure financial management, behavioral and emotional stability, criminogenic attitudes, and substance abuse. Items are scored on a 0-3 scale indicating the extent to which an item functionally impairs a client’s path to desistance. The ACRNA has been coded for 868 participants who averaged 5.89 (SD = 4.20) assessments (see Figure 2). This allowed for the examination of, for example, whether within-individual changes dynamic factors predicted within-individual changes in offending (McCuish et al., 2021). Data collection is ongoing and therefore sample sizes in Figure 2 are underestimated, especially at later ages.

--Insert Figure 2 about Here--

Social Network Data. Social network analysis is useful for testing principles of theories that specify the importance of interrelationships among people, including those who co-offend together (Bouchard, 2020). We became interested in social network data after discovering that a participant, who was the leader of a prominent gang, was involved in two separate homicide offenses approximately ten years apart. One of his co-offenders in the first homicide was also his co-offender in the second homicide. Another co-offender from the first homicide was a victim in the second homicide (McCuish et al., 2015). From this case study, we examined whether networks could be coded for all participants. In reading participants’ CORNET files, it was clear that participants knew one another. They co-offended together, socialized together in the community, dated, hung out in prison, and offended against one another.

In 2019, the ISVYOS received funding to complete social network coding for all participants. Especially when participants are incarcerated, social network information is provided daily, sometimes every few hours. Timestamps help establish temporal order between network ties and other outcomes, including offending. Network data have been coded so that researchers can examine changes in networks at each year of age. Data from CORNET facilitated the coding of the criminogenic networks of participants, where each participant represented an “ego” and all persons connected to them represented “alters”. Coding allowed for the inspection of participants’ network characteristics, including degree centrality (i.e., a participant’s number of criminogenic connections), betweenness centrality (i.e., the extent to which a participant falls along the shortest path between unconnected individuals), and network density (i.e., the percentage of all possible ties present in a participant’s network). The primary sources of information were community and prison logs and CORNET alerts. CORNET alerts provided correctional officials with information on, for example, who an individual co-offended with or had conflict with, including whether the conflict was mutual or one-way. These alerts identified when participants were the subject of a police duty-to-warn notification due to their risk of perpetrating gun violence or being a target of a gang-related ‘hit’. Papachristos et al. (2015) showed that social network data were valuable for understanding risk of victimization and perpetration of gun violence. The longitudinal component of the ISVYOS has helped identify a reciprocal relationship between social networks and such outcomes (Ryu & McCuish, 2022).

The type of tie between egos and alters were categorized as either co-offending, social, mutual conflict, or victimization. Social ties were coded only if the alter was also involved in criminal behavior. For community ties, the city that a tie was formed in was coded to provide information on residential mobility among participants. For prison ties (e.g., socializing with a cellmate, conflicts with others, prison-based co-offending), the name of the institution was recorded to facilitate the examination of networks within and between prison settings. As part of a study funded by Public Safety Canada, criminogenic networks were coded for 99 participants. The egonetworks of these seeds revealed connections to more than 2,000 unique alters (Ryu & McCuish, 2022), demonstrating that participants were well-connected to others in conflict with the law. The magnitude of participants’ criminogenic social networks and the time required to code them (approximately eight hours per participant) highlight the level of detail on CORNET. A sociogram of the criminogenic networks of this subsample showed that they were virtually entirely connected by a single network component, highlighting the interdependent nature of the sample.

What has it Found? Key Findings and Publications

Four main themes have been addressed using ISVYOS data: (1) the reliability and predictive validity of features of psychopathy in adolescence, (2) justice system criminal careers through adulthood, (3) youth risk and protective factors associated with adult outcomes, and (4) the relationship between specific crime types and future offending in adulthood. Findings contrasted with widely accepted conclusions about offending that emerged from community-based samples (Farrington, 2003), which helped illustrate the unique needs of incarcerated youth.

