Labeling theory

Labeling and the Effect of Adolescent Legal System Involvement on Adult Outcomes for Foster Youth

Aging Out of Care

JoAnn S. Lee University of Washington

Mark E. Courtney University of Chicago

Tracy W. Harachi and Emiko A. Tajima University of Washington

This study uses labeling theory to examine the role that adolescent legal system involvement may play in initiating a process of social exclusion, leading to higher levels of adult criminal activities among foster youth who have aged out of care. We used data from the Midwest Evaluation of the Adult Functioning of Former Foster Youth (Midwest Study), a prospective study that sampled 732 youth from Illinois, Iowa, and Wisconsin as they were preparing to leave the foster care system at ages 17 or 18. The youth were interviewed again at ages 19, 21, and 23 or 24. We used structural equation modeling to examine pathways to self-reported adult criminal behaviors from juvenile legal system involvement. The path model indicated that legal system involvement as a juvenile was associated with a lower likelihood of having a high school diploma at age 19, which was associated with a reduced likelihood of employment and increased criminal activities at age 21. Legal system involvement is more common among foster youth aging out of care, and this legal system involvement appears to contribute to a process of social exclusion by excluding former foster youth from conventional opportunities.

A bout 23,000 youth emancipated from the foster caresystem in fiscal year 2013 (Children’s Bureau, 2014).Foster youth aging out of care report high levels of engagement in delinquent behavior and legal system involve- ment (e.g., arrests, convictions, and/or incarceration) as a juve- nile and are at high risk for continued engagement in criminal activities and legal system involvement as adults (Courtney et al., 2011; Reilly, 2003; Southerland, Casanueva, & Ringeisen, 2009; Vaughn, Shook, & McMillen, 2008). Chronic risk factors present in many foster youth may explain both the high rates of legal system involvement as a juvenile and later adult criminal behavior. These chronic risk factors include histories of mal- treatment (Maxfield & Widom, 1996; Smith & Thornberry, 1995), a history of mental health or substance abuse (Maschi, Hatcher, Schwalbe, & Rosato, 2008), and/or low academic achievement or trouble in school (Blomberg, Bales, Mann, Piquero, & Berk, 2011; Maschi et al., 2008). In addition,

experiences in the child welfare system, such as placement instability and placement type, may also be related to both juvenile legal system involvement and adult criminal behaviors (Jonson-Reid & Barth, 2000; Ryan, Marshall, Herz, & Hernan- dez, 2008; Ryan & Testa, 2005). However, less studied is the possibility that legal system involvement as a juvenile may be an independent risk factor for later criminal behavior.

Foster youth who are aging out of care are at a critical cross- roads in their lives because they are transitioning out of care and into adulthood simultaneously. These foster youth are making choices about their adult commitments for the first time, choices that may affect the rest of their lives. The transition to adulthood has been characterized as a period of freedom from social norms and obligations to allow individuals the opportunity to explore possibilities before making long-term commitments (Arnett, 2006), although this latitude may also heighten the risk of engage- ment in criminal behaviors for some youth. Successful transitions into work, marriage, and parenting have been associated with rerouting previously negative trajectories (Crutchfield & Pitchford, 1997; Kreager, Matsueda, & Erosheva, 2010; Sampson & Laub, 1990). In contrast, a criminal record has a lasting, negative impact on the lives of youths, and may interfere with their ability to participate fully in society (Alexander, 2011). This study seeks to learn whether these high rates of involvement in the juvenile justice system among youth aging out of care may increase their risk for engaging in criminal activities in adulthood by initiating a process of social exclusion.

JoAnn S. Lee, School of Social Work, University of Washington; Mark E. Courtney, School of Social Service Administration, University of Chi- cago; Tracy W. Harachi and Emiko A. Tajima, School of Social Work, University of Washington.

Correspondence concerning this article should be addressed to JoAnn S. Lee, who is now at the Department of Social Work, George Mason University, 4400 University Drive, MSN-1F8, Fairfax, VA 22030. E-mail: [email protected]

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an Ps

yc ho

lo gi

ca l

A ss

oc ia

tio n

or on

e of

its al

lie d

pu bl

is he

rs .

T hi

s ar

tic le

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

American Journal of Orthopsychiatry © 2015 American Orthopsychiatric Association 2015, Vol. 85, No. 5, 441–451 http://dx.doi.org/10.1037/ort0000090

441

Criminal Behaviors in Adulthood: Theoretical Rationale

Labeling theory claims that the experience of negative social labels, such as “deviant,” “delinquent,” or “criminal,” leads to future deviant behavior (Becker, 1991; Lemert, 1951). Thus, legal system involvement is hypothesized to operate as a negative social label that increases the likelihood of later adult criminal behaviors. Recent labeling theorists have hypothesized that the application of a label may initiate a process of social exclusion where conven- tional opportunity structures, such as educational and employment opportunities, become blocked (Paternoster & Iovanni, 1989; Sampson & Laub, 1997). This process also encompasses the potential stigmatization an individual may experience in their daily lives and the individual’s consequent withdrawal from conven- tional society (Bernburg, 2009). In this sense, the label is hypoth- esized to initiate a process that weakens the individual’s bond to conventional social institutions (Sampson & Laub, 1997), such as an ex-felon who loses the right to vote, faces limited job oppor- tunities, and cannot access crucial safety net programs such as housing and food stamps (Alexander, 2011). There is evidence for the relationship between a label and later deviance generally (Huizinga, Schumann, Ehret, & Elliott, 2004; McAra & McVie, 2007) and more specifically between labeling in the form of legal system involvement as a juvenile and later adult criminal outcomes (Bernburg & Krohn, 2003; Gatti, Tremblay, & Vitaro, 2009), as well as evidence supporting the social exclusion mechanism (Bernburg & Krohn, 2003; Sampson & Laub, 1997). To our knowledge, labeling theory has not been tested within a sample of foster youth.

Juvenile Justice Involvement Among Foster Care Youth

Among studies that have tested the labeling hypothesis, legal system involvement is commonly operationalized as a formal label (Bernburg, Krohn, & Rivera, 2006; Huizinga et al., 2004; Jackson & Hay, 2013; Lopes et al., 2012; Restivo & Lanier, 2013). Label- ing theory predicts that disadvantaged youth are more likely to be labeled (Paternoster & Iovanni, 1989), and consistent with this prediction, foster youth appear to experience higher levels of juvenile justice involvement than youth from the general popula- tion. Studies indicate that there is a disproportionate percentage of child welfare involved youth in the juvenile justice system (Herz, Ryan, & Bilchik, 2010; Ross, Conger, & Armstrong, 2002; Ryan, Herz, Hernandez, & Marshall, 2007), with an increasing percent- age of child welfare youth at deeper levels (e.g., arrest, detention, incarceration) of the juvenile justice system (Halemba, Siegel, Lord, & Zawacki, 2004). These patterns in juvenile justice system involvement do not appear to be related to differences in delin- quent behaviors. One study using juvenile justice data from New York City found that although foster youth were detained at a higher rate than those not involved in the foster care system, there was no evidence that the foster youth committed more crimes or more severe crimes (Ross et al., 2002). Higher rates of legal system involvement have also been documented among foster youth aging out of care in comparison to a general population sample, although differences in potentially confounding factors such as self-reported delinquent behavior were not taken into

account in the comparison (Cusick, Courtney, Havlicek, & Hess, 2011).

