Unpacking Adultification: Institutional Experiences and Misconduct of Adult Court and Juvenile Court Youth Living Under the Same Roof

Abstract

The present study explores the ability of juvenile waiver policy and the resulting adult court status to impact the experiences and behavior of male youth originating from differing jurisdictions but incarcerated together within state-level juvenile correctional facilities. Using official agency data and youth survey data, this research examines how “adult” status influences behavior for waived youth compared with their juvenile court counterparts. Structural equation modeling and multiple group analysis is used to determine the extent to which adult status moderates the measurement structure of models related to individual characteristics, institutional experiences, and institutional misconduct. Results reveal an improvement in model fit to occur when estimated across the two groups of juvenile court and adult court youth, beyond that which occurs when parameters are constrained to be equal for these two groups. This work reveals adult status to be powerful enough to moderate the effects of who a youth is and what that youth experiences, on how that youth behaves during a term of juvenile incarceration.

Keywords

juvenile waiver adult court institutional misconduct rehabilitation multiple group analysis

Researchers have long focused on understanding the causes and correlates of institutional misconduct, and they have increasingly focused on exploring this type of behavior at the juvenile level. This study is one of the first efforts to quantitatively discern the extent to which adult court status influences a youth’s institutional experiences, including participation in institutional misconduct, during juvenile incarceration. Studies often provide evidence to demonstrate that misconduct is a robust predictor of recidivism (Huebner, Varano, & Bynum, 2007; Langan, Camp, & Saylor, 2004; Lattimore, MacDonald, Piquero, Linster, & Visher, 2004; Maruna & Toch, 2005; Trulson, DeLisi, & Marquart, 2009; Trulson, Marquart, Mullings, & Caeti, 2005). Other literature has explored institutional misconduct as the outcome of interest and has consistently found prior criminal behavior and gang membership to be robust predictors of such misconduct among adult and juvenile inmates alike (Cunningham & Sorensen, 2007; DeLisi, Berg, & Hochstetler, 2004; DeLisi, Spruill, Peters, Caudill, & Trulson, 2013; Griffin & Hepburn, 2006; Jiang & Fisher-Giorlando, 2002; MacDonald, 1999; Tasca, Griffin, & Rodriguez, 2010; Trulson, 2007).

Adult court status is often missing as a covariate in this research, possibly because these adult court youth are often exceptional within juvenile offender populations. Available data suggest that just less than 1% of all juvenile delinquency cases processed through the American juvenile justice system are waived to adult criminal court (Griffin, Addie, Adams, & Firestine, 2011). However, it is worth noting that most states do not publicly report their waivers, which means that the true proportion of juvenile cases waived is likely more than 1% (Griffin et al., 2011). These waivers effectively transfer jurisdiction from the more rehabilitative juvenile system to what has increasingly become a punishment-oriented adult system.

In addition, though waiver to adult court was originally reserved for the most violent juvenile offenders, the landscape has changed. The advent of more punitive legislation has continued to widen the net of eligible offenses and broadens the discretion by which to waive youth to adult court. Originally, judicial discretion provided the mechanism by which to waive youth, but increasingly prosecutorial discretion and automatic statutory exclusion waivers are utilized (Griffin et al., 2011). As a result, samples emerge such as those under study here, which include upward of 40% adult court youth incarcerated alongside juvenile court youth (JCY).

The role of adult status on behavior (usually defined as recidivism) is an area of research that has grown considerably during the last few decades, but this study aims to fill the void in research that examines behavior defined as institutional misconduct for these waived youth. This line of research is necessary not only because the proportion of adult court youth at state level correctional facilities is increasing but also because the increasing severity of youth at the state level brings with it a need to evaluate the safety of youth and staff in these facilities.

Theoretical Framework

Prison life has been described in ways that address adaptation to carceral life, institutional culture, pains of imprisonment, and evolution of identity behind prison walls that often involves the loss of autonomy (Clemmer, 1940; Goffman, 1961; Sykes, 1958; Sykes & Messinger, 1960; Wheeler, 1961). Different theoretical efforts have been made to explain processes of institutionalization and prisonization, and a discussion of this earlier work is warranted to provide a lens through which the current study examines a juvenile carceral environment.

Clemmer (1940) and Goffman (1961) agree that all inmates are subjected to similar conditions during imprisonment and ultimately react to these conditions by assimilating into the new role of inmate. As they become conditioned into their new role within this environment—a role more often believed to be passive than active—inmates typically surrender to the authority of the institution. Related to the pains of imprisonment, the deprivation model of behavior proposes that conditions within the institution related to loss of liberty and institutional structure can encourage aggression and violence among inmates (Poole & Regoli, 1983; Sykes, 1958). Those subjected to mass incarceration are often thought to lose most autonomy or agency that would normally factor into how they respond to a given situation or circumstance, causing them to become more automated parts of a larger and systematic, carceral machine.

This transition into incarceration begins with initial stages of institutionalization, which includes inmates possessing various pre-prison characteristics that may significantly influence misconduct, supporting an importation model of behavior (DeLisi, Trulson, Marquart, Drury, & Kosloski, 2010; Gover, MacKenzie, & Armstrong, 2000; Irwin & Cressey, 1962). In addition to pre-prison characteristics such as prior violence and gang membership, a newly adopted “adult” status is arguably another imported characteristic that may influence behavior. The present study capitalizes on a new venue in which to study experiences and behavior among youth who are experiencing the “adultification” brought on by the resulting sanction of juvenile waiver to and conviction within adult criminal court.

Wheeler (1961) introduced the concept of anticipatory socialization, demonstrated by a U-shaped curve where inmates find societal values important at intake and at the end of incarceration (prior to release), but adopt inmate values more during the middle of their prison terms. The anticipatory socialization introduced by Wheeler (1961) has gained considerable empirical support, with studies finding pre-prison experiences (Irwin & Cressey, 1962; Schrag, 1961; Thomas, 1977), level of connection to the extra-prison world during confinement (Tittle & Tittle, 1964), and inmates’ post-prison expectations (Thomas & Foster, 1972; Wellford, 1967) to influence level of prisonization. Barak-Glantz (1983) and others eventually acknowledged the glaring omission of indicators related to discipline within this work, and shed light on the importance of including behavioral indicators such as institutional misconduct, in addition to attitudinal indicators of the prison experience, within such evaluations of the lived prison experience.

The current project uses this foundation of prisonization and anticipatory socialization as a departure point for using institutional behavior as the key outcome for how “adultification” is lived on the ground of juvenile prisons. In addition to the importance of inmates’ pre-prison characteristics and the depriving conditions of correctional facilities, anticipatory socialization provides a relevant framework for this study because adult court youth likely view their approaching “release” very differently than do JCY. Although JCY may anticipate being released to their communities, adult court youth face a release from a juvenile facility only to be transferred to an adult prison. To provide additional context for the present study, the following review of literature provides an overview of the adult and juvenile misconduct research.

Review of Misconduct Literature

The majority of empirical research on institutional misconduct has focused on the adult population. Prior to discussing the smaller body of research related to misconduct among youth, it is necessary to first highlight the key findings of prior research in the adult literature. First, this work consistently finds younger age to predict involvement in violent misconduct (Cunningham & Sorensen, 2007; Flanagan, 1983; Griffin & Hepburn, 2006). For example, using a sample of adult male inmates, Flanagan (1983) found that inmates who were younger than 25 years of age at intake accrued a significantly higher rate of infractions during incarceration than did older inmates. Cunningham and Sorensen (2007) examined disciplinary records of more than 24,000 adult inmates in Florida to find that younger age predicted violence misconduct.

Findings are inconsistent with regard to a relationship between race and institutional misconduct. Some work finds racial minorities to engage in a higher level of misconduct (Harer & Steffensmeier, 1996; Huebner, 2003), whereas other research fails to find race as a predictor of misconduct (Jiang & Fisher-Giorlando, 2002). Most often, studies determine that African Americans are the most likely to engage in misconduct (Harer & Steffensmeier, 1996; Huebner, 2003). Some of this work is limited, however, in that it only compares African American and White inmates. Other research that includes Latinos in its analysis has found Latinos to be two to four times more likely to be involved in violent misconduct compared with African American or White inmates (Berg & DeLisi, 2006).

Prior research consistently finds criminal history and gang membership to be robust predictors of misconduct among adult inmates (Cunningham & Sorensen, 2007; DeLisi et al., 2004; Griffin & Hepburn, 2006; Jiang & Fisher-Giorlando, 2002; MacDonald, 1999; Tasca et al., 2010). More specifically, a prior violent criminal history increases one’s propensity to commit institutional violence (Kolivoski & Shook, 2016; MacDonald, 1999).

