Skip to main content
PMC home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2014 Aug 12.
Published in final edited form as: J Interpers Violence. 2011 Nov 11;27(7):1199–1224. doi: 10.1177/0886260511425240

The Mediating Role of Self-Regulation Between Intrafamilial Violence and Mental Health Adjustment in Incarcerated Male Adolescents

Suzanne C Perkins 1, Kai S Cortina 1, Joanne Smith-Darden 1, Sandra Graham-Bermann 1
PMCID: PMC4129515  NIHMSID: NIHMS594889  PMID: 22080578

Abstract

Background

This paper investigates the relation between history of intrafamilial violence and self-regulatory capacity, cognitive processing and mental health adjustment in incarcerated adolescents. Adolescents were incarcerated at the time of the study for various violent offenses, ranging from persistent delinquency to sexual assault (n = 115). A model is proposed that posits that self-regulation, cognitive ability and cognitive processing are integral to the relation between intrafamilial violence and mental health function. The primary hypothesis of the study tests this mediation model. The relations between mental health, cognitive processing, self-regulation and intrafamilial violence are also examined.

Methods

The study was conducted during two sessions at a juvenile facility in the Midwest using survey measures, academic and intelligence testing and cognitive tasks.

Results

Youth were between the ages of 13 and 20. Approximately 70% were previously diagnosed with a disability. Significant Pearson’s correlations were found between seven out of eight mental health subscales of the Youth Self-Report (YSR) and intrafamilial violence history. Structural equation modeling was used to examine the role of cognitive processing in the association between intrafamilial violence and mental health function. Non-verbal or performance deficits, a significant difference between verbal skills and non-verbal skills, were related to intrafamilial violence. Self-regulation partially mediated the relation between intrafamilial violence and mental health function.

Conclusions

Self-regulation ability may be compromised by intrafamilial violence and be a precursor to both internalizing and externalizing mental health problem in incarcerated youth. Educational, clinical and research implications are discussed.

Introduction

Intrafamilial violence is often a precursor to later mental health problems in children (Higgins, 2004), including both internalizing problems, like depression and anxiety, and externalizing problems, like aggression and delinquency. The mechanism that leads from intrafamilial violence to later internalizing and/or externalizing mental health problems remains unclear. Self-regulation and cognitive processing deficits are related to both intrafamilial violence and mental health problems and may mediate the relation between those two constructs (Buckner, Mezzacappa, & Beardslee, 2009; Jonson-Reid, et al, 2004).

Cognitive processing is a broad term used to describe the ability to manage information using capacities including intelligence, planning, processing speed, memory, attention and executive control, including self-regulation. Self-regulation is an essential underlying component to all cognitive processing, necessary for the down-regulation of emotion and up-regulation cognitive skills necessary for the completion of academic or social goals (Efklides, 2008). Self-regulatory deficits are demonstrated in children exposed to violence (Dodge & Crick, 1990; Merwe & Dawes, 2000), which suggests that one possible contributing factor to the development of mental health problems after childhood exposure to violence is deficit in self-regulation.

Intrafamilial violence, Mental Health and Achievement

Multiple studies have documented the negative influence of intrafamilial violence on subsequent internalizing, externalizing and academic achievement. Although many studies examine internalizing and externalizing mental health separately, the overlap in these concepts leads us to take a more holistic approach to mental health function. Overall levels of intrafamilial violence{Margolin:2010jt}, Hanson et al., 2008), exposure to domestic violence (Graham-Bermann, Gruber, Girz, & Howell, 2009), maltreatment (Hosser, Raddatz, & Windzio, 2007; Haapasalo, & Virtanen, 1999), sexual violence (Reed, et al., 2009), and community violence (Ruchkin, et al, 2007) are all related to deficits in mental health functioning. Exposure to family violence is also related to cognitive capacity including, lower achievement scores (Thompson & Wyatt, 1999) and verbal and full-scale intelligence (Graham-Bermann, Howell, Miller, Kwek, & Lilly, 2010).

Mechanisms of Cognitive and Self-Regulation Mediation

Criminal justice theory focuses on self-control as a factor in the development of criminal behavior (Gottfredson & Hirschi, 1990). Self-control theory posits that low self-control is a mediator between various ecological factors, including schooling and parenting, and later criminal behavior (Gottfredson & Hirschi, 1990). Low self-control is defined in the literature as a range of both psychological processes and behaviors and researchers postulate that self-control is fixed by the age of about 10 (Winfree, Taylor & Esbensen, 2006). In contrast, Bandura’s social cognitive theory (1986) defines self-regulation as involving self-observation and behavioral monitoring, and as a component of cognitive control, which is neuroanatomically centered in the prefrontal cortex; self-regulation is known to develop well into the 20’s (Casey, Janes & Hare, 2008).