Regarding psychopathy, data from the ISVYOS helped show that the PCL:YV had high interrater reliability, internal reliability, and construct validity. These data were used to help develop the PCL:YV manual (Forth et al., 2003). PCL:YV scores predicted recidivism (Corrado et al., 2004), but not for girls (Vincent et al., 2008). The CAPP-IRS, which measures a more exhaustive set of features of psychopathy compared to the PCL:YV, was adapted for use with youth and was identified as a reliable and valid measure of psychopathy (Dawson et al., 2012; McCuish et al., 2019a, 2019b). Contrary to assertions that risk factors in childhood and adolescence were uninformative of persistent offending in adulthood (Laub & Sampson, 2003), PCL:YV scores predicted chronic general offending (McCuish et al., 2014) and violent offending (McCuish et al., 2015b) nearly two decades later, even when controlling for past offending and other risk factors (Lussier et al., 2020).

Regarding justice system criminal careers, modeling conviction trajectories showed that participants were involved in such a high rate of offending that controlling for exposure time (e.g., time incarcerated) was insufficient for avoiding false desistance, which is where an individual appears to desist solely because of inopportunity to offend due to lengthy periods of incarceration (McCuish, 2020). Repeated measures of crime-types allowed for the examination of the dynamic nature of crime versatility and specialization. Increases in offending specialization were associated with desistance (Lussier et al., 2017). Contrary to assertions that offending declines in adulthood even for youth involved in the highest rate of offending (Laub & Sampson, 2003), for about 30% of ISVYOS participants, once accounting for time incarcerated, their rate of convictions escalated after adolescence. Further, the number of convictions needed to meet the threshold for a pattern of chronic offending in mature adulthood (ages 30-35) resembled the threshold for being identified by a pattern of chronic offending in adolescence (McCuish et al., 2021). The inexorable effect of aging was not observed for this sample. Contrary to Laub and Sampson’s (2003) assertion that the relationship between past and future offending grew weaker with age, for ISVYOS participants, the relationship between age of onset and convictions in mature adulthood was stronger than the relationship between age of onset and convictions in emerging adulthood (McCuish et al., 2021).

Regarding the relationship between youth risk factors and adult outcomes, foster care (Yang et al., 2020), fetal alcohol spectrum disorder (Corrado & McCuish, 2015), and self-identity (McCuish et al., 2018b) were found to increase ISVYOS participants’ risk of persistent justice system involvement. For girls, a negative self-identity and running away from home were informative of persistent justice system involvement (Gushue et al., 2021). McCuish et al. (2021) found that higher levels of risk in adolescence decreased the likelihood of positive sources of informal social control in adulthood (i.e., selection effects). Further, although informal social controls in adulthood influenced desistance, a series of moderation analyses showed that informal social controls did not influence desistance for higher-risk sample members, such as those who spent lengthy periods of time incarcerated in adolescence (i.e., treatment effect heterogeneity). Such findings contrasted with assertions about informal social controls as a normative part of the life course that influenced desistance regardless of a person’s past (Laub & Sampson, 2003).

Regarding specific crime-types, youth involved in sexual offenses resembled youth involved in non-sexual offenses in terms of patterns of (1) antisocial behavior prior to age 12 (McCuish et al., 2015c), (2) conviction trajectories through adulthood (McCuish et al., 2016; Reale et al., 2019), and (3) likelihood of sexual offending in adulthood (e.g., Lussier et al., 2016). Despite policy and political discourse surrounding ‘out-of-control’ youth (Corrado & Markwart, 1994), McCuish et al. (2018a) found that the risk of recidivism in adulthood was similar between youth involved in homicide offenses and youth involved in violent non-homicide offenses.

What are the Main Weaknesses and Strengths?

The ISVYOS lacks data on the percent of the youth custody population we interviewed. Nevertheless, the gender and ethnic composition of the sample resembles data on youth custody at the population-level (Malakieh, 2017). Differences in youth justice legislation should be considered before generalizing ISVYOS findings to other incarcerated youth. For example, Canada’s youth justice system handles all youth in conflict with the law, regardless of the severity of their offense. This differs from jurisdictions in the United States where youth involved in especially serious offenses are typically transferred to the adult system, are incarcerated with adults, and thus excluded from samples of adjudicated youth (Loughran et al., 2009). A main weakness is that the ISVYOS departs from traditional cohort studies because it did not use RAs to administer follow-up interviews. Instead, follow-up data were extracted from administrative files that included interviews and risk assessments completed by various justice system practitioners when participants were in adolescence and adulthood. The level of detail and repeated implementation of risk assessment tools allowed the ISVYOS to remain true to DLC principles while navigating the unique context of conducting research with high-risk, marginalized youth in need of protection. Although data in adulthood were available for virtually all participants, those who were repeatedly involved in the justice system have more data available to code.