In addition, research on length of confinement and sentence length has produced mixed findings. Berg and DeLisi (2006) found a longer term of incarceration to predict a higher likelihood for involvement in violent misconduct. However, Fernandez and Neiman (1998) found that sentence length failed to predict rule infractions and misconduct among a large sample of 13,000 California inmates. While perhaps counterintuitive, Cunningham and Sorensen (2007) found a longer sentence length to actually predict lower rates of violent misconduct.

Other research focuses on depriving conditions of confinement, some of which can be mitigated by providing inmates with institutional support in the form of rehabilitative opportunities and social support. Using a sample of 4,168 inmates across 185 state prisons, Huebner (2003) found involvement in prison work programs to significantly decrease the likelihood of inmate violence. Cochran (2012) found that the frequency and most notably the consistency of visitation are associated with a decreased level of institutional misconduct. Evidence consistently reveals that the inclusion of both importation and deprivation variables explains a greater amount of variance in consequences of confinement than either model taken on its own (Hochstetler & DeLisi, 2005; Huebner, 2003; Thomas, 1977).

There is arguably a dearth of misconduct literature on juveniles within state-level correctional facilities relative to the prison misconduct research on adults. Poole and Regoli (1983) conducted one such study to explore violence of male youth within four different juvenile correctional facilities and found preinstitutional violence to most strongly predict aggression during terms of institutionalization. Echoing the adult literature, prior work also finds gang affiliation to also influence institutional misconduct within a juvenile correctional environment (DeLisi et al., 2013; Griffin & Hepburn, 2006; MacDonald, 1999; Trulson, 2007).

Trulson, Haerle, Caudill, and DeLisi (2016) examined a sample of more than 3,000 serious and violent juvenile offenders determinately sentenced to state-level juvenile facilities in Texas over the course of 25 years. They found younger age at commitment, African American race, and a higher number of prior out-of-home placements to predict a higher frequency of misconduct. They also found an increase in the number of years of education completed to predict less misconduct, revealing education to be rehabilitative and protective against misconduct involvement. Consistent with Cunningham and Sorenson’s (2007) research on adult inmates, Trulson and colleagues found lengthier determinate sentences to predict a lower frequency of misconduct.

Related, research has shown the pains of imprisonment to be exacerbated by a higher level of security within juvenile institutions (Feld, 1981; Gover et al., 2000; MacDonald, 1999; McShane & Williams, 1989). Feld (1981) coded facilities as having primarily a rehabilitative or custodial orientation and found that youth were more likely to engage in misconduct within the custody-oriented facilities. MacDonald (1999) found that being committed to institutions of higher security significantly increased the odds for youth to engage in violent behavior while institutionalized. Gover et al. (2000) found that commitment to a more strictly structured boot camp facility (rather than a more traditional juvenile facility) reported significantly higher levels of anxiety. McShane and Williams (1989) found juvenile offenders who entered Texas prisons as adults (ages 17-21 years) to be less likely than a similar young adult comparison group to acquire work assignments, or achieve perceived trust with institutional staff. This work also revealed these offenders to be less eligible for visitation, revealed by prior work to be protective against misconduct involvement (Cochran, 2012).

The research that focuses specifically on youth waived to and convicted within adult court is even scarcer. Tasca et al. (2010) provide one of few studies that examine predictors of violence among youth who have been waived to and convicted within adult court, and subsequently transferred to adult prison. Their examination of in-depth interviews with 95 male youths, aged 16 to 18 years, improved upon prior research by being able to code substantive self-reported deprivation variables such as having property victimized or being threatened with a weapon by fellow inmates. Findings reveal one importation variable (gang membership) and one deprivation variable (being threatened with a weapon) as significantly predictive of a self-reported, self-initiated assault on another inmate.

Kolivoski and Shook (2016) examined a sample of 763 waived youth committed to adult prisons in Michigan to determine the extent to which a youth’s developmental stage (measured as age at commitment) predicts involvement in institutional misconduct. They found younger age at commitment to significantly predict a higher total number of misconduct infractions during a youth’s term of commitment. Related, a study by Kuanliang, Sorensen, and Cunningham (2008) found younger age to significantly predict a higher likelihood for involvement in prison violence, when comparing a sample of juveniles below 18 years to adult prisoners, and specifically to a subsample of youthful adults aged 18 to 20 years.

Gover et al. (2000) examined carceral adjustment of juvenile offenders by exploring the relationship of certain importation and deprivation variables to the outcome of anxiety levels. With this research, they produced one of few studies that analyze primary self-report survey data collected from juvenile correctional facilities. Among the 3,986 youths committed to 48 facilities nationwide, this cross-sectional research revealed that White youth, those exposed to family violence, younger youth, and those who perceived a higher level of procedural justice experience significantly lower levels of anxiety.

The outcome of institutional misconduct within juvenile correctional facilities has been neglected within the body of research on juvenile waiver policy, and it is not known whether the same factors influence misconduct for youth from juvenile court compared with adult court youth. Only one study could be identified that compared misconduct of adult court youth to JCY while committed to the same facility. Bechtold and Cauffman (2013) found that adult court youth engage in less misconduct overall than their juvenile court counterparts. This single study provides a sound research design but is limited because it only examines a sample of youth from one facility and evaluates behavior exclusively during the first two months of incarceration.

The current study aims to fill a gap in the extant research by examining waived youth who are still incarcerated alongside their juvenile court counterparts. This work expands on the work of others (Bechtold & Cauffman, 2013; McShane & Williams, 1989) by examining the extent to which juvenile offenders and “adult” juvenile offenders behave differently under the same roof. The ways in which their “adult” status influences their behavior could detrimentally affect the institutional culture and rehabilitative focus of juvenile correctional facilities.

“Adult” Status, Experiences, and Behavior

Waiver to adult court originated from arguments such as deterrence and incapacitation theory, but efficacy of this policy has been a subject of debate since its inception. Contrary to arguments that aim to justify waiver, numerous negative consequences of waiver have emerged in the research as scholars conduct different empirical evaluations of this policy (Bishop, 2000; Howell, 1996). There is surprisingly little known about the various impacts of juvenile waiver to adult court (Krisberg, 2005). The empirical research that addresses this issue tends to focus on sentencing and recidivism outcomes of juveniles waived to adult court, compared with either juveniles who were not waived or to similar young adult counterparts.

To begin with, punishment severity across court type is clearly disproportionate for waived youth compared with others. Research suggests that waived youth are more often detained pre-trial, convicted, and incarcerated than youth who remain in juvenile court (Kupchik, 2006). Waived youth also receive more severe punishments than their juvenile court counterparts, in the form of not only lengthier sentences but also often an inevitable transfer to adult prison (Barnes & Franz, 1989; Bishop, 2000; Fagan & Kupchik, 2011; Kupchik, 2006).

Sentences delivered in adult court often place youth at a greater risk of physical and sexual victimization within adult prisons, compared with youth who are incarcerated within juvenile correctional facilities (Bishop, 2000; Kupchik, 2006). While providing due process to juveniles within the adult venue of criminal court, waiver often subjects juveniles to adult incarceration that arguably violates other constitutional rights, such as the prohibition of cruel and unusual punishment. Forst, Fagan, and Vivona (1989, p. 9) eloquently support this position by suggesting that “a rather cruel and ironic form of punishment is accorded to waived youth, where retribution for crimes against society occurs through victimization by staff and inmates” (in adult prisons).

In addition to potential victimization, waived juveniles are also less likely to receive counseling and therapeutic services during their terms of incarceration (Penney & Moretti, 2005). This suggests the potential marginalization of waived youth during juvenile incarceration, a point on which the current project will elaborate. Kupchik (2007) found that, though counterintuitive, waived youth in adult facilities reported better access to treatment and education services than waived youth committed to juvenile facilities. His finding supports the notion that youth housed within facilities that promote a rehabilitative intent should have equal access to such services. His study does not, however, compare the availability of services in juvenile facilities for waived youth with those retained in juvenile court. The current project explores this line of inquiry by comparing youth who are under different jurisdictions but housed within the same juvenile facilities. It becomes even more important to equalize the playing field at the level of juvenile corrections if some youth are headed to adult prisons, to be inevitably exposed to environmental destabilizers shown to promote persistence of a lifelong criminal career (Penney & Moretti, 2005).