In addition to empirical evidence, the increased rate of involve- ment into the juvenile justice system by child welfare involved individuals compared with nonchild welfare children and youth is arguably consistent with the ideological underpinnings of the juvenile justice system. The juvenile justice system was designed to consider the youth’s social circumstances more strongly than the youth’s guilt (Feld, 1999), and factors such as child and family functioning are considered “paramount” in helping authorities make decisions in the best interest of the child (Fader, Harris, Jones, & Poulin, 2001; Gebo, 2005). In fact, Fader et al. (2001) examined administrative data in Philadelphia to identify the factors that influenced decisions to commit juvenile offenders to delin- quency programs, using only information that would have been available to judges. They found that a history of being reported to the child welfare system was an important factor in determining whether an individual was committed to a residential intervention program, and was the most important factor for decisions involv- ing first time offenders with a history of drug use (Fader et al., 2001). These higher rates of juvenile justice involvement are troubling if legal system involvement acts as a label to initiate a process that increases, rather than decreases, the likelihood of later adult criminal behaviors.

Processes of Social Exclusion Social exclusion can be “defined as chronic, multidimensional

disadvantage resulting in catastrophic detachment from society” (Axford, 2010, p. 738). Social exclusion is not simply a state of lacking resources (i.e., poverty), but is a relational term that reflects an “inability to participate fully in society” (Axford, 2010, p. 738). Former foster youth report a range of negative outcomes as they age out of care, including limited educational attainment, limited employment experiences, substance use and abuse, home- lessness, economic instability and involvement in criminal activ- ities (Berzin, 2008; Courtney et al., 2011; Pecora et al., 2006; Reilly, 2003). These negative outcomes challenge former foster youths’ ability to successfully transition into adult roles and con- ventional social institutions, such as the labor market and mar- riage. Thus, a process of social exclusion may be operating among many former foster youth.

Studies that have examined the social exclusion mechanism hypothesized by labeling theory have been limited but promising. There is evidence that incarceration as an adolescent is associated with weakened bonds to conventional social institutions, as evi- denced by an increased likelihood of dropping out of high school and a lower likelihood of high school graduation (Hjalmarsson, 2008; Kirk & Sampson, 2013) and lower likelihood of being employed and if employed, stably employed (Apel & Sweeten, 2010; Sampson & Laub, 1997).

Two studies have examined the full pathway from legal system involvement in adolescence to later adult crime, providing support for the social exclusion mechanism of the labeling theory hypoth- esis. Both studies used data from the Rochester Youth Develop- ment Study to fully address the developmental nature of labeling theory by examining crime outcomes in adulthood. Bernburg and Krohn (2003) tested paths from police and official intervention in adolescence to later high school graduation and attendance, unem-

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an Ps

yc ho

lo gi

ca l

A ss

oc ia

tio n

or on

e of

its al

lie d

pu bl

is he

rs .

T hi

s ar

tic le

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

442 LEE, COURTNEY, HARACHI, AND TAJIMA

ployment, and adult crime in early adulthood (ages 19–22). They found that while police and official intervention continues to have a direct effect on later adult crime while controlling for adolescent self-reported delinquency, academic aptitude, and demographic factors, the association was partially mediated by educational attainment and periods of nonemployment. However, their analy- ses only included males.

Lopes and colleagues’ (2012) study is unique in that it examined outcomes in later adulthood (ages 29–31), after the age-graded peak in offending during late adolescence or early adulthood (Agnew, 2003). Similar to Bernburg and Krohn’s earlier study, they found that, while controlling for earlier delinquency and risk factors, labeling through police contact or an arrest in adolescence was associated with increased unemployment and lower rates of high school graduation in early adulthood, which in turn was associated with economic unsustainability and adult drug use. Thus, legal system involvement in adolescence may be related to educational and employment outcomes in early adulthood, which in turn may be related to negative trajectories through adulthood.

There is growing evidence that legal system involvement as an adolescent may be associated with a process of exclusion leading to ongoing adult criminal behaviors, but this theory has not been tested in a sample of foster care youth. Foster youth aging out of care are a vulnerable population who are disproportionately rep- resented in the juvenile justice system (Herz et al., 2010; Ross et al., 2002; Ryan et al., 2007) and for whom the state bears respon- sibility as a parent following the parens patriae principle of the state as parent (Courtney, 2009). It is important to understand whether and how system intervention potentially contributes to processes of social exclusion.

The Current Study This study provides a test of the labeling theory hypothesis in a

new, high-risk population using a sample of foster youth as they are aging out of care and transitioning to adulthood (ages 17–24). The prospective sample of foster youth allows for a test of the theory using data that spans developmental periods, from adoles- cence to early adulthood. This sample of foster youth includes both males and females, and those with and without legal system involvement.

Much of the work on relationships between maltreatment, child welfare involvement, and criminal outcomes looks broadly at criminal outcomes and often operationalizes outcomes using ad- ministrative data on arrests or other forms of legal system involve- ment (Herz et al., 2010; Jonson-Reid, 2002; Jonson-Reid & Barth, 2000; Ryan et al., 2008; Yampolskaya, Armstrong, & McNeish, 2011). However, delinquent/criminal behaviors and legal system involvement may each contribute uniquely to the relationships being tested.1 While delinquent behavior reflects the actions of an individual, referred to as “behaviors of individuals,” systemic biases may be captured within legal system involvement, referred to as “the behavior of law” (Uggen & Kruttschnitt, 1998, p. 339). For example, an individual living in a more disadvantaged and heavily policed neighborhood could be more likely to be arrested

for a given delinquent act in comparison to an individual perpe- trating the same act but living in a less heavily policed neighbor- hood (Sampson, 1986). Furthermore, individuals may be arrested for crimes they did not commit. Thus, in contrast to prior studies that use either criminal behaviors or arrests as a single measure of criminal outcomes, we include measures of both in this study to identify how delinquent behaviors and legal system involvement may differentially influence the processes being tested.

This study asks two questions. First, among foster youth aging out of care, is legal system involvement as a juvenile associated with increased criminal behaviors in adulthood, even when taking into account prior delinquent behaviors? Second, if so, is this relationship mediated by bonds to conventional social institutions, namely education, employment, and parenthood? We hypothesize that among foster youth, legal system involvement will be associ- ated with later adult criminal behaviors, and that this association will be mediated by education and employment outcomes.