The use of administrative data did not preclude examining within-individual change. For example, ISVYOS data were used in fixed-effects and cross-lagged dynamic panel modeling to examine within-individual change in informal social control across emerging adulthood (McCuish et al., 2021). We have also used these data to examine how serious victimization impacts change in criminogenic networks between adolescence and adulthood (Ryu & McCuish, 2022). The ISVYOS shows that administrative data can answer DLC criminology questions within a group of individuals rarely found in community-based studies but whom are repeatedly in contact with the justice system. This includes questions about selection effects, state dependence, treatment effect heterogeneity, and the magnitude of stability or change in offending over the life course (McCuish et al., 2021). The ISVYOS was initiated to answer questions held by youth justice system practitioners and ongoing partnerships with various local agencies helped ensure that findings were communicated to key stakeholders.

Can I get Hold of the Data? Where can I Find out More?

Due to ethics requirements, the data are not publicly available but can be acquired by researchers seeking to verify ISVYOS findings or to conduct independent investigations of research questions not being actively addressed by members of the ISVYOS research team. Instruments and coding forms are also available by request. Requests should be sent to the Principal Investigator (first author). Researchers must sign an access agreement that includes an acknowledgement that the data will not be used for reasons other than those agreed to by the Principal Investigator. All ISVYOS publications data are posted to its ResearchGate Profile.

Profile in a Nutshell

  • Why the cohort was set up and/or unique feature(s) of the cohort.

    • Policymakers in British Columbia wanted to know more about youth involved in serious and violent offending. The ISVYOS was supported by provincial youth justice practitioners to address the lack of knowledge about such youth.

  • Location, year(s) of baseline data collection, number of participants at baseline, composition of the study population including age range.

    • The ISVYOS began in 1998 and recruited male and female youth who were incarcerated in various custody centers throughout British Columbia, Canada (n = 1,719). Members were between ages 20-40 (M = 30.57; SD = 4.65) at the December 2019 wave of criminal history data collection. Data collection is ongoing.

  • Frequency of follow-up, attrition, number of participants currently in the cohort.

    • Follow-up data are available for 1,625 participants (94.5% of the sample). For some individuals, these data are incomplete because they died (n = 131) or moved outside the province (n = 128). Criminal history and social network data are coded for each year of age. Risk assessment data are coded when made available by practitioners.

  • Main categories of data collected.

    • Psychopathy, substance use, self-identity, abuse and exposure to violence, criminal history data, informal social controls, and social network data.

  • Collaboration and data access.

    • Please contact the first author regarding data access.

References

Ahonen, L., Jennings, W., Loeber, R., & Farrington, D. (2016). The relationship between developmental trajectories of girls’ offending and police charges: Results from the Pittsburgh Girls Study. Journal of Developmental and Life-Course Criminology, 2, 262-274.

Bala, N., Carrington, P., & Roberts, J. (2009). Evaluating the Youth Criminal Justice Act after five years: A qualified success. Canadian Journal of Criminology and Criminal Justice51, 131-167.

Bell, S. (2014). Young offenders and juvenile justice: A century after the fact. Nelson.

Borum, R., Bartel, P., & Forth, A. (2002). Manual for the Structured Assessment for Violence Risk in Youth (SAVRY). Consultation version. Tampa: Florida Mental Health Institute, University of South Florida.

Bouchard, M. (2020). Collaboration and boundaries in organized crime: A network perspective. Crime and Justice49, 425-469.

Corrado, R., & Markwart, A. (1994). The need to reform the YOA in response to violent young offenders: Confusion, reality or myth. Canadian Journal of Criminology36, 343-378.

Corrado, R., & McCuish, E. (2015). The development of early onset, chronic, and versatile offending: The role of fetal alcohol spectrum disorder and mediating factors. International Journal of Child and Adolescent Health8, 241.

Corrado, R., Odgers, C., & Cohen, I. (2000). The incarceration of female young offenders: Protection for whom? Canadian Journal of Criminology42, 189-207.