Yet another unintended consequence is the increased recidivism following adult incarceration for adult court commitments compared with similar offenders sentenced and incarcerated under juvenile court jurisdiction (Bishop, 2000; Kupchik, 2006). These recidivism findings clearly indicate that the policy of waiver misses its mark in at least two respects: specific deterrence and enhanced public safety. Clearly, during their term of incarceration, the general public is safe from these serious and violent juvenile offenders, by virtue of these youth being incapacitated and removed from society entirely during this period of time. This may lend support to incapacitation theory as justification for waiver, but this benefit of incapacitation is overshadowed if the juveniles who are eventually released from adult prison—potentially following extensive criminalization—recidivate at higher rates than comparable offenders released from juvenile institutionalization, offsetting any benefits of incapacitation (Bishop, 2000; Kupchik, 2006).

Given that institutional misconduct is an established, robust predictor of recidivism, any policy that affects rates of institutional misconduct (for better or for worse) should be examined for efficacy or lack thereof. It is important to further explore misconduct as an outcome not only because it predicts recidivism but also because levels of misconduct are a safety concern for youth and staff within juvenile correctional facilities. More broadly, such research can inform the evolution of institutional theory, especially as it relates to prisonization of young offenders. This study will contribute to this body of knowledge by exploring differences in institutional misconduct for adult court youth as compared with JCY across five different state juvenile correctional facilities in California, and during a period of more than 2 years. The focal population of this study is juvenile offenders who are waived to adult court, convicted within adult court, and thus “adultified” prior to beginning a term of juvenile incarceration.

Data Sources and Methodology

The Research Site and Context

The data for the current study were gathered as part of a larger research study on gangs and violence in California juvenile correctional facilities. Access to these facilities was granted as a result of a consent decree in the Farrell v. Allen lawsuit of 2004. This consent decree mandated that the California Division of Juvenile Justice (DJJ) recruit of a nationally recognized expert to develop strategies to reduce gang and racial violence within DJJ facilities.

At the time of the study, the population within DJJ had already changed dramatically as a result of lower crime rates, policy changes, facility closures, and the gradual realignment of juvenile custody from the state to the county level. Between 1999 and 2009, the population of DJJ decreased by 79%, from 7,761 to 1,602 (Office of Research, 2010). In addition to the size of the population changing dramatically, the nature of this population also changed during this time because of legislative changes that widened the net for incarcerating specifically serious and violent offenders.

Originally, juvenile offenders deemed to be inappropriate for county-level services (due to extensive delinquent histories, lack of amenability to treatment, etc.) were committed to DJJ facilities. Both youth processed through juvenile court and those waived to adult court could previously be committed to DJJ until the age of 25 years, if need be. However, policy now requires that DJJ transfer all adult court youth to an adult California Department of Corrections and Rehabilitation (CDCR) facility at the age of 18 years.

In 2007, the California legislature mandated that only violent or serious felonies (defined under PC 707b¹) that were eligible for waiver to adult court, and certain sex offenses that required registration as a sex offender (defined under PC 290²) could be committed to DJJ (Office of Research, 2010). This significantly narrowed the scope of offenses eligible for state commitment and ultimately produced a population of juvenile court and adult court youth that was, for all intents and purposes, extremely similar.

Data Collection and Sample

Data for this project derive from two main data sources. The first is official agency data provided by the DJJ, which includes various datasets with information on youth characteristics (demographics, facility classifications such adult court status, gang member designation, etc.), delinquent histories, and institutional misconduct (including violent incidents). These cross-sectional data were acquired on the full DJJ population as of September 15, 2010 (N = 1,153 male youths³) to approximately align with the youth interview dates (which occurred during September and October 2010). The official data include cross-sectional data on some characteristics and classifications, as well as historical data on movement through facilities and institutional misconduct. A second set of official data was acquired that covered the time period of September 16, 2010, through February 23, 2012. It is this post-interview official dataset that provides institutional misconduct data for the current project.

The second main data source consists of primary data in the form of self-report youth surveys. This survey was administered in all five youth facilities (three in northern California and two in the southern part of the state) to a random sample of the 1,153 male youths, and resulted in a survey sample of 306 male youths.⁴ We generated a random order list of all youth housed at a given facility, and then designated the first 80 youths on the list as our initial sample, with the goal of completing at least 50 to 60 interviews at each facility. Our research team was granted limited access to each facility, with the expectation of interviewing approximately 25% of the total population of male youth committed to these five facilities at the time of data collection. We made face-to-face contact with each youth individually to provide an explanation of the study, secure consent, or to confirm a refusal. Any youth who were off-site (at court, hospitalized, transferred to another facility) were noted and, when possible, our team circled back to attempt contacts a second time.

We conducted these youth interviews individually and in private settings such as empty offices or classrooms within the facilities to ensure confidentiality. Youth were able to skip questions as needed, could end the interview at any time, and the average length of these interviews was 70 min (Maxson et al., 2011). These interviews covered general topics related to demographics, family dynamics, perceptions of safety in the facility, and concerns about returning home or transferring to adult prison once released from juvenile incarceration. Youth interviews also included items related to social support and rehabilitation, which are of interest for the present study. We ultimately experienced a participation rate of 84%, and descriptive statistical analyses confirmed that our random sample was more than adequately representative of the full DJJ population.⁵ Related to the current project, the full population consisted of 189 adult court youth (16%), and the interview sample consisted of 60 (20%).

What Is an “EnM?”

During initial site visits prior to data collection, we met with key administrators and staff members to discuss logistics, to familiarize ourselves with potential interview locations at each facility, and to develop initial rapport with our staff member points of contact. Consistent across all facilities, and revealed in our field notes of these visits, various staff members referred to “E and M’s” in passing (pronounced EnMs). Staff typically attributed a negative connotation to this status, implying that this group of youth required a higher level of structure due to the fact that they were more severe offenders and displayed worse behavior overall than other youth. Upon asking different individuals to explain what they meant by “EnM,” staff provided similar explanations that often consisted of one phrase: the wards from adult court.

The youth within the general population of JCY receive an identification number upon intake that consists of five digits, much like a zip code. “EnMs,” however, receive an identification number that starts with either an “E” or an “M,” followed by four numbers.⁶ It is worth noting that during this study’s data collection, EnMs were all awaiting transfer to adult prison at 18 years, whereas JCY were allowed to remain in DJJ facilities (or under juvenile jurisdiction on parole) until the age of 25 years.

This single difference in their respective fates effectively highlights the fact that these groups are subjected to two very different models of accountability-based justice (Ward & Kupchik, 2009). JCY are afforded the opportunity to enjoy system-based accountability, as provided by an additional 7 years of protection by the juvenile system under DJJ jurisdiction, which aligns more closely with the philosophy of parens patriae that ensured a state or county’s system would take responsibility and care for its youth. EnM youth, on the other hand, are subjected to a newer and more punitive model of person-based accountability—a model that holds them to a higher level of personal responsibility, as evidenced by eviction from the juvenile justice system and an inevitable term of incarceration within an adult prison (that begins a full 7 years earlier than the end of juvenile jurisdiction).

Research Question and Hypotheses

This project utilizes data that can examine a combination of the actual behavior and experiences that occur following the application of juvenile waiver policy in California. This larger framework provides the following research question:

Research Question 1: Does EnM status moderate differences in institutional experiences and misconduct of incarcerated youth, and which factors predict any differences between these two groups?

This question is answered first using secondary official data for the full population of youth below the age of 18 years, and then by supplementing those analyses with self-report interview data that will allow for examination of how additional constructs related to institutional experiences influence misconduct for our representative interview sample. This determines the ways in which individual characteristics and experiences during incarceration predict variation in institutional misconduct by court of commitment, and which of these two measurement structures is more strongly moderated by EnM status. Differences for these groups are expected because prior research has shown that adult court youth engage in less institutional misconduct than JCY, and that availability to services also varies by court of commitment (Bechtold & Cauffman, 2013; Kupchik, 2007). Evidence from the extant research drives my theory that “adultification” initiates a different process of adaptation to juvenile correctional environments for EnMs than for JCY.

The primary hypothesis derived from this research question is that adult court status (EnM status) will moderate the influence of individual characteristics and institutional experiences on future misconduct. In addition, differential effects of these individual characteristics and institutional experiences on future misconduct will exist across the two groups of adult court (EnMs) and JCY. Two broad competing hypotheses emerge from this aforementioned research question and rationale:

Hypothesis 1: Adultification → institutionalization → negative experiences → higher levels of misconduct

Hypothesis 2: Adultification → institutionalization → positive experiences → lower levels of misconduct

The above hypotheses make an assumption that some level of institutionalization is an inevitable result of incarceration, prior to other experiences that may influence behavior during that term of incarceration. Before narrating each hypothesis, a brief explanation of these key concepts is warranted. The concept of “adultification” is the most straightforward and is operationalized as a dichotomous indicator of adult court commitment (EnM) status.