Method

Sample

This study uses survey data from the Midwest Evaluation of the Adult Functioning of Former Foster Youth (Midwest Study), a prospective study that followed 732 youth from Illinois, Iowa and Wisconsin as they were preparing to age out of the foster care system. The Midwest Study drew a random sample of two thirds of youth from Illinois and all eligible youth from Wisconsin and Iowa who had been in the foster care system for at least a year and were age 17 and thus preparing to transition to independence. After the initial interview, participants were reinterviewed approxi- mately every 2 years through age 26 with an over 80% retention rate at each interview. Youth were asked questions about a range of topics, including prefoster care experiences of abuse and ne- glect, foster care experiences, and outcomes including education, employment, parenting, and delinquency and legal system involve- ment. This study uses interviews from Waves 1–4 because these covered both the peak years in offending (late adolescence and early adulthood) and the pivotal initial transition out of care; the first wave of interviews was conducted between May 2002 and March 2003 when participants were 17–18 years old, and the fourth wave of interviews was conducted in July 2008–April 2009 when participants were 23–24 years old.

Measures

Self-reported delinquent behaviors. A baseline de- linquency variable was created from 11 items in the baseline interview. These questions, which were also asked at subsequent interviews, asked about the frequency of engagement in delinquent behaviors in the last 12 months, such as damaged property that didn’t belong to the participant, stole something worth more than $50, and took part in a gang fight. Participants were given four response categories: 0 � never, 1 � 1 or 2 times, 2 � 3 or 4 times, and 3 � 5 or more times. Responses to these 11 questions were summed for a baseline measure of delinquency and had values ranging from 0–28 because no participant reported the highest level of participation for all of the questions.

1Delinquency refers to illegal behavior of adolescents while crime refers to illegal behavior of adults.

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an Ps

yc ho

lo gi

ca l

A ss

oc ia

tio n

or on

e of

its al

lie d

pu bl

is he

rs .

T hi

s ar

tic le

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

443LABELING AMONG FORMER FOSTER YOUTH

The engagement in criminal activities variables used as the out- come for the multivariate regression models were constructed from 17 questions in the adult interviews at waves 3 and 4, when participants were 21 and 23 or 24 years old. Participants were asked about the frequency of engagement in criminal behaviors in the last 12 months, such as the frequency with which they bought, sold, or held stolen property; deliberately wrote a bad check; and used a weapon in a fight. Participants were given the same four response categories as in the baseline delinquency questions. Responses to these questions were summed to create a composite variable of the types and frequency of engagement in those acts, and the created variables had values ranging from 0 to 34 because no participant reported the highest level of engagement for all of the questions.

Juvenile/criminal justice system involvement. A single dichotomous variable (1 � yes, 0 � no) was created to indicate whether the youth reported that they had been arrested, convicted, or incarcerated prior to the first interview.2 This base- line measure, legal system involvement as a juvenile, was included in all models.

Conventional social institutions. Four dichotomous items were included that measured an individual’s bonds to con- ventional social institutions (e.g., school, work, and family) at Waves 1, 2 and 3, when participants were ages 17, 19, and 21. The first was educational attainment, a dichotomous measure that either indicated less than a high school degree (0) or at least a high school diploma or equivalence (1). Also included were, school enrollment, a measure that indicated whether they were enrolled in a school or vocational training program at the time of the inter- view; employment status, a measure that indicated whether they were employed at the time of the interview; and parenthood, a measure that indicated whether they had a child at baseline or whether they had a child living with them in subsequent inter- views. Although parenthood for participants at baseline may be considered “early” and potentially problematic, there is some evidence that the transition to parenthood may be associated with desistance from crime and reduction in arrests among disadvan- taged populations (Kreager, Matsueda, & Erosheva, 2010; Lee, Courtney, & Hook, 2012; Lee, Courtney, & Tajima, 2014). As such, parenthood may provide informal social control for these youth in the same way as bonds to other social institutions. At baseline, none of the participants reported that they had a child living with them and almost all of the participants were still in care. In subsequent interviews, some or all of the participants had exited care depending on the interview; thus, after baseline, a resident child was more likely to indicate that the participant was actively parenting and thus bonded to their new family. These last three variables, school enrollment, employment status, and parent-

hood, were dichotomous variables where 1 � yes and 0 � no. Educational and employment outcomes also reflect conventional opportunity structures.

Control variables. Control variables included in the mod- els consisted of demographic variables (gender and race) and state indicators, whether the youth was from Illinois, Wisconsin, or Iowa. Scores on the word recognition portion of the Wide Range Achievement Test (WRAT) were included as a measure of reading ability. Indicators of early risk factors include a measure of family conflict, which was based on whether the participant said that their caregiver abused their spouse (yes or no), measured at baseline. Their histories of maltreatment were also measured using the Lifetime Experiences Questionnaire (Rose, Abramson, & Kaupie, 2000). Histories of neglect (nine items such as caretaker fails to provide regular meals and caretaker failed to protect participant from harm) and physical abuse (7 items such as caretaker beat up participant and caretaker hit or kicked participant) were measured at baseline, while histories of sexual abuse (8 items such as unwanted sexual contact with penis and unwanted finger contact in private sexual parts) were measured at Wave 2, when all the participants were over age 18. In addition, two measures of mental well-being at baseline were included: a dichotomous measure of symptoms of posttraumatic stress syndrome (PTSD) or depression and a dichotomous measure of symptoms of alcohol or substance use or abuse.

Two measures of child welfare experience, measured at base- line, were also included in the models: number of placements, and last placement at baseline. The last placement at baseline is a categorical variable with four categories: foster family with or without relatives and adoption, group home, independent living, or other including 34 youth who had already emancipated from the system. The first category, foster family with or without relatives, was collapsed after the coefficient for foster family with relatives was not significant in reference to foster family without relatives.

Two interaction terms were also included in the models to test the potential moderating effect of gender on legal system involve- ment as a juvenile and self-reported delinquency at baseline. See Table 1 for a summary of the descriptive statistics.

Analyses

Analyses were conducted in MPlus 6.12 (Muthén & Muthén, 1998). Maximum likelihood robust estimators were used in all models to take into account the non-normality of the outcome variables. To consider how system involvement is related to future criminal adult behaviors, two negative binomial regression models were estimated. Self-reported engagement in criminal activities at ages 21 and 23 or 24 were regressed on legal system involvement as a juvenile, while controlling for earlier self-reported delin- quency or criminal activities. Because the two criminal activities variables have a count distribution, negative binomial models were estimated (Long, 1997). In all estimated models, the dispersion parameter was statistically significant, indicating that the negative binomial, rather than the Poisson model, was the appropriate choice. When exponentiated, the coefficients in a negative bino- mial regression model are incidence rate ratios (IRRs), and can be interpreted as the likelihood of the incidence of an event. Full models were also estimated that included legal system involve-

2The dichotomous measure of legal system involvement provides a comparison between those who have experienced any type and degree of formal label and those who have not. We also tested these models with a more nuanced measure, that of legal system involvement as a juvenile. This variable was treated as an ordinal variable, and included the following four categories: no legal system involvement as a juvenile, arrests, convictions, and/or incarceration as a juvenile. Youth were classified based on their deepest level of involvement into the system. The use of this more nuanced measure did not substantially change the results for the estimated models.