Corrado, R., Vincent, G., Hart, S., & Cohen, I. (2004). Predictive validity of the Psychopathy Checklist: Youth Version for general and violent recidivism. Behavioral Sciences & the Law22, 5-22.

Cullen, F. (2011). Beyond adolescence‐limited criminology: Choosing our future—the American Society of Criminology 2010 Sutherland address. Criminology49, 287-330.

Cullen, F. T., Pratt, T. C., & Graham, A. (2019). Why longitudinal research is hurting criminology. The Criminologist44(2), 1-7.

Dawson, S., McCuish, E., Hart, S., & Corrado, R. (2012). Critical issues in the assessment of adolescent psychopathy: An illustration using two case studies. International Journal of Forensic Mental Health11, 63-79.

DiIulio, J. (1995, December). Moral poverty: The coming of the Superpredators should scare us into wanting to get to the root causes of crime a lot faster. Chicago Tribune, p. 31.

Drefahl, S., Wallace, M., Mussino, E., Aradhya, S., Kolk, M., Brandén, M., ... & Andersson, G. (2020). A population-based cohort study of socio-demographic risk factors for COVID-19 deaths in Sweden. Nature Communications11(1), 1-7.

Farrington, D. (2003). Developmental and life‐course criminology: Key theoretical and empirical issues‐the 2002 Sutherland Award address. Criminology41, 221-225.

Farrington, D.P. (2015). Prospective longitudinal research on the development of offending. Australian & New Zealand Journal of Criminology48, 314-335.

Feld, B. (1993). Criminalizing the American juvenile court. Crime and Justice17, 197-280.

Forth, A., Kosson, D., & Hare, R. (2003). The Hare Psychopathy Checklist: Youth Version. Multi-Health Systems.

Gottfredson, M. & Hirschi, T. (1990). A general theory of crime. Stanford University Press.

Gress, C. (2010). Revealing research & evaluation. https://www2.gov.bc.ca/assets/gov/law-crime-and-justice/criminal-justice/corrections/research-evaluation/issue-3.pdf.

Gushue, K., McCuish, E. C., & Corrado, R. R. (2021). Developmental offending patterns: Female offending beyond the reference category. Criminal Justice and Behavior4, 139-156.

Hirschi, T. (1979). Separate but unequal is better. Journal of Research in Crime and Delinquency, 16, 34-38.

Latessa, E., Lugo, M., Pompoco, A., Sulliva, C., & Wooldredge, J. (2015) Evaluation of Ohio’s prison programs. Center for Crime Justice Research, University of Cincinnati, Cincinnati.

Laub, J., & Sampson, R. (1993). Turning points in the life course: Why change matters to the study of crime. Criminology31, 301-325.

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

Le Blanc, M. et al. (1996). Mesures de l’adaptation sociale et personnelle pour les adolescents québécois: manuel et guide d’utilisation [Social and personal adjustment measures for Québec adolescents: User guide and manual] (3rd ed.). Université de Montréal.

Le Blanc, M., & Fréchette, M. (1989). Male criminal activity from childhood through youth: Multilevel and developmental perspective. Springer-Verlag.

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

Loughran, T. A., Mulvey, E. P., Schubert, C. A., Fagan, J., Piquero, A. R., & Losoya, S. H. (2009). Estimating a dose‐response relationship between length of stay and future recidivism in serious juvenile offenders. Criminology47(3), 699-740.

Lussier, P., Corrado, R. R., & McCuish, E. (2016). A criminal career study of the continuity and discontinuity of sex offending during the adolescence-adulthood transition: A prospective longitudinal study of incarcerated youth. Justice Quarterly33(7), 1123-1153.

Lussier, P., McCuish, E., & Corrado, R. (2020). Psychopathy and the prospective prediction of adult offending through age 29: Revisiting unfulfilled promises of developmental criminology. Journal of Criminal Justice, online first, 1-14.

Lussier, P., McCuish, E., Deslauriers-Varin, N., & Corrado, R. (2017). Crime specialization as a dynamic process? Criminal careers, crime mix, and crime specialization in chronic, serious, and violent offenders. In The Routledge International Handbook of Life-Course Criminology (pp. 112-139). Routledge.