The concept of institutionalization is included in the hypotheses to highlight the fact that every youth experiences an initial phase of institutionalization when they are committed to DJJ. Goffman’s (1961) work is relevant here, as he provided thoughts on processes that every youth (or any inmate) inevitably experiences during this initial phase. These processes include role dispossession, trimming, and contaminative exposure (Goffman, 1961). For the youth under study, contaminative exposure of the physical form is particularly threatening because they may be forced to interact with gang members, sex offenders, youth in need of mental health treatment, and other youth who are incarcerated for commission of a violent crime.

During this initial period of institutionalization within DJJ facilities, specifically, youth also gain knowledge (or have prior knowledge confirmed) about the fact that DJJ norms are very different from societal norms. Within the facilities under study here, misconduct (including violence) is very normalized, and avoiding such behavior is known among youth and staff to be the exception, rather than the rule.⁷ During this term of incarceration, they may either “conform” to these norms of misconduct and violence or deviate from these institutional norms by withdrawing from such disruptive behavior.

Hypothesis 1 suggests that this “adultification” may encourage EnMs to have more negative experiences and to then respond with conformity (i.e., embracing norms and expectations of misconduct and violence) as they await transfer to adult prison. Alternatively, this accelerated “maturity” may cause adult court youth to adapt in a way that encourages them to conceptualize themselves differently, and to utilize agency by deviating from institutional norms. Influenced by more positive experiences, this type of adaptation may influence lower rates of misconduct for this group overall, compared with similarly serious and violent JCY.

Analysis Plan

These competing hypotheses are tested here by exploring ways in which institutional misconduct differs by court of commitment, individual characteristics, and institutional experiences. “Experiences” in these hypotheses include any opportunity that a youth has for self-improvement (education, employment, treatment services) or any opportunity a youth has to develop and maintain social support in a way that should foster rehabilitation (Forst et al., 1989; Kupchik, 2007). The operationalization of these different experiences will be explained in more detail within the measures section to follow.

Following a descriptive examination of the sample, differences in characteristics, experiences, and behavior of these two groups are tested using the quantitative methodology of structural equation modeling (SEM). This is the ideal methodology because SEM has the ability to test multiple relationships between predictors (exogenous variables) and multiple outcomes (endogenous variables) simultaneously, while controlling for measurement error (Bollen, 1989).

Most importantly, SEM also allows for analysis of multiple groups to determine the differences between estimated parameters in a given model of one sample compared with another. Multiple groups analysis is often used if a status indicator such as gender, age, or race is believed to exert a more significant overall effect on an outcome than is reflected by including that indicator as only a covariate in the model. Otherwise put, the measurement structure underlying the models is hypothesized to differ by court of commitment (EnM status). This method will provide a straightforward manner by which to determine whether parameter estimates within each model are statistically significantly different for EnM youth compared with JCY. Rather than a variable simply mediating a relationship between another variable and the outcome, this variable is believed to moderate the parameters that are specified in the model (Kline, 2011).

Descriptive Analysis

Sample Characteristics

As previously discussed, the current population of DJJ includes only the most severe and violent juvenile offenders in the state of California. In fact, for commitment to DJJ, a youth must have committed a waiver-eligible offense. This means that, waived or not, every youth in DJJ is similarly severe and violent.

To produce the most valid comparison group possible, this sample excludes any youth more than the age of 18 years and includes only those within the developmental stage of mid-adolescence that are 13 to 17 years of age. Few EnMs are incarcerated within DJJ who are more than the age of 18 years, primarily because adult court youth are transferred to adult prison at that age. The official data are explored first for the full population of those below 18 years (N = 425, 174 EnM youth and 251 JCY), and emphasis is given to the following categories: demographics, criminal history, and institutional misconduct.

Table 1 below displays descriptive statistics for the official data variables used in the structural model that examines the influence of individual characteristics on institutional misconduct.

Table 1.

Descriptive Statistics for Full Population.

	EnMs (n = 174)		JCY (n = 251)		Test of difference	Total (N = 425)
	N	%	N	%	Test of difference	N	%
Race
Latino	110	63	153	61	NS	263	62
Black/African American	48	28	81	32	NS	129	30
Other (including White and Asian)	16	9	17	7	NS	33	8
Mean age in years (SE)	17.4 (0.04)		17.0 (0.05)		***	17.1 (0.04)
Gang member	104	60	183	73	**	287	68
Registered sex offender	80	46	118	47	NS	198	47
Mental health treatment need	41	24	97	39	**	138	33
Mean age (years) crime committed (SE)	15.6 (0.07)		15.2 (0.07)		***	15.4 (0.05)
Sentence length in years	11.3 (0.88)		13.7 (0.90)		^†	12.7 (0.64)
Total number of sentencing enhancements	0.85 (0.06)		0.67 (0.05)		*	0.75 (0.04)
Sentencing enhancements
Felony gang	46	26	69	28	NS	115	27
Firearm	67	39	54	22	***	121	28
Great bodily injury	29	17	37	15	NS	66	16
Other	5	3	7	3	NS	12	3

Note. JCY = juvenile court youth; NS = not significant.

†

Marginally significant p < .10. *p < .05. **p < .01. ***p < .001.

Individual Characteristics Model—Measures

To build on the framework and the competing hypotheses discussed above, the first model proposes that those individual characteristics such as demographics, facility classifications, and offender severity influence institutional misconduct during juvenile incarceration. Multiple group analysis allows for determination of whether or not EnM status moderates the effect of each exogenous predictor on the endogenous outcome of institutional misconduct. Analysis is conducted using the full population (N = 425) for this first model.

Endogenous Variables—Institutional Misconduct

For the purposes of this research, the dependent variables that measure misconduct are constructed using official agency data from DJJ. Institutional misconduct rates were calculated using each facility’s Disciplinary Decision-Making System (hereafter, DDMS) database. DJJ provided DDMS data recorded prior to and following our interview date of September 16, 2010, providing the ability to use this behavior as the key outcome variable for this study.⁸ DDMS includes Level 2 and Level 3 allegations of misconduct (the higher the level, the more severe the incident).⁹ Level 2 incidents are typically nonviolent or at worst involve a minor physical altercation. Level 3 incidents are more severe in that they jeopardize the safety of youth or staff within DJJ (whether through a more violent physical altercation or through smuggling dangerous contraband into the facility). The endogenous variables for this analysis are two categories of institutional misconduct involvement: Any DDMS and Violent DDMS.

The use of Any misconduct emphasizes the need to consider chronicity of behavior overall (combining nonviolent and violent), whereas Violent provides a more narrow outcome that captures a higher severity of misconduct. For the Violent category of misconduct, Level 3 allegations and any Level 2 allegations that involve even a minor physical altercation are combined. These outcome variables are aggregate rates of these categories of misconduct incidents officially reported by DJJ staff during the period of institutionalization following our interviews. These misconduct rates were calculated by summing the individual DDMS allegation codes¹⁰ for a given category for each youth that occurred following the interview date and then dividing that by the time each youth has been incarcerated within DJJ following that interview date. The time incarcerated takes into account any temporary release(s) and/or re-incarcerations that may have occurred during the 524-day follow-up period post-interview.

Exogenous Variables

Figure 1 illustrates the structural model that explores the ability of certain observed individual characteristics variables to influence differences in institutional misconduct, while controlling for other factors. This model examines group differences using variables from the official data related to individual characteristics. The analysis for Figure 1 will utilize the full population of those below 18 years of age (N = 425), meaning that all exogenous and endogenous variables for this model are derived from DJJ’s official agency data, allowing me to maximize the sample size of youth examined in this section of analysis.

Figure 1.

Predicting institutional misconduct with individual characteristics.

Key Variables

Demographics:

Age: The official age of each youth at the time of interview/data collection.

Race: The race of each youth is categorized into African American, Latino, and Other (includes White and Asian). For the analysis, African American is the reference group.

Facility classifications:

EnM Status: This is a dichotomous indicator of adult court of commitment that is used as the grouping variable in the following analysis.

Gang membership: Gang member status is a classification assigned to each youth upon intake to DJJ. This status is determined by a variety of factors that occur either alone or in combination (i.e., self-identified gang member, gang enhancement, previously recorded as a gang member within California’s gang database, etc.).