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an Ps

yc ho

lo gi

ca l

A ss

oc ia

tio n

or on

e of

its al

lie d

pu bl

is he

rs .

T hi

s ar

tic le

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

444 LEE, COURTNEY, HARACHI, AND TAJIMA

ment and the four measures of conventional social institutions (educational attainment, school enrollment, employment status, and parenthood) while controlling for earlier delinquency. The covariance between legal system involvement and baseline delin- quency was also estimated, to take into account the association between these two variables. Other variables, such as symptoms of depression or PTSD, were included in the models based on whether model fit indices indicated improved fit. Model fit was evaluated using the Akaike information criterion (AIC) and the Bayesian information criterion (BIC), which identifies optimal models by evaluating how closely values predicted by the esti- mated model match up to observed data (Agresti, 2002). Although both provide similar information, the BIC adds a penalty for additional parameters, thus favoring more parsimonious models. In addition, the Wald �2 test was used to compare nested models, more specifically, to evaluate the impact on model fit statistics of adding the four bonds to social institutions variables (i.e., school enrollment, educational attainment, employment status, and par- enthood) to the model.

To test whether the relationship between legal system involve- ment as a juvenile and later criminal activities is mediated by bonds to social institutions, a structural model was estimated including the multivariate models already estimated. Mediating pathways were modeled by regressing school enrollment, educa-

tional attainment (no attainment of a high school diploma vs. at least a high school diploma or equivalence), employment status, and parenthood on legal system involvement as a juvenile while controlling for delinquency. Because these four outcomes were dichotomous variables, logistic regression was used to model these mediating models. Covariance between baseline delinquency and legal system involvement as a juvenile, as well as the covariance between school enrollment and the lack of a high school diploma or equivalency at age 19 were estimated in the models, to take into account the associations between both pairs of variables. Model fit was evaluated with the AIC and BIC. Direct, indirect, and total pathways between adult criminal activities and legal system in- volvement as a juvenile through the four bonds to social institu- tions variables were calculated.

Multiple imputation was used to address missing values on a range of variables including the indicators of family conflict and child maltreatment, which were missing approximately 5% of values, and the measures of mental well-being that were missing 7%–9% of values. Multiple imputation in MPlus uses Bayesian estimation using available information in the vari- ables included and considered in the model, as well as addi- tional auxiliary variables that were not included in the models but were related to the imputed variables (age, born in the United States, emotional or psychological counseling, and sub-

Table 1. Descriptive Statistics

Baseline Age 19 Age 21 Age 23/24

% or M (SD) % or M (SD) % or M (SD) % or M (SD)

Delinquency/criminal activities 3.22 (4.15) 1.81 (3.27) 1.44 (2.85) Legal system involvement 0.56 Employed 0.31 0.39 0.54 Enrolled 0.95 0.48 0.24 No high school diploma or equivalency 0.88 0.40 0.23 Parenthood 0.86 0.17 0.32 Male 0.49 Race

White 0.29 Black 0.55 Other 0.16

State Illinois 0.64 Wisconsin 0.27 Iowa 0.09

WRAT Score 39.36 (8.41) Family conflict 0.25 History of maltreatment

Neglect 0.17 Physical Abuse 0.31 Sexual Abuse 0.16 Depression/PTSD symptoms 0.25 Drug/alcohol symptoms 0.22 Number of placements 5.83 (5.85)

Last placement Foster home, relatives, or adoption 0.67 Group home 0.18 Independent living 0.09 Other 0.06

Note. WRAT � Wide Range Achievement Text; PTSD � posttraumatic stress disorder.

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an Ps

yc ho

lo gi

ca l

A ss

oc ia

tio n

or on

e of

its al

lie d

pu bl

is he

rs .

T hi

s ar

tic le

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

445LABELING AMONG FORMER FOSTER YOUTH

stance abuse treatment) to produce 50 imputed datasets (Muthén & Muthén, 1998). Models were estimated for all 50 datasets using maximum likelihood estimators and averaged across datasets, and standard errors were calculated using the Rubin formula (Muthén & Muthén, 1998).

Results

Adult Criminal Behaviors

The youth who reported legal system involvement as a juvenile reported higher levels of adult criminal activities at age 21, Mlegalsystem � 2.05 versus Mnolegalsystem � 0.87, t(503) � 3.96, p � .001, and at age 23 or 24, Mlegalsystem � 1.71 versus Mnolegalsystem � .84, t(587) � 3.61, p � .0002. This relationship persists at age 21 when controlling for baseline delinquency (B � .380, p � .05). However, at ages 23 or 24, the coefficient for whether the youth reported legal system involvement as a juvenile was not statistically significant when controlling for baseline de- linquency and criminal behaviors at age 21.

Table 2 presents the results for the two full negative binomial multivariate models estimated for criminal activities at age 21 and

age 23 or 24. At age 21, the coefficient for legal system involve- ment does not remain significant with the addition of covariates. The coefficients for self-reported delinquency at baseline, being male, a history of neglect, a history of sexual abuse, and more foster care placements all had a positive and significant relation- ship with later adult criminal activity. The interaction term be- tween gender and delinquency was significant, indicating that the relationship between delinquency and engagement in criminal activities at age 21 was slightly smaller for men (B � �.075, p � .05), but men start at a higher level of overall criminal activities than women as indicated by the coefficient for being male (B � .889, p � .001). The interaction term between gender and legal system involvement was not significant.

Of the four variables measuring bonds to conventional social institutions (employment status, school enrollment, educational attainment, and parenthood), only the coefficient for educational attainment (no high school diploma or equivalency) was signifi- cant (B � .393, p � .01), where the lack of a high school diploma or equivalency at age 19 was associated with an increased likeli- hood of engagement in criminal activities at age 21. The Wald test comparing model fit for nested models, where the model that estimates the four coefficients was compared with the model

Table 2. Legal System Involvement as a Juvenile and Adult Criminal Activities (N � 732)

Variable

Full model

Age 21 Ages 23 or 24

B SE B SE

Legal system involvement �0.009 0.218 0.334 0.252 Delinquency (baseline) 0.145 0.026��� 0.087 0.036�

Prior criminal activities (age 21) 0.163 0.028���

Covariance 3.747 0.458��� 3.747 0.458���

Male 0.889 0.229��� 1.028 0.274���

Race White (ref) Black 0.188 0.159 0.133 0.182 Other 0.144 0.193 0.484 0.210�

Family conflict 0.013 0.165 0.487 0.191�

History of maltreatment Neglect 0.549 0.190�� �0.308 0.198 Physical abuse �0.142 0.158 �0.047 0.186 Sexual abuse 0.457 0.190� �0.146 0.251 Number of placements 0.029 0.01�� 0.024 0.011�

Bonds to society/conventional opportunities Employed 0.064 0.146 �0.275 0.166� Enrolled �0.173 0.153 �0.059 0.199 No HS diploma or equivalency 0.393 0.149�� 0.042 0.185 Parenthood 0.073 0.206 �0.058 0.186

Interaction terms Male � Legal System 0.214 0.278 �0.767 0.310�

Male � Delinquency �0.075 0.029� �0.026 0.039 Dispersion parameter 1.443 0.158��� 1.745 0.205���

Log-likelihood �3752 22 �3582 23 AIC 7,547 7,210 BIC 7,648 7,316 Adjusted BIC 7,578 7,243 Wald test 11.059 4� 3.718 4

Note. HS � high school; AIC � Akaike information criterion; BIC � Bayesian information criterion. � p � .10. � p � .05. �� p � .01. ��� p � .001.