Malakieh, J. (2017). Youth correctional statistics in Canada, 2015/2016. Ottawa, ON. Retrieved from https://www150.statcan.gc.ca/n1/pub/85-002-x/2017001/article/14702-eng.htm on October 19, 2020.

McCuish, E. C. (2020). Seductions of exposure time: The mismeasurement of desistance among persons involved in frequent and serious offending. Canadian Journal of Criminology and Criminal Justice62, 29-50.

McCuish, E., Bouchard, M., & Corrado, R. (2015a). The search for suitable homicide co-offenders among gang members. Journal of Contemporary Criminal Justice31, 319-336.

McCuish, E., Bouchard, M., Beauregard, E., & Corrado, R. (2019a). A network approach to understanding the structure of core symptoms of psychopathic personality disturbance in adolescent offenders. Journal of Abnormal Child Psychology47, 1467-1482.

McCuish, E., Cale, J., & Corrado, R. (2018a). A prospective study of offending patterns of youth homicide offenders into adulthood: An examination of offending trajectories and the crime mix posthomicide. Youth Violence and Juvenile Justice16, 18-36.

McCuish, E., Corrado, R., Hart, S., & DeLisi, M. (2015b). The role of symptoms of psychopathy in persistent violence over the criminal career into full adulthood. Journal of Criminal Justice43, 345-356.

McCuish, E., Corrado, R., Lussier, P., & Hart, S. (2014). Psychopathic traits and offending trajectories from early adolescence to adulthood. Journal of Criminal Justice42, 66-76.

McCuish, E., Hanniball, K., & Corrado, R. (2019b). The assessment of psychopathic personality disturbance among adolescent male offenders: Interview strategies and recommendations. International Journal of Forensic Mental Health18, 35-49.

McCuish, E., Lussier, P., & Corrado, R. (2015c). Examining antisocial behavioral antecedents of juvenile sexual offenders and juvenile non-sexual offenders. Sexual Abuse27, 414-438.

McCuish, E., Lussier, P., & Corrado, R. (2016). Criminal careers of juvenile sex and nonsex offenders: Evidence from a prospective longitudinal study. Youth Violence and Juvenile Justice14, 199-224.

McCuish, E., Lussier, P., & Corrado, R. (2018b). Incarceration as a turning point? The impact of custody experiences and identity change on community reentry. Journal of Developmental and Life-Course Criminology4, 427-448.

McCuish, E.C., Lussier, P., & Corrado, R.R. (2021). The life-course of serious and violent youth grown up: A 20-year longitudinal study. Routledge.

Millon, T., & Davis, R. (1993). The Millon Adolescent Personality Inventory and the Millon Adolescent Clinical Inventory. Journal of Counseling and Development, 71, 570-574.

Neil, R., & Sampson, R. (2021). The Birth lottery of history: Arrest over the life course of multiple cohorts coming of age, 1995–2018. American Journal of Sociology126, 1127-1178.

Ouimet, M. (1993). L’aigle et le castor: étude de la distribution spatiale de la criminalité aux États-Unis et au Canada. Criminologie, 26, 85-102.

Papachristos, A. V., Braga, A. A., Piza, E., & Grossman, L. S. (2015). The company you keep? The spillover effects of gang membership on individual gunshot victimization in a co‐offending network. Criminology53, 624-649.

Piquero, A. R. (2015). Understanding race/ethnicity differences in offending across the life course: Gaps and opportunities. Journal of Developmental and Life-Course Criminology1, 21-32

Reale, K., McCuish, E., & Corrado, R. (2020). The impact of juvenile sex offending on the adult criminal career. Sexual Abuse32, 400-422.

Schneider, A.L. (1990). Deterrence and juvenile crime. Springer.

Scott, E. & Steinberg, L. (2009). Rethinking juvenile justice. Harvard University Press.

Statistics Canada. (2013). British Columbia (Code 59) (table). National Household Survey Aboriginal Population Profile. Statistics Canada: Ottawa. www.statcan.gc.ca (accessed February 2020).

Statistics Canada. (2020). Average counts of young persons in provincial and territorial correctional services (Table: 35-10-0003-01 (formerly CANSIM 251-0008)). Retrieved from: https://www150.statcan.gc.ca/t1/tbl1/en/tv.action?pid=3510000301.