Severity of offender:

Age crime committed: The official age of each youth at the time of commitment to term of DJJ incarceration.

Sentence length: Length of sentence in years.

Number of sentencing enhancements: Total number of enhancements, including felony gang, firearm, great bodily injury, and any other miscellaneous category of enhancement.

Grievances filed: This is a logged count of all grievances filed during a youth’s current term of incarceration.

The observed variables of age at commission of crime, total number of sentencing enhancements, and sentence length are used to build the construct of offender severity in this model, with sentence length being predicted by the total number of sentencing enhancements.¹¹ Preliminary regression analyses revealed sentencing enhancements to be a robust predictor of sentence length but showed that these enhancements have no effect on behavior. The total number of enhancements is introduced into this model because it is theoretically related to offender severity and because it provides the degree of freedom necessary to assess model fit for this recursive structural equation model (Kline, 2011).

Individual Characteristics Model—Multiple Group Analysis Results

Multiple group analysis generates the results displayed in Table 2 below and allows for estimated parameters of the structural model to be compared across these two groups. Older age marginally predicts a lower rate of both misconduct categories for EnMs, but this effect was statistically significant for JCY (and more robust for JCY when predicting Violent misconduct). The age at commission of crime marginally predicts a higher rate of Any misconduct but only for EnMs. In addition, EnMs in the “Other” category of race (which includes White and Asian youth) have a significantly lower rate of Violent misconduct than African American EnMs. There were no significant findings related to race for JCY, or for the outcome of Any misconduct.

Table 2.

Standardized Regression Results of Multiple Group Analysis of Individual Characteristics Model.

	Any misconduct				Violent misconduct
	EnMs (n = 174)		JCY (n = 251)		EnMs (n = 174)		JCY (n = 251)
	Coefficient	SE	Coefficient	SE	Coefficient	SE	Coefficient	SE
Age	−0.064^†	0.036	−0.037*	0.017	−0.062^†	0.032	−0.063***	0.015
Age crime committed	0.042^†	0.022	0.005	0.012	0.019	0.020	0.014	0.010
Race (African American reference)
Latino	0.056	0.034	−0.033	0.025	0.017	0.031	0.002	0.022
Other (including White & Asian)	−0.081	0.058	−0.077	0.051	−0.116*	0.051	−0.065	0.044
Gang member	0.023	0.034	−0.007	0.027	−0.041	0.030	0.007	0.023
Sentence length	0.003*	0.001	0.000	0.001	0.001	0.001	−0.001	0.001
Grievances filed (log)	−0.069	0.037	−0.036	0.028	0.034	0.033	−0.042^†	0.025
Enhancements → Sentence length	7.567***	0.976	6.469***	1.119	7.567***	0.976	6.469***	1.119
R²—misconduct	R² = .092		R² = .037		R² = .080		R² = .094
R²—sentence length	R² = .257		R² = .118		R² = .257		R² = .118

Note. JCY = juvenile court youth.

†

Marginally significant p < .10. *p < .05. **p < .01. ***p < .001.

A longer sentence length significantly predicts an increase in the rate of Any misconduct for EnMs but fails to be significant elsewhere. This analysis reveals total number of sentencing enhancements to be the most robust predictor to explain variance in sentence length for both EnMs and JCY. Table 1 revealed that these groups do not differ descriptively in the proportion subjected to felony gang, great bodily injury, or “other” types of enhancements. However, a significantly higher proportion of EnM youth have a firearm enhancement. Given that these two groups are similarly serious and violent offenders who do not differ in any commitment offense category (i.e., homicide, assault, robbery, or burglary categories), the difference in proportion of firearm enhancements for these two groups is likely driven by prosecutorial discretion rather than offense type. This is further validated by the fact that total number of enhancements significantly predicts a longer sentence length for EnMs and JCY alike.

Model Fit for Individual Characteristics Model

For the full population structural model that predicts misconduct categories using individual characteristics, approximate fit indices improve substantially when estimating a multiple groups model, as displayed in Table 3 below. This improvement in overall model fit is consistent, regardless of the misconduct outcome (Any or Violent).

Table 3.

Model Approximate Fit Statistics: Individual Characteristics Models.

	df	χ² (p value)	CFI	RMSEA	RMSEA CI
Any misconduct
Full model	8	26.193 (.001)	0.829	0.073	[0.043, 0.105]
Multiple groups	14	23.602 (.051)	0.906	0.057	[0.000, 0.095]
Violent misconduct
Full model	8	26.342 (.001)	0.831	0.073	[0.043, 0.106]
Multiple groups	14	25.059 (.034)	0.906	0.061	[0.017, 0.099]

Note. CFI = comparative fit index; RMSEA = root mean square error of approximation; CI = confidence interval.

For both of these outcomes, the models reject the null exact-fit hypothesis that the model-implied covariance matrix is equal to the sample covariance matrix, as indicated by the significant chi-square values. This indicates that any differences between the covariance matrix implied by the model and the sampling covariance matrix are not due to factors other than sampling error—discrepancies between the two may be more than that expected by chance. However, with EnM status as the grouping variable, the chi-square value is no longer significant for the model predicting Any misconduct, and also improves for the model predicting Violent misconduct (although still significant at the .05 level). Although this insignificant model test statistic provides preliminary evidence in favor of the model being consistent with the covariance data, other aspects of the correspondence between the model and the data need to be considered.

Approximate fit indexes are continuous measures of the correspondence between a model and the data that provide qualitatively descriptive information about model fit (Kline, 2011). Most of these indices are scaled as goodness-of-fit statistics (the higher the values, the better the model fit). The comparative fit index (CFI) provides an incremental fit index to determine the improvement in the fit of the utilized model relative to that of a baseline model. The root mean square error of approximation (RMSEA) provides a point estimate that is scaled as a badness-of-fit index, with a value closer to zero indicating better model fit.

As shown, significant improvements in fit occur, as indicated by CFI values that reveal an 8% improvement over that of the independence fit model. In addition, RMSEA values also decrease when moving from the full structural model to the multiple group analysis. Given these fit statistics, it can be concluded that the same constructs in this model are manifested in different ways for these two groups. Otherwise put, EnM status is powerful enough to moderate the overall measurement structure using individual characteristics to predict misconduct for these two groups.

Institutional Experiences Model—Measures

This second section of analysis uses the youth interview data to determine if institutional experiences reveal additional differences for interviewed EnM youth (n = 56) compared with their juvenile court counterparts (n = 64), and if those previously established parameter differences in rates of misconduct hold for the interview sample of those below 18 years of age (N = 120).

Endogenous Variables

The outcome variables for this analysis are the same two aggregate rates of misconduct incidents reported by DJJ staff during the period of institutionalization following our interviews: Any DDMS and Violent DDMS. Similar to the previous model, these outcome variables are constructed from official agency data.

Exogenous Variables

Figure 2 below provides the structural model for this analysis and introduces new constructs provided by the youth interview data. Although this limits the sample for analysis here to the youth interview sample (N = 120), it allows for modification of the model to estimate parameters related to institutional experiences for these two groups.

Figure 2.

Predicting institutional misconduct with institutional experiences.

In addition to including the grouping variable of EnM status, various youth interview items are recoded into a composite index for each of the following constructs¹²: social connectedness, institutional support, self-esteem, stock in rehabilitation, and treatment services received.

As previously mentioned, institutional violence is very normalized within DJJ facilities. Any lack of rehabilitation (or negative experience) that a youth endures could encourage him to conduct himself in a way that conforms to these violent institutional norms. Alternatively, more positive institutional experiences such as greater exposure to rehabilitation could encourage him to deviate from these norms by engaging in less violence. In addition to potential parameter differences among these two groups, the model fit for this structural model that incorporates experiences is hypothesized improve when estimated across two groups than when parameters are constrained to be equal for these two groups. Identical to Figure 1, EnM status is used as the grouping variable.

Responses from the youth interview are utilized in this second model to operationalize the concept of “institutional experiences” as introduced within the previously outlined hypotheses and theoretical framework. These constructs are developed using items from the youth interviews that are theoretically related to each topic, and these indicators are loaded onto an index to measure each of these variables.¹³ The value of each construct is then used as a covariate in models to predict Any misconduct and Violent misconduct. Items from the youth interview used in the composite for each construct are listed below in Table 4.

Table 4.

Variable Descriptions & Indicators for Composite Self-Report Constructs.