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an Ps

yc ho

lo gi

ca l

A ss

oc ia

tio n

or on

e of

its al

lie d

pu bl

is he

rs .

T hi

s ar

tic le

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

446 LEE, COURTNEY, HARACHI, AND TAJIMA

where they were not estimated, was statistically significant, �2(4) � 11.059, p � .05, indicating improved model fit with the addition of these bonds to social institution variables. In addition, the coefficient for legal system involvement as a juvenile, although not statistically significant, shrinks from 0.043 to �0.009 after adding these four variables.

At ages 23 or 24, criminal activities were related to prior delinquency (reported at baseline) and criminal activities (reported at age 21). In addition, men, as compared with women, and youth in the racial category “other” (Asian, Hispanic, Native, and other), compared with White youth, were associated with a higher likeli- hood of engagement in criminal activities. In addition, family conflict (measured as caregiver domestic violence) and more foster care placements were also associated with an increased likelihood of later engagement in criminal activities. The interaction term between gender and delinquency was not significant, although the interaction term between gender and legal system involvement as a juvenile was statistically significant. The relationship between legal system involvement as a juvenile and later criminal activities was smaller for men (B � �.767, p � .05), although men start at a higher level of overall criminal activities than women as indi- cated by the coefficient for gender (B � 1.03, p � .001). This may suggest that while women may experience a labeling effect, where women with legal system involvement as a juvenile was related to higher levels of criminal activity at age 23 or 24 compared with women without legal system involvement as a juvenile, men experience a deterrence effect where legal system involvement as a juvenile was related to lower levels of criminal activity at age 23 or 24 compared with men without legal system involvement as a juvenile.

Of the four variables measuring bonds to social institutions (employment status, school enrollment, educational attainment, and parenthood), none of the coefficients was statistically signif- icant, although the coefficient for employment was marginally significant (B � �.275, p � .10). The Wald test comparing model

fit between the model that estimates coefficients for these four variables and the model that does not, was not significant, indi- cating that the more parsimonious model, where these coefficients were not estimated, provided comparable fit.

Mediation Models

Figure 1 presents the results for the key constructs in the overall structural model. Legal system involvement as a juvenile was associated with a lower odds (odds ratio [OR] � 0.578) of em- ployment and over twice the likelihood (OR � 2.378) of not having a high school diploma or equivalency by age 19, but does not appear to be associated with school enrollment or parenthood. Self-reported delinquency at baseline did not have a significant relationship with conventional opportunities (education or employ- ment), although delinquency was associated with a higher likeli- hood (OR � 1.113) of parenthood at age 19.

The relationship between not having a high school diploma or equivalency and criminal activities at age 21 continues to be significant in this mediation model, so that an individual without a high school diploma or equivalency was more likely (IRR � 1.488) to report criminal activities at age 21. The coefficients for the other three conventional opportunities/bonds to social institu- tions variables were not significant.

Legal system involvement as a juvenile was associated with a lower likelihood (OR � .554) of being enrolled in school at age 21. In considering the indirect relationship between legal system in- volvement as a juvenile and employment at age 21 via the two educational variables at age 19, school enrollment was associated with a higher likelihood (OR � 1.692) of employment at age 21, and not having a high school diploma or equivalency was associ- ated with a lower likelihood (OR � .385) of employment at age 21.

Looking at the relationship between criminal activities at age 23 or 24 and the four conventional bonds to society variables at age 21, only the coefficient for employment was marginally significant

gA)2W( 91 egA )1W( ecnecselodA/doohdlihC )4W(42/32 egA )3W(12 e

Legal System Involvement as a

Juvenile

Delinquency School Enrollment

Employed

No HS

School Enrollment

Employed

No HS

Criminal Activities

Criminal Activities

Parenthood

Parenthood

-.551

.862

.107

-.596

.526-.954

.145

.087

.163

.393

-.275

1.084 3.881

-.714

Figure 1. Legal system involvement as a juvenile and later criminal activities with mediating pathways. Unstandardized coefficients are presented. Nonsignificant pathways are not depicted.

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an Ps

yc ho

lo gi

ca l

A ss

oc ia

tio n

or on

e of

its al

lie d

pu bl

is he

rs .

T hi

s ar

tic le

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

447LABELING AMONG FORMER FOSTER YOUTH

(IRR � .763). None of the other variables was significantly related to criminal activities at age 23 or 24.

While the indirect pathway from legal system involvement as a juvenile and criminal activities at age 21 through educational attainment was significant (B � .339, p � .05), the indirect pathways through employment and school enrollment were not significant. The direct effect of legal system involvement and criminal activities at age 21 was not significant. At age 23 or 24, neither the indirect pathways through the education and employ- ment variables, nor the direct effect were significant. However, the total effect, including the direct effect and the indirect pathways through education and employment at age 21, as well as through education at age 19 then employment at 21, was significant (B � .625, p � .05).

Discussion To answer the question whether legal system involvement as a

juvenile is associated with adult criminal behaviors, we found that legal system involvement as a juvenile was related to criminal activities at age 21, but the relationship between legal system involvement as a juvenile and criminal activity at age 23 or 24 was not significant when controlling for earlier delinquency. This may suggest that the association between juvenile legal system involve- ment and adult criminal behavior declines over time. As an alter- native, the results from the full model with the complete set of control variables raises the possibility that this lack of significance may reflect opposing processes operating between men and women at age 23 or 24. The significant interaction term between being male and legal system involvement indicates different pro- cesses for men and women, where the coefficient for women was positive while the coefficient for men was negative. The direction of the coefficients suggests that juvenile legal system involvement initiates a labeling effect for women, although this coefficient was not significant, but a deterrence effect for men. The lack of significance for the interaction term at age 21 is curious, although it may suggest that gender differences in the labeling process grow over time. In contrast, the interaction term for gender and delin- quency was significant at age 21, but not age 23 or 24. However, given the age-graded crime curve, where offending declines in early adulthood (Agnew, 2003), perhaps the lack of gender differ- ences in slope at age 23 or 24 reflects a convergence that results from a general process of desistance. More research is needed to test these gender differences in other samples to verify whether there are opposing processes operating, and if so, whether this finding can be generalized to other populations.

To answer the question whether the association between juve- nile legal system involvement and adult criminal behaviors is mediated by bonds to conventional social institutions, this study finds evidence that the labeling effect of legal system involvement as a juvenile on engagement in criminal activities at age 21 is mediated by educational attainment at age 19. In addition, there is limited evidence that the lack of a high school diploma or equiv- alency at age 19 is related to a lower likelihood of employment at age 21. In turn, employment at age 21 is marginally related to a lower likelihood of engagement in criminal activity at age 23 or 24. This process of social exclusion appears to operate through

education early in the transition to adulthood, and may continue to operate through employment later during the transition, similar to findings from previous studies (Bernburg & Krohn, 2003; Lopes et al., 2012).