Sutherland, E. (1947). Criminology (4th ed.). Philadelphia: Lippincott.

Vincent, G., Odgers, C., McCormick, A., & Corrado, R. (2008). The PCL: YV and recidivism in male and female juveniles: A follow-up into young adulthood. International Journal of Law and Psychiatry31, 287-296.

Waldenlind, K., Eriksson, J. K., Grewin, B., & Askling, J. (2014). Validation of the rheumatoid arthritis diagnosis in the Swedish National Patient Register: a cohort study from Stockholm County. BMC Musculoskeletal Disorders15, 1-6.

Wall-Wieler, E., Roos, L. L., Bolton, J., Brownell, M., Nickel, N. C., & Chateau, D. (2017). Maternal health and social outcomes after having a child taken into care: population-based longitudinal cohort study using linkable administrative data. J Epidemiol Community Health71, 1145-1151.

Yang, J., McCuish, E., & Corrado, R. (2020). Is the foster care-crime relationship a consequence of exposure? Examining potential moderating factors. Youth Violence and Juvenile Justice19, 94-112.

Zylbersztejn, A., Gilbert, R., Hjern, A., Wijlaars, L., & Hardelid, P. (2018). Child mortality in England compared with Sweden: A birth cohort study. The Lancet391, 2008-2018.

Tables and Figures

Table 1. Non-exhaustive description of information included within CORNET administrative data

Type of data

Information included

Source of information

Longitudinal component

Official offending

Charges, convictions, admissions/releases from custody, court location, custody location, each time participant appears in court, type(s) of sanction, length of sanction(s). Specifies transfers to different institutions and the probation office responsible for supervision.

Court records provided by JUSTIN, a system used in the province of British Columbia for managing court appearances/case flow, etc.

Dates (YYYY/MM/DD) for all entries.

Risk assessment tools

The Community Risk-Needs Assessment (youth and adult versions), the Institutional Risk Assessment (youth and adult), the Structured Assessment of Violence Risk in Youth (Borum et al., 2002), various sexual offending and intimate partner violence risk assessment tools.

Probation officers, institutional social workers, correctional officers, institutional and community case managers. Practitioners interview participant and collaterals.

Completion date (YYYY/MM/DD) included for all risk assessment tools; does not necessarily correspond with date information acquired.

Tombstone data

Height, weight of participant. Home address, emergency contacts, list of persons approved to visit in custody.

Completed by institutional correctional officers at each admission to custody.

Dates (YYYY/MM/DD) for each updated entry.

Client logs (Community)

Pre-sentence reports detailing participant’s substance use history, mental health, family history, etc. Case management plans detailing participant’s strengths/weaknesses, goals for intervention, improvements. Identifies when a participant moved outside the province or died. Identifies who the person interacts with positively/negatively.

Case managers and probation officers interviewing participant and collateral contacts. Also includes the practitioner’s observations of participant behavior and attitudes.

The time (date, hour, minute) for each entry. Includes contact while incarcerated or under some form of community supervision.

Client logs (Prison)

Behavior in institution, including institutional charges for misbehavior, who the participant associates with, attitude towards staff and other residents. Identifies who the person interacts with positively/negatively. Injury forms (e.g., nature of injury, how it happened).

Correctional officers, institutional probation officers, youth social workers. Includes collateral contacts from lawyers, family members, intimate partners.

The time (date, hour, minute) for each entry. Completed daily when participant is incarcerated.

Institutional alerts

Names of persons participant is to have no contact with (e.g., participant assaulted another person; a person assaulted participant; participant co-offended with this person).

Correctional officers, intel from police, institutional and community probation officers, case managers. No-contact orders.

Dates (YYYY/MM/DD) for each unique alert. Reflects date alert added and not necessarily date of incident.