Social connectedness: The extent to which a youth maintains connections to family and friends outside of DJJ.

Have any family members or adults you’ve lived with ever been incarcerated? (reverse coded)

Do you have any children?

Do family members or friends visit you in a typical month?

Do you receive letters during a typical month?

Do you have phone conversations with family or friends?

Institutional support: The level of support from and strength of rapport with DJJ facility staff that each youth expresses.

Are there people in this facility you talk to about personal matters?

If you were in trouble, is there someone here you could talk to?

Dichotomous indicator of paid employment within current facility

Staff here trust me with extra responsibilities.

Staff here allow me more freedoms than other youth in the facility.

Staff here believe I usually behave responsibly.

Treatment services received: Measurement of the total number of services that a youth either has received or is currently receiving.

Do you receive all of the treatment services that you’re supposed to receive according to your Individual Change Plan (ICP)?

Do you receive individual counseling on a regularly scheduled basis?

Do you receive group counseling on a regularly scheduled basis?

Total number of specialized treatment programs received.

Total number of resource groups received.

Incentive level (A = 3, B = 2, C = 1)

Stock in rehabilitation: The extent to which a youth personally believes treatment services to be necessary and effective.

In general, do you have a positive or negative feeling about the treatment services that you receive here?

Do you need treatment for any specific problem?

Are there treatment services not currently offered here that might help youth?

Are there services in this facility that reduce violence between youth?

Self-esteem: Measurement of self-worth using a subset of items from the Rosenberg (1965) self-esteem scale.

I feel that I have a number of good qualities.

I am able to do things as well as most other people.

I take a positive attitude toward myself.

Institutional Experiences Model—Multiple Group Analysis Results

The constructs related to incarceration experiences produce noteworthy differences in parameter estimates for EnMs and JCY, and these results are reported below in Table 5. Social connectedness is significantly predictive of decreased misconduct (both categories of Any and Violent) for JCY but has no effect for EnMs. Institutional support significantly predicts a lower rate of Violent misconduct for JCY, but this significant effect on Any misconduct is much more robust for EnMs.

Table 5.

Standardized Regression Results of Multiple Group Analysis of Institutional Experiences Model.

	Any misconduct				Violent misconduct
	EnMs (n = 56)		JCY (n = 64)		EnMs (n = 56)		JCY (n = 64)
	Coefficient	SE	Coefficient	SE	Coefficient	SE	Coefficient	SE
Institutional experiences
Social connectedness	−0.003	0.005	−0.016***	0.005	0.004	0.005	−0.011**	0.004
Institutional support	−0.026***	0.005	−0.008^†	0.005	−0.002	0.006	−0.009*	0.004
Stock in rehabilitation	0.003	0.006	−0.001	0.005	−0.011^†	0.006	−0.004	0.005
Treatment services received	−0.001	0.006	−0.004	0.004	−0.014*	0.006	−0.006	0.004
Self-esteem	0.007	0,018	0.021^†	0.013	−0.012	0.019	0.004	0.011
Treatment services received → Stock in rehabilitation	0.402**	0.116	0.152	0.104	0.402**	0.116	0.152	0.104
R²—misconduct	.331		.237		.115		.233
R²—stock in rehabilitation	.177		.032		.177		.032

Note. JCY = juvenile court youth.

†

Marginally significant p < .10. *p < .05. **p < .01. ***p < .001.

Stock in rehabilitation is marginally predictive of a lower rate of Violent misconduct for EnMs only, but does not approach significance elsewhere. This implies that a stronger belief in the rehabilitative ideal exerts a more substantial effect in decreasing violence for EnMs than for JCY. A higher number of treatment services received significantly predicts a lower rate of Violent misconduct for EnMs, but fails to be significant elsewhere. Related, a higher number of treatment services received significantly predicts a higher level of stock in rehabilitation for EnMs in this multiple group analysis, but is not significant for JCY. Compared with those previously displayed in Table 2, the R² values displayed above in Table 5 reflect the fact that the structural model built using institutional experiences explains more variance in misconduct outcomes than does the model built with individual characteristics.

Model Fit for Institutional Experiences Model

For the interview sample structural model that predicts misconduct using experiential constructs, approximate fit indices once again improve substantially when estimating a multiple groups model. As displayed in Table 6 below, this improvement in overall model fit is consistent, regardless of the misconduct outcome category.

Table 6.

Model Approximate Fit Statistics: Institutional Experiences Models.

	df	χ² (p value)	CFI	RMSEA	RMSEA CI
Any misconduct
Full model	4	5.662 (.226)	0.960	0.059	[0.000, 0.160]
Multiple groups	6	6.950 (.326)	0.977	0.051	[0.000, 0.181]
Violent misconduct
Full model	4	5.662 (.226)	0.917	0.059	[0.000, 0.160]
Multiple groups	6	6.950 (.326)	0.965	0.051	[0.000, 0.181]

Note. CFI = comparative fit index; RMSEA = root mean square error of approximation; CI = confidence interval.

Similar to the individual characteristics model, the nonsignificant chi-square values indicate that the degree of discrepancies between the covariance matrix implied by the model and the sampling covariance matrix is less than that expected by chance. This test statistic takes sampling error into account, which is especially important here because the sample size of N = 120 is relatively small. Significant improvements in fit occur, as indicated by the shift in CFI and RMSEA values when estimated across two groups, beyond that which occurs when parameters are constrained to be equal for these two groups.

Similar to the fit statistics produced from the individual characteristics model, it can be concluded that the same constructs in this institutional experiences model are manifested in different ways for these two groups. This further validates the ability of adult court status, here EnM status, to moderate the influence of multiple factors in this measurement structure on the outcome of institutional misconduct. In other words, EnM status exerts a more significant overall effect on variables related to misconduct than is reflected by including that status as a covariate in the model.

Discussion

Summary

The results of this study indicate that adult court status has the power to moderate the effects of some individual characteristics and individual experiences previously established as robust predictors of institutional misconduct. The fact that different factors predict Any misconduct and Violent misconduct for EnMs compared with JCY implies that the motives that drive behavior during juvenile incarceration differ considerably for “adult” juveniles compared with “juvenile” juveniles. In addition, staff across five juvenile correctional facilities in California may report misconduct differentially for EnMs compared with JCY. A longer sentence length significantly predicts an increase in the rate of Any misconduct for EnMs but fails to be significant for JCY. Given that these groups only marginally differ in sentence length descriptively, staff may be more aware of EnMs with lengthier sentences, perceive them to be more serious offenders, and write them up for more misconduct allegations overall as a result.

The significant findings from the multiple group analysis of the full population model suggest that EnM youth belonging to the “Other” category of race (which includes White and Asian youth) engage in a lower rate of Violent misconduct than African American EnMs. Any significant effects of race were absent for JCY. Considering that the proportion of EnMs and JCY do not differ by any racial category, this finding indicates that either (a) EnMs of certain races engage in more violence compared with others or (b) the classification of EnM status further aggravates the existing racialization of violence within DJJ facilities. The latter would be consistent with prior research that has shown African American youth to receive significantly more officially reported incidents of misconduct than White youth (Harer & Steffensmeier, 1996; Huebner, 2003; Kolivoski & Shook, 2016; Poole & Regoli, 1983; Trulson et al., 2016). The potential exists for discriminatory reporting practices toward African American EnMs, producing this intra-racial difference in Violent misconduct for EnMs, while reporting across racial groups appears to be more consistent for JCY.

Social connectedness is a measure of the extent to which youth maintain connections with those outside of DJJ. This construct theoretically mitigates the effects of Goffman’s (1961) “role dispossession” that encourages the development of barriers to the outside world during institutionalization. For JCY, social connectedness is significantly predictive of decreased misconduct (Any and Violent). Given that this construct includes measures of visitation and other contact with friends and family, this is consistent with prior research that has found increased visitation to decrease frequency of institutional misconduct over time (Cochran, 2012). This finding may speak to the fact that JCY will go home when released, which may motivate them to more diligently maintain those outside connections while incarcerated within DJJ. This ability of social connectedness to buffer misconduct did not extend to EnMs, however. Because they are awaiting a transfer to adult prison, this may decrease their motivation to strengthen social connections elsewhere, and may cause them to allow barriers between their roles in society and their roles during incarceration to be solidified.