As other studies have shown, foster youth are more likely to be placed into a detention- or probation-related placement than are nonfoster youth (Ryan et al., 2007). Unfortunately, placement in detention may result in changing schools or interrupted educa- tional experiences that may have already been interrupted by the foster care placement. In fact, placement in detention may also contribute to foster care placement instability, which was consis- tently significantly associated with adult criminal activities at both age 21 and 23 or 24. Disrupted educational experiences may result in the decreased likelihood of completing high school as well as decreased later employment, thus, weakening the individual’s bond to society (Apel & Sweeten, 2010; Hjalmarsson, 2008; Kirk & Sampson, 2013). This suggests that when a juvenile makes any contact with the legal system, greater attention can be paid to ensuring continuity in the youth’s educational experiences when making dispositional decisions to increase the likelihood of earn- ing a high school diploma. Perhaps more of these youth can receive educational and other services tailored to the youth’s specific needs through the foster care system rather than being detained by the juvenile justice system, especially for the girls who may be more likely to experience legal system involvement as a negative label. Collaborative programs between the juvenile jus- tice and child welfare system, such as the Crossover Youth Prac- tice Model (Lutz, Stewart, & Legters, n.d.) should be carefully implemented and evaluated.

While in this study there appears to be some relationship be- tween legal system involvement as a juvenile and later educational and employment outcomes, parenthood does not appear to be related to legal system involvement. Although there was no rela- tionship between parenthood and both legal system involvement and adult criminal behaviors, we included this measure for two reasons. First, this finding is in contrast to recent studies that have found a relationship between parenthood and decreased criminal activities and legal system involvement (Kreager et al., 2010; Lee, Courtney, & Hook, 2012; Lee, Courtney, & Tajima, 2014). Sec- ond, the hypothesized mediating process between legal system involvement and later criminal activities has been described alter- natively as exclusion from conventional opportunities or as weak- ened social bonds. Since there appear to be some relationships between legal system involvement and school and employment, but not parenthood, this may suggest that the mechanism operating is through exclusion from opportunity structures more specifically, and not stigma and weakened bonds in general. At the same time, an early transition into parenthood, as is the case for most of the youth in this study who were between the ages of 17–24 for the duration of the study, may limit gains to human capital and foreclose other opportunities (Berzin & De Marco, 2010; Guldi, Page, & Stevens, 2007). In this sense, a concern with early par- enthood is the possibility that opportunity structures will be blocked for the individual, and thus may not contribute to their overall bond to society. However, our estimated models do not support this possibility, at least for this sample of former foster youth during early adulthood, since there was not a significant pathway from parenthood to later education and employment

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an Ps

yc ho

lo gi

ca l

A ss

oc ia

tio n

or on

e of

its al

lie d

pu bl

is he

rs .

T hi

s ar

tic le

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

448 LEE, COURTNEY, HARACHI, AND TAJIMA

outcomes. The effect of parenthood is likely to change over time and warrants future research.

The mediating models also suggest that legal system involve- ment and delinquency should be treated as two distinct concepts within this sample of foster youth aging out of care. Legal system involvement had some significant relationships with employment and educational opportunities, and thus appears to impact lives by weakening individual attachments to the institutions of the labor market and education. On the other hand, delinquency had a direct relationship with later criminal activities, but no relationship with these conventional opportunity structures. Future work should carefully consider this distinction when framing research questions and choosing which measures to include. In particular, this dis- tinction between the two concepts should be considered when using administrative data, since measures of legal system involve- ment and delinquent behavior do not appear to be equivalent. Future research might also explore whether this distinction holds in other populations.

There are some limitations to this study. Our measures of conventional social institutions were limited in that some were point in time measures. School enrollment and employment status were measured at the time of interview, but approximately 2 years elapsed between interviews. Furthermore, these measures did not take into account the quality of those bonds, which may have resulted in more robust findings. For example, some participants with nonresident children may be strongly bonded to their child even if they do not live together; a measure of the participant’s commitment, attachment, and involvement with their child(ren), regardless of residential status, would have provided a more ac- curate measure of the youth’s social bond. Similarly, our measure of legal system involvement could have been enhanced with ad- ditional information about the frequency and duration of their involvement with the legal system, as this would have indicated the strength of the labeling event.

Conclusion This study finds evidence to support the labeling theory: label-

ing in the form of juvenile justice involvement was associated with higher rates of engagement in adult criminal behaviors. Moreover, labeling appears to contribute to a process of social exclusion involving limited educational and employment outcomes. The ability of these foster youth to successfully adopt roles of adult- hood is dependent on their ability to establish commitments to institutions of adulthood (Lee, 2014; Sampson & Laub, 1990). Research has demonstrated that the transition to adulthood can be facilitated by the foster care system in the form of extended care services, in increasing educational attainment, increasing earnings, and reducing the likelihood of adult arrest (Courtney, Dworsky, & Pollack, 2007; Hook & Courtney, 2011; Lee, Courtney, & Tajima, 2014). This transition process is difficult for former foster youth, many of whom lack the supports of family and postsecondary education during the transition period, but it appears that if they have experienced legal system involvement as a juvenile, then their ability to transition into adulthood may become more difficult and complicated by their involvement in two systems.

Keywords: labeling theory; social exclusion; aging out of foster care; transition to adulthood; crime and delinquency

References Agnew, R. (2003). An integrated theory of the adolescent peak in offend-

ing. Youth & Society, 34, 263–299. http://dx.doi.org/10.1177/ 0044118X02250094

Agresti, A. (2002). Categorical data analysis. New York, NY: Wiley- Interscience. http://dx.doi.org/10.1002/0471249688

Alexander, M. (2011). The new Jim Crow. Ohio State Journal of Criminal Law, 9, 7–26.