Table 2. Demographic characteristics, risk/need factors, and youth justice system involvement for the sample


Full Sample

(n =1,719)

M (SD) / n (%)

Male Participants

(n = 1,339)

M (SD) / n (%)

Female Participants

(n = 372)

M (SD) / n (%)

t/χ2, df, p, φ/d

Demographics





Age at Interview

16.06 (1.28)

16.16 (1.27)

15.72 (1.25)

t [1523] = 5.63, p < .001, d = 0.35

White

860 (54.8%)

681 (55.5%)

179 (52.5%)

χ2 [2] = 43.28, p < .001, φ = 0.17

Indigenous

464 (29.6%)

324 (26.4%)

140 (41.1%)

Non-Indigenous Minority

245 (15.6%)

223 (18.2%)

22 (6.5%)

School Behavior





Attending School

730 (47.9%)

581 (48.8%)

149 (44.6%)

χ2 [1] = 1.86, p = .173, φ = 0.04

Skipping School (Before Age 12)

321 (22.9%)

247 (22.6%)

74 (24.0%)

χ2 [1] = 0.26, p = .609, φ = 0.01

Number of Different Schools

5.61 (5.54)

5.55 (5.53)

5.82 (5.57)

t [1424] = -0.78, p = .438, d = 0.05

Substance Use





Substance Use Versatility

4.98 (2.20)

4.74 (2.12)

5.82 (2.30)

t [495.56] = -7.65, p < .001, d = 0.49

Hard Drug Use

722 (48.0%)

489 (41.7%)

233 (70.8%)

χ2 [1] = 87.59, p < .001, φ = 0.24

Family Dynamics





Left Home (Before Age 12)

223 (16.7%)

160 (15.3%)

63 (21.8%)

χ2 [1] = 6.92, p = .009, φ = 0.07

Kicked out of Home (Before Age 12)

101 (7.6%)

77 (7.4%)

24 (8.3%)

χ2 [1] = 0.26, p = .608, φ = 0.01

Family Dysfunction Scale

3.04 (1.71)

2.86 (1.66)

3.70 (1.77)

t [420.77] = -7.16, p < .001, d = 0.50

Sexuality and Abuse





Sexual Activity (Before Age 12)

132 (10.7%)

117 (12.0%)

15 (5.8%)

χ2 [1] = 8.28, p = .004, φ = 0.08

Physical Abuse

593 (46.3%)

420 (41.3%)

173 (65.8%)

χ2 [1] = 50.55, p < .001, φ = 0.20

Sexual Abuse

198 (17.5%)

87 (9.6%)

111 (49.3%)

χ2 [1] = 197.58, p < .001, φ = 0.42

Self-Identity





Positive Self-Identity

70.31 (10.16)

70.70 (10.06)

68.76 (10.44)

t [1345] = 2.81, p = .005, d = 0.19

Positive Other-Identity

64.54 (13.14)

64.55 (13.32)

64.50 (12.33)

t [1290] = 0.06, p = .955, d = 0.00

Youth Offending (Ages 12-17)





Age at First Conviction

14.89 (1.76)

14.95 (1.81)

14.63 (1.43)

t [560.05] = 3.29, p = .001, d = 0.20

Number of Convictions

11.17 (8.66)

11.43 (9.07)

10.20 (6.62)

t [609.41] = 2.66, p = .008, d = 0.16

Days Incarcerated

276.56 (280.06)

292.76 (293.30)

212.23 (206.50)

t [633.03] = 5.53, p < .001, d = 0.32

Any Violent Conviction

991 (61.2%)

792 (62.4%)

198 (60.7%)

χ2 [1] = 0.31, p = .578, φ = 0.01

Sex Offending

110 (7.2%)

104 (8.4%)

7 (2.3%)

χ2 [1] = 13.79, p < .001, φ = 0.10

Levene’s test of equal variance violated.

Notes. The sum of female participants (n = 361) and male participants (n = 1,318) does not equal the total sample (n =1,719) because of missing data. The full sample does not reflect the number of members that received an intake interview (n = 1,548) nor the number of members with available criminal history data in adolescence (n = 1,534). Versatility of substance use reflects the number of different types of drugs used. φ = phi effect size. d = Cohen’s d effect size.

[CHART]

Figure 1. Number of participants from Cohort I and Cohort II who received a particular interview

[CHART]

Figure 2. Number of Adult Community Risk-Need Assessments completed at each year of age for the ISVYOS sample.

Notes. Data collection remains ongoing and therefore sample sizes are underestimated, especially at later ages.

Comments
0
comment

No comments here

Why not start the discussion?