Institutional support exerts beneficial effects for these two groups, depending on the outcome. A higher level of institutional support predicts a lower rate of Any misconduct for EnMs and alternatively predicts a lower rate of Violent misconduct for JCY. For EnMs, a higher level of support predicts less engagement in the more chronic, composite rate of any and all misconduct. This same type of support predicts a lower rate of specifically Violent misconduct for JCY. All five state-level facilities in this study would undoubtedly be categorized as custodial institutions with a high level of security. Anecdotally, it was clear during site visits that staff perceived EnMs to need even an additional level of supervision and security. This finding, then, suggests that institutional support from staff may mitigate the previously established relationship between high level of security and the resulting high levels of misconduct (Feld, 1981; Gover et al., 2000; MacDonald, 1999; McShane & Williams, 1989). Staff rapport exerts a positive influence on these two groups, but of different types: one of social control on EnMs, and one of discipline and deterrence for JCY.

Stock in rehabilitation marginally predicts a lower rate of specifically Violent misconduct for EnMs. This construct failed to be significant in the full structural model, which implies that this construct is moderated by EnM status. This is noteworthy because it validates the assumption that—to the extent we can instill a belief in rehabilitation within “adult” juvenile offenders—this belief has the ability to decrease involvement in violence for this “adultified” group during juvenile incarceration. In addition, the ability of treatment services received to significantly predict a higher level of stock in rehabilitation and to significantly predict a lower rate of Violent misconduct only held true for EnMs. Previous work in this area has revealed waived youth to receive fewer therapeutic and educational services during terms of incarceration (Kupchik, 2007; Penney & Moretti, 2005). The present study provides clear support to suggest that when provided with treatment services during juvenile incarceration, waived youth develop a stronger belief in the rehabilitative ideal, which decreases violence involvement during this term of incarceration.

Beyond that, it initially seems surprising that EnM youth who are facing an adult prison sentence would translate a belief in rehabilitation into improved behavior. Given that JCY are likely to go home sooner than EnMs, JCY have a much more tangible incentive to (a) find belief in treatment and rehabilitation and (b) to behave better as a result of that belief. However, these data provide the opposite finding, as stock in rehabilitation has no effect on JCY. The EnMs who are often viewed as “lost causes,” and who are headed to adult prison, actually engage in less violent behavior as a result of their belief in rehabilitation and the receipt of treatment services. These findings suggest that the knowledge of their inevitable transfer to a “worse” adult prison is a catalyst for EnMs motivation to capitalize on as much treatment in this “better” venue while they can. For EnMs, at least, the more treatment services you receive, the more you believe in your ability to change, and perhaps the more you believe in your ability to survive adult incarceration, which may influence better behavior during juvenile incarceration. Taken together, these findings suggest that if EnM status exerts its own pressure as a “pain of imprisonment,” there are ways to mitigate the effects of that status. Rather than EnM status paralyze any efforts toward rehabilitation during juvenile incarceration, the knowledge of an impending transfer to adult prison could be a catalyst for them to capitalize on benefits of institutional support and rehabilitative opportunities (Sykes, 1958). The fact that EnMs with belief in and access to rehabilitation engaged in less misconduct provides further support for the need to equalize the rehabilitative playing field within juvenile facilities for EnMs and JCY alike.

Strengths and Limitations

Using two different samples to estimate two theoretical models, this work contributes to juvenile policy research by providing some important new findings related to how adult court status, other individual characteristics, and institutional experiences may influence institutional misconduct. This study is the first to utilize multiple group analysis within SEM to examine the impact of adult court status as a moderating variable for two different measurement structures, the first focusing on individual characteristics and the second focusing on institutional experiences. Although some prior research has used a dichotomous measure of institutional misconduct (Kuanliang et al., 2008; Tasca et al., 2010; Trulson, 2007) or a total number of misconduct incidents during one’s term of commitment (Kolivoski & Shook, 2016), this study improves upon this by calculating a rate of misconduct over time for each youth. This allows for calculation of misconduct rates for youth at any point in their term of incarceration, which provides a more nuanced measure of this key outcome variable.

Although the present study is one of the first to utilize SEM and multiple group analysis to explore the ability of adult court status to moderate various factors for institutionalized youth, it is not without limitations. The study has a modest sample size, specifically where the institutional experiences model is concerned. As reliance on youth survey data increases for that model, the overall sample size decreases. However, despite a decrease in sample size, the multiple group analysis and associated model fit statistics suggest there to be sufficient statistical power to provide valid findings to demonstrate that self-reported experiences differentially predict behavior for adult court youth compared with their juvenile court counterparts.

Future research should work toward a larger sample, but this challenge should not diminish the strengths of the study. This study is strengthened by its ability to utilize both official agency data and primary self-report youth survey data in its analyses. Gaining access to interview institutionalized youth is very challenging, and this study is further strengthened by the primary youth survey data that shed light on individual experiences and youth perceptions during this term of incarceration within a state-level juvenile correctional facility.

Conclusion

The extent to which EnMs and JCY differ on parameters in these multiple group models provides a quantitative context on which to build with future qualitative analysis. However, the quantitative findings of this study stand on their own. In addition to parameter differences, the model fit findings are very important. Whether examining a theoretical structure related to individual characteristics or individual experiences during incarceration, both measurement structures differ substantially by court of commitment. Within a juvenile correctional environment, this “adult” status is powerful enough to moderate the influence of various factors on the outcome of misconduct. By acknowledging the power of this adult court status, we can appreciate the extent to which adultification impacts the overall dynamics and culture of juvenile correctional facilities. In most states, adult court youth are transferred to adult prison at 18 years, whereas JCY are allowed to remain under juvenile jurisdiction until the age of at least 21 years, or in California until 25 years of age. If the looming transfer to adult prison has the ability to influence behavior of adult court youth (here, EnMs) in way that might impact the institutional culture of juvenile correctional facilities, then the findings of this study have significant policy implications for more efficient operation of juvenile correctional facilities. Since the time of our data collection, California has begun to acknowledge the importance of keeping youth and even young adults (those aged 18-25 years) under juvenile correctional jurisdiction beyond their 18th birthdays. Although waiver to adult court is still utilized, the detrimental effects of adult prison are well known, and policymakers within DJJ and at the county level have taken initial steps to divert young adults away from state prisons and jails to serve their sentences within juvenile correctional environments.

Related to the earlier implication of equalizing the playing field for all youth committed to juvenile correctional facilities, this research provides considerable support for acknowledging the ability of waived youth to take advantage of rehabilitative opportunities in a way that can positively impact the institutional culture of juvenile facilities. These findings support efforts toward increasing access of treatment services for waived youth. More broadly, this work should encourage juvenile correctional facilities to make a concerted effort to provide equal opportunities for rehabilitation to “juvenile” juveniles and “adult” juveniles in the hopes of decreasing rates of misconduct and restoring the rehabilitative focus of juvenile correctional facilities.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: California Department of Corrections and Rehabilitation (CDCR), Agreement #5600001351. The opinions expressed herein represent those of the author(s) and do not necessarily represent the positions of the CDCR.

Notes

References

Barak-Glantz

I. L.

(1983). Patterns of prisoner misconduct: Toward a behavioral test of prisonization. Sociological Focus, 16, 129-146.

Barnes

C. W.

Franz

R. S.

(1989). Questionably adult: Determinants and effects of the juvenile waiver decision. Justice Quarterly, 6, 117-136.

Bechtold

Cauffman

(2013). Tried as an adult, housed as a juvenile: A tale of youth from two courts incarcerated together. Law and Human Behavior, 38, 126-138.

Berg

M. T.

DeLisi

(2006). The correctional melting pot: Race, ethnicity, citizenship, and prison violence. Journal of Criminal Justice, 34, 631-642.

Bishop

(2000). Juvenile offenders in the adult criminal justice system. Crime and Justice, 27, 81-167.

Bollen

K. A.

(1989). Structural equations with latent variables. New York, NY: John Wiley.

Clemmer

(1940). The prison community. Boston, MA: Christopher.

Cochran

J. C.

(2012). The ties that bind or the ties that break: Examining the relationship between visitation and prisoner misconduct. Journal of Criminal Justice, 40, 433-440.

Cunningham

M. D.

Sorensen

J. R.

(2007). Predictive factors for violent misconduct in close custody. The Prison Journal, 87, 241-253.

10.

DeLisi

Berg

M. T.

Hochstetler

(2004). Gang members, career criminals and prison violence: Further specification of the importation model of inmate behavior. Criminal Justice Studies, 17, 369-383.

11.

DeLisi

Spruill

J. O.

Peters

D. J.

Caudill

J. W.

Trulson

C. R.

(2013). “Half in, half out”: Gang families, gang affiliation, and gang misconduct. American Journal of Criminal Justice, 38, 602-615.