Apel, R., & Sweeten, G. (2010). The impact of incarceration on employ- ment during the transition to adulthood. Social Problems, 57, 448–479. http://dx.doi.org/10.1525/sp.2010.57.3.448

Arnett, J. J. (2006). Emerging adulthood: Understanding the new way of coming of age. In J. J. Arnett & J. L. Tanner (Eds.), Emerging adults in America: Coming of age in the 21st century (pp. 3–19). Washington, DC: American Psychological Association. Retrieved from http://a1 .phobos.apple.com/us/r30/CobaltPublic/v4/e8/17/bc/e817bc94-cf40-5cf2- b7a0-0ffa5bbf8e96/108-2370347507287973706-emerging_adulthood .pdf http://dx.doi.org/10.1037/11381-001

Axford, N. (2010). Is social exclusion a useful concept in children’s services? British Journal of Social Work, 40, 737–754. http://dx.doi.org/ 10.1093/bjsw/bcn121

Becker, H. S. (1991). Outsiders: Studies in the sociology of deviance. New York, NY: The Free Press. (Original work published 1963)

Bernburg, J. G. (2009). Labeling theory. In M. D. Krohn, A. J. Lizotte, & G. P. Hall (Eds.), Handbook on crime and deviance (pp. 187–207). New York, NY: Springer. http://dx.doi.org/10.1007/978-1-4419-0245-0_10

Bernburg, J. G., & Krohn, M. D. (2003). Labeling, life chances, and adult crime: The direct and indirect effects of official intervention in adoles- cence on crime in early adulthood. Criminology, 41, 1287–1318. http:// dx.doi.org/10.1111/j.1745-9125.2003.tb01020.x

Bernburg, J. G., Krohn, M. D., & Rivera, C. J. (2006). Official labeling, criminal embeddedness, and subsequent delinquency a longitudinal test of labeling theory. Journal of Research in Crime and Delinquency, 43, 67–88. http://dx.doi.org/10.1177/0022427805280068

Berzin, S. C. (2008). Difficulties in the transition to adulthood: Using propensity scoring to understand what makes foster youth vulnerable. Social Service Review, 82, 171–196. http://dx.doi.org/10.1086/588 417

Berzin, S. C., & De Marco, A. C. (2010). Understanding the impact of poverty on critical events in emerging adulthood. Youth & Society, 42, 278–300. http://dx.doi.org/10.1177/0044118X09351909

Blomberg, T. G., Bales, W. D., Mann, K., Piquero, A. R., & Berk, R. A. (2011). Incarceration, education and transition from delinquency. Jour- nal of Criminal Justice, 39, 355–365. http://dx.doi.org/10.1016/j.jcrimjus .2011.04.003

Children’s Bureau. (2014). The AFCARS report: Preliminary estimates FY 2013 as of June 2014 (No. 21). Washington, DC: U.S. Department of Health and Human Services, Administration for Children, Youth and Families.

Courtney, M. E. (2009). The difficult transition to adulthood for foster youth in the US: Implications for the state as corporate parent. Social Policy Report, 23, 3–18.

Courtney, M. E., Dworsky, A., Brown, A., Cary, C., Love, K., & Vorhies, V. (2011). Midwest evaluation of the adult functioning of former foster youth: Outcomes at age 26. Chicago, IL: Chapin Hall Center for Chil- dren at the University of Chicago.

Courtney, M. E., Dworsky, A., & Pollack, H. (2007). When should the state cease parenting? Evidence from the Midwest Study. Chicago, IL: Chapin Hall Center for Children at the University of Chicago. Retrieved from http://jimcaseyyouth.org/sites/default/files/documents/When%20 Should%20the%20State%20Cease%20Parenting.chall_.pdf

Crutchfield, R. D., & Pitchford, S. R. (1997). Work and crime: The effects of labor stratification. Social Forces, 76, 93–118. http://dx.doi.org/ 10.1093/sf/76.1.93

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an Ps

yc ho

lo gi

ca l

A ss

oc ia

tio n

or on

e of

its al

lie d

pu bl

is he

rs .

T hi

s ar

tic le

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

449LABELING AMONG FORMER FOSTER YOUTH

Cusick, G. R., Courtney, M. E., Havlicek, J., & Hess, N. (2011). Crime during the transition to adulthood: How youth fare as they leave out-of-home care. Chicago, IL: Chapin Hall Center for Children at the University of Chicago. Retrieved from http://www.chapinhall.org/sites/ default/files/Crime%20During%20Transition_03_16_11.pdf

Fader, J. J., Harris, P. W., Jones, P. R., & Poulin, M. E. (2001). Factors involved in decisions on commitment to delinquency programs for first-time juvenile offenders. Justice Quarterly, 18, 323–341. http://dx .doi.org/10.1080/07418820100094921

Feld, B. C. (1999). Bad kids: Race and the transformation of the juvenile court. Studies in crime and public policy. Retrieved from http://eric.ed .gov/?id�ED434977.

Gatti, U., Tremblay, R. E., & Vitaro, F. (2009). Iatrogenic effect of juvenile justice. Journal of Child Psychology and Psychiatry, 50, 991–998. http://dx.doi.org/10.1111/j.1469-7610.2008.02057.x

Gebo, E. (2005). Do family courts administer individualized justice in delinquency cases? Criminal Justice Policy Review, 16, 190–210. http:// dx.doi.org/10.1177/0887403404267773

Guldi, M., Page, M. E., & Stevens, H. (2007). Family background and children’s transitions to adulthood over time. In S. Danziger & C. E. Rouse (Eds.), The price of independence: The economics of early adulthood (pp. 261–277). New York, NY: Russell Sage Foundation.

Halemba, G. J., Siegel, G. C., Lord, R. D., & Zawacki, S. (2004). Arizona Dual Jurisdiction Study. Retrieved from http://www.cwla.org/programs/ juvenilejustice/AZDualJurStudy.pdf

Herz, D. C., Ryan, J. P., & Bilchik, S. (2010). Challenges facing crossover youth: An examination of juvenile-justice decision making and recidi- vism. Family Court Review, 48, 305–321. http://dx.doi.org/10.1111/j .1744-1617.2010.01312.x

Hjalmarsson, R. (2008). Criminal justice involvement and high school completion. Journal of Urban Economics, 63, 613–630. http://dx.doi .org/10.1016/j.jue.2007.04.003

Hook, J. L., & Courtney, M. E. (2011). Employment outcomes of former foster youth as young adults: The importance of human, personal, and social capital. Children and Youth Services Review, 33, 1855–1865. http://dx.doi.org/10.1016/j.childyouth.2011.05.004

Huizinga, D., Schumann, K., Ehret, B., & Elliott, A. (2004). The effect of juvenile justice system processing on subsequent delinquent and crimi- nal behavior: A cross-national study. Washington, DC: United States Department of Justice. Retrieved from http://enj.org/portal/biblioteca/ penal/penal_juvenil/44.pdf

Jackson, D. B., & Hay, C. (2013). The conditional impact of official labeling on subsequent delinquency: Considering the attenuating role of family attachment. Journal of Research in Crime and Delinquency, 50, 300–322. http://dx.doi.org/10.1177/0022427812444931

Jonson-Reid, M. (2002). Exploring the relationship between child welfare intervention and juvenile corrections involvement. American Journal of Orthopsychiatry, 72, 559–576. http://dx.doi.org/10.1037/0002-9432.72 .4.559

Jonson-Reid, M., & Barth, R. P. (2000). From placement to prison: The path to adolescent incarceration from child welfare supervised foster or group care. Children and Youth Services Review, 22, 493–516. http:// dx.doi.org/10.1016/S0190-7409(00)00100-6

Kirk, D. S., & Sampson, R. J. (2013). Juvenile arrest and collateral educational damage in the transition to adulthood. Sociology of Educa- tion, 86, 36–62. http://dx.doi.org/10.1177/0038040712448862

Kreager, D. A., Matsueda, R. L., & Erosheva, E. A. (2010). Motherhood and criminal desistance in disadvantaged neighborhoods. Criminology: An Interdisciplinary Journal, 48, 221–258. http://dx.doi.org/10.1111/j .1745-9125.2010.00184.x

Lee, J. S. (2014). An institutional framework for the study of the transition to adulthood. Youth & Society, 46, 706–730. http://dx.doi.org/10.1177/ 0044118X12450643

Lee, J. S., Courtney, M. E., & Hook, J. L. (2012). Formal bonds during the transition to adulthood: Extended foster care support and criminal/legal involvement. Journal of Public Child Welfare, 6, 255–279. http://dx.doi .org/10.1080/15548732.2012.683336

Lee, J. S., Courtney, M. E., & Tajima, E. (2014). Extended foster care support during the transition to adulthood: Effect on the risk of arrest. Children and Youth Services Review, 42, 34–42. http://dx.doi.org/ 10.1016/j.childyouth.2014.03.018

Lemert, E. M. (1951). Social pathology: A systematic approach to the theory of sociopathic behavior. New York, NY: McGraw-Hill.

Long, J. S. (1997). Regression models for categorical and limited depen- dent variables. Atlanta, GA: Sage.

Lopes, G., Krohn, M. D., Lizotte, A. J., Schmidt, N. M., Vasquez, B. E., & Bernburg, J. G. (2012). Labeling and cumulative disadvantage: The impact of formal police intervention on life chances and crime during emerging adulthood. Crime & Delinquency, 58, 456–488. http://dx.doi .org/10.1177/0011128712436414

Lutz, L., Stewart, M., Legters, L., & Herz, D. (n.d.). Crossover youth practice model. Washington, DC: Center for Juvenile Justice Reform.

Maschi, T., Hatcher, S. S., Schwalbe, C. S., & Rosato, N. S. (2008). Mapping the social service pathways of youth to and through the juvenile justice system: A comprehensive review. Children and Youth Services Review, 30, 1376–1385. http://dx.doi.org/10.1016/j.childyouth .2008.04.006

Maxfield, M. G., & Widom, C. S. (1996). The cycle of violence: Revisited 6 years later. Archives of Pediatrics & Adolescent Medicine, 150, 390 –395. http://dx.doi.org/10.1001/archpedi.1996.021702900 56009

McAra, L., & McVie, S. (2007). Youth justice? The impact of system contact on patterns of desistance from offending. European Journal of Criminology, 4, 315–345. http://dx.doi.org/10.1177/1477370807077186

Muthén, L. K., & Muthén, B. O. (1998). Mplus user’s guide (6th ed.). Los Angeles, CA: Author.

Paternoster, R., & Iovanni, L. (1989). The labeling perspective and delin- quency: An elaboration of the theory and an assessment of the evidence. Justice Quarterly, 6, 359–394. http://dx.doi.org/10.1080/074188289000 90261

Pecora, P. J., Kessler, R. C., O’Brien, K., White, C. R., Williams, J., Hiripi, E., . . . Herrick, M. A. (2006). Educational and employment outcomes of adults formerly placed in foster care: Results from the Northwest Foster Care Alumni Study. Children and Youth Services Review, 28, 1459– 1481. http://dx.doi.org/10.1016/j.childyouth.2006.04.003

Reilly, T. (2003). Transition from care: Status and outcomes of youth who age out of foster care. Child Welfare: Journal of Policy, Practice, and Program, 82, 727–746.

Restivo, E., & Lanier, M. M. (2015). Measuring the contextual effects and mitigating factors of labeling theory. Justice Quarterly, 32, 116–141.

Rose, D. T., Abramson, L. Y., & Kaupie, C. A. (2000). The Lifetime Experiences Questionnaire: A measure of history of emotional, physical, and sexual maltreatment. Manuscript in preparation.

Ross, T., Conger, D., & Armstrong, M. (2002). Bridging child welfare and juvenile justice: Preventing unnecessary detention of foster children. Child Welfare, 81, 471–494.

Ryan, J. P., Herz, D., Hernandez, P. M., & Marshall, J. M. (2007). Maltreatment and delinquency: Investigating child welfare bias in juve- nile justice processing. Children and Youth Services Review, 29, 1035– 1050. http://dx.doi.org/10.1016/j.childyouth.2007.04.002

Ryan, J. P., Marshall, J. M., Herz, D., & Hernandez, P. M. (2008). Juvenile delinquency in child welfare: Investigating group home effects. Chil- dren and Youth Services Review, 30, 1088–1099. http://dx.doi.org/ 10.1016/j.childyouth.2008.02.004

Ryan, J. P., & Testa, M. F. (2005). Child maltreatment and juvenile delinquency: Investigating the role of placement and placement insta-

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an Ps

yc ho

lo gi

ca l

A ss

oc ia

tio n

or on

e of

its al

lie d

pu bl

is he

rs .

T hi

s ar

tic le

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

450 LEE, COURTNEY, HARACHI, AND TAJIMA

bility. Children and Youth Services Review, 27, 227–249. http://dx.doi .org/10.1016/j.childyouth.2004.05.007

Sampson, R. J. (1986). Crime in cities: The effects of formal and informal social control. Crime and Justice, 8, 271–311. http://dx.doi.org/10.1086/ 449125

Sampson, R. J., & Laub, J. H. (1990). Crime and deviance over the life course: The salience of adult social bonds. American Sociological Re- view, 55, 609–627. http://dx.doi.org/10.2307/2095859

Sampson, R. J., & Laub, J. H. (1997). A life-course theory of cumulative disadvantage and the stability of delinquency. Developmental Theories of Crime and Delinquency, 7, 133–161.

Smith, C., & Thornberry, T. P. (1995). The relationship between childhood maltreatment and adolescent involvement in delinquency. Criminology, 33, 451–481. http://dx.doi.org/10.1111/j.1745-9125.1995 .tb01186.x

Southerland, D., Casanueva, C. E., & Ringeisen, H. (2009). Young adult outcomes and mental health problems among transition age youth in- vestigated for maltreatment during adolescence. Children and Youth Services Review, 31, 947–956. http://dx.doi.org/10.1016/j.childyouth .2009.03.010

Uggen, C., & Kruttschnitt, C. (1998). Crime in the breaking: Gender differences in desistance. Law & Society Review, 32, 339–366. http:// dx.doi.org/10.2307/827766

Vaughn, M. G., Shook, J. J., & McMillen, J. C. (2008). Aging out of foster care and legal involvement: Toward a typology of risk. Social Service Review, 82, 419–446. http://dx.doi.org/10.1086/592535

Yampolskaya, S., Armstrong, M. I., & McNeish, R. (2011). Children placed in out-of-home care: Risk factors for involvement with the juvenile justice system. Violence and Victims, 26, 231–245. http://dx.doi .org/10.1891/0886-6708.26.2.231

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an Ps

yc ho

lo gi

ca l

A ss

oc ia

tio n

or on

e of

its al

lie d

pu bl

is he

rs .

T hi

s ar

tic le

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

451LABELING AMONG FORMER FOSTER YOUTH