In schools, self-regulation is necessary for self-judgment of academic progress and reaction to performance outcomes (Bernier, Carlson, & Whipple, 2010; Buckner, Mezzacappa, & Beardslee, 2009; Mischel et al., 2010). Self-regulation is a broad term used to encompass a variety of cognitive volitional activities in humans and is dependent on a complex neurocognitive network (Mischel et al., 2010). The development of self-regulation is dependent on early childhood experiences within the family (Bernier, Carlson, & Whipple, 2010) and is influenced throughout the lifespan by experience, including schooling, and can be improved through training (Tang et al., 2007). The ability to self-regulate emotional response during times of high cognitive load is central to flexible responding during learning (Eccles & Wigfield, 2002). Self-regulated learners use active learning and possess a belief that they can perform efficaciously and set achievable academic goals (Eccles & Wigfield, 2002). In this sense, self-regulation is a mechanism can be seen as an extension of self-control theory that links ecological factors with psychological processes.

Miller and Brickman (2004) suggest a component of past experience in the development of future oriented self-regulation. Past experience, success or failure in learning situations leads to value for the task and knowledge of the possibility for success. Taken more broadly, past experience of abuse and other negative early life stress lead to a diminished capacity for self-regulation (Merwe & Dawes, 2000). According to social information processing theory, children with violence histories and experience with harsh parenting are biased toward threatening stimuli and interpret neutral social cues as negative or threatening (Dodge & Crick, 1990). This heightened threat appraisal is related to avoidant coping, hostility, depression, anxiety and negative affect, suggesting that the inability to self-regulate negative associations and the subsequent use of avoidant coping may mediate the relation between intrafamilial violence and later mental health function. Self-regulation is a critical component of optimal mental health development and history of intrafamilial violence has a significant negative influence on the ability to develop self-regulatory skills.

A higher rate of both intrafamilial and community violence exposure has been found in both adult and adolescent offenders. In one recent study of adults, using a self-reported measure, sexual abuse by a family member was experienced by 18% of inmate and physical abuse by fifty percent (Carlson & Shafer, 2010). In incarcerated adolescents, 14–19% have substantiated histories of maltreatment (Jonson-Reid & Way, 2001), which suggests that self-report rates would be higher. Intrafamilial violence is a risk factor for later criminal behavior. For adults, having been touched sexually prior to puberty results in a 7–10% increase in the probability of being incarcerated later in life (Curtis, Leung, Sullivan, Eschbach, & Stinson, 2001). Incarcerated populations also exhibit deficits in self-regulation (Ross & Fontao, 2007). Deficits in self-regulation and any of the broad areas of cognition often result in developmental problems collectively called disabilities.

Terms, like learning disability, encompass multiple facets of cognitive processing, such as intelligence, attention and processing speed. Approximately 13.4% of the general school age population is served under the Individuals with Disabilities Education Act (IDEA) (Department of Education, 2009). Among juvenile offenders national rates of disabilities are between 33% and 45% (Krezmien, Mulcahy, & Leone, 2008; Quinn, Rutherford, Leone, Osher, & Poirier, 2005). Deficits in self-regulation may be related to the high prevalence of learning and behavioral disabilities in incarcerated populations (Quinn, Rutherford, Leone, Osher, & Poirier, 2005). Direct relations between violence and mental health problems (Hosser, Raddatz, & Windzio, 2007,{Margolin:2010jt}, violence and cognitive processing (Graham-Bermann, Howell, Miller, Kwek, & Lilly, 2010), and cognitive processing and mental health suggests that cognitive processing deficits may be an underlying mechanism in the relation between intrafamilial violence and mental health.

In the present study, we conceptualize self-regulatory capacity as a developmental construct and measure the current level of functioning within a framework of developmental psychopathology rather than as a fixed capacity. We believe that differences in current level of cognitive processing ability (including self-regulation) may mediate the relation between past history of intrafamilial violence and current level of mental health functioning in an incarcerated population. Community violence and the cumulative exposure to different forms of violence also have deleterious effects (Margolin, Vickerman, Oliver, & Gordis, 2010). In the current study we focus on intrafamilial violence to isolate the effect of intrafamilial violence from the very high rates of community violence exposure in this population. The high rates of intrafamilial violence, cognitive processing deficits and mental health problems in this subpopulation make this group especially appropriate for examining these complex interactions. The current study aims to further our understanding of the interdependence between these constructs in a population of violent and non-violent youth offenders. Specifically we postulate that cognitive processing deficits, in general, and self-regulation, in particular, may mediate the relation between intrafamilial violence and mental health problems. It is first hypothesized that incarcerated adolescents will show a direct relation between intrafamilial violence and mental health functioning. In addition, is hypothesized that this relation will be partially mediated by cognitive processing, learning or self-regulatory problems.

Method

Design

The study was conducted at a youth incarceration facility in Michigan and included survey data, individual cognitive and neuropsychological assessment and educational records review. Permission to conduct the study