12.

DeLisi

Trulson

C. R.

Marquart

J. W.

Drury

A. J.

Kosloski

A. E.

(2010). Inside the prison black box: Toward a life course importation model of inmate behavior. International Journal of Offender Therapy and Comparative Criminology, 55, 1186-1207.

13.

Fagan

Kupchik

(2011, April). Juvenile incarceration and the pains of imprisonment (Duke Forum for Law & Social Change, Columbia Public Law Research Paper No. 11-263). New York, NY: Columbia Law School.

14.

Feld

B. C.

(1981). A comparative analysis of organizational structure and inmate subcultures in institutions for juvenile offenders. Crime and Delinquency, 27, 336-363.

15.

Fernandez

K. E.

Neiman

(1998). California’s inmate classification system: Predicting inmate misconduct. The Prison Journal, 78, 40-422.

16.

Flanagan

T. J.

(1983). Correlates of institutional misconduct among state prisoners. Criminology, 21, 29-39.

17.

Forst

Fagan

Vivona

T. S.

(1989). Youth in prisons and training schools: Perceptions and consequences of the treatment-custody dichotomy. Juvenile & Family Court Journal, 40, 1-14.

18.

Goffman

(1961). Asylums: Essays on the social situation of mental patients and other inmates. New York, NY: Random House.

19.

Gover

A. R.

MacKenzie

D. L.

Armstrong

G. S.

(2000). Importation and deprivation explanations of juveniles’ adjustment to correctional facilities. International Journal of Offender Therapy and Comparative Criminology, 44, 450-467.

20.

Griffin

M. L.

Hepburn

J. R.

(2006). The effect of gang affiliation on violent misconduct among inmates during the early years of confinement. Criminal Justice and Behavior, 33, 419-466.

21.

Griffin

Addie

Adams

Firestine

(2011). Trying juveniles as adults: An analysis of state transfer laws and reporting. Washington, DC: U.S. Department of Justice, Office of Justice Programs, Office of Juvenile Justice and Delinquency Prevention.

22.

Harer

M. D.

Steffensmeier

D. J.

(1996). Race and prison violence. Criminology, 34, 323-356.

23.

Hochstetler

DeLisi

(2005). Importation, deprivation, and varieties of serving time: An integrated-lifestyle-exposure model of prison offending. Journal of Criminal Justice, 33, 257-266.

24.

Howell

J. C.

(1996). Juvenile transfers to the criminal justice system: State of the art. Law & Policy, 18(1&2), 17-60.

25.

Huebner

B. M.

(2003). Administrative determinants of inmate violence: A multilevel analysis. Journal of Criminal Justice, 31, 107-117.

26.

Huebner

B. M.

Varano

S. P.

Bynum

T. S.

(2007). Gangs, guns, and drugs: Recidivism among serious, young offenders. Criminology & Public Policy, 6, 187-222.

27.

Irwin

Cressey

D. R.

(1962). Thieves, convicts and the inmate culture. Social Problems, 10, 142-155.

28.

Jiang

Fisher-Giorlando

(2002). Inmate misconduct: A test of the deprivation, importation, and situational models. The Prison Journal, 82, 335-358.

29.

Kline

R. B.

(2011). Principles of structural equation modeling (3rd ed.). New York, NY: The Guilford Press.

30.

Kolivoski

K. M.

Shook

J. J.

(2016). Examining the relationship between age and prison behavior in transferred juveniles. Criminal Justice and Behavior, 43, 1242-1259.

31.

Krisberg

(2005). Juvenile justice: Redeeming our children. Thousand Oaks, CA: Sage.

32.

Kuanliang

Sorensen

J. R.

Cunningham

M. D.

(2008). Juvenile inmates in an adult prison system: Rates of disciplinary misconduct and violence. Criminal Justice and Behavior, 35, 1186-1201.

33.

Kupchik

(2006). Judging juveniles: Prosecuting adolescents in adult and juvenile courts. New York: New York University Press.

34.

Kupchik

(2007). The correctional experiences of youth in adult and juvenile prisons. Justice Quarterly, 2, 247-270.

35.

Langan

N. P.

Camp

S. D.

Saylor

W. G.

(2004). Is prison misconduct analogous to crime. Washington, DC: Federal Bureau of Prisons.

36.

Lattimore

P. K.

MacDonald

J. M.

Piquero

A. R.

Linster

R. L.

Visher

C. A.

(2004). Studying the characteristics of arrest frequency among paroled youthful offenders. Journal of Research in Crime and Delinquency, 41, 37-57.

37.

MacDonald

J. M.

(1999). Violence and drug use in juvenile institutions. Journal of Criminal Justice, 27, 33-44.

38.

Maruna

Toch

(2005). The impact of imprisonment on the desistance process. In Travis

Visher

(Eds.), Prisoner reentry and crime in America (pp. 139-178). New York, NY: Cambridge University Press.

39.

Maxson

Bradstreet

Gascon

Gerlinger

Grebenkemper

Haerle

. . . Scott

(2011). Gangs and violence in California’s youth correctional facilities: A research foundation for developing effective gang policies. Report submitted to the California Department of Corrections and Rehabilitation’s Division of Juvenile Justice, Sacramento, CA.

40.

McShane

M. D.

Williams

F. P.

(1989). The prison adjustment of juvenile offenders. Crime and Delinquency, 35, 254-269.

41.

Office of Research. (2010). Juvenile justice outcome evaluation report. Sacramento: California Department of Corrections and Rehabilitation.

42.

Penney

S. R.

Moretti

M. M.

(2005). The transfer of juveniles to adult court in Canada and the United States: Confused agendas and compromised assessment procedures. International Journal of Forensic Mental Health, 4(1), 19-37.

43.

Poole

E. D.

Regoli

R. M.

(1983). Violence in juvenile institutions: A comparative study. Criminology, 21, 213-232.

44.

Rosenberg

(1965). Society and the adolescent self-image. Princeton, NJ: Princeton University Press.

45.

Schrag

(1961). Some foundations for a theory of correction. In Cressey

D. R.

(Ed.), The prison: Studies in institutional organization and change (pp. 309-357). New York, NY: Holt.

46.

Sykes

G. M.

(1958). The society of captives: A study of a maximum security prison. Princeton, NJ: Princeton University Press.

47.

Sykes

G. M.

Messinger

S. L.

(1960). The inmate social system. In Cloward

Cressey

D. R.

Grosser

G. H.

McCleery

Ohlin

L. E.

Sykes

G. M.

Messinger

S. L.

(Eds.), Theoretical studies in social organizations of the prison (pp. 5-119). New York, NY: Social Science Research Council.

48.

Tasca

Griffin

M. L.

Rodriguez

(2010). The effect of importation and deprivation factors on violent misconduct: An examination of Black and Latino youth in prison. Youth Violence and Juvenile Justice, 8, 234-249.

49.

Thomas

C. L.

Foster

S. C.

(1972). Prisonization in the inmate contraculture. Social Problems, 20, 229-239.

50.

Thomas

C. W.

(1977). Theoretical perspectives on prisonization. Journal of Criminal Law and Criminology, 68, 135-145.

51.

Tittle

C. R.

Tittle

D. P.

(1964). Social organization of prisoners: An empirical test. Social Forces, 43, 216-221.

52.

Trulson

C. R.

(2007). Determinants of disruption: Institutional misconduct among state-committed delinquents. Youth Violence and Juvenile Justice, 5(1), 7-34.

53.

Trulson

C. R.

DeLisi

Marquart

J. W.

(2009). Institutional misconduct, delinquent background, and rearrest frequency among serious and violent delinquent offenders. Crime & Delinquency, 57, 709-731.

54.

Trulson

C. R.

Haerle

D. R.

Caudill

J. W.

DeLisi

(2016). Lost causes: Blended sentencing, second chances, and the Texas Youth Commission. Austin: University of Texas Press.

55.

Trulson

C. R.

Marquart

J. W.

Mullings

J. L.

Caeti

T. J.

(2005). In between adolescence and adulthood: Recidivism outcomes of a cohort of state delinquents. Youth Violence and Juvenile Justice, 3, 355-387.

56.

Ward

Kupchik

(2009). Accountable to what? Professional orientations towards accountability-based juvenile justice. Punishment & Society, 11, 85-109.

57.

Wellford

(1967). Factors associated with adoption of the inmate code. Journal of Criminal Law, Criminology, and Police Science, 58, 197-203.

58.

Wheeler

(1961). Socialization in correctional communities. American Sociological Review, 26, 697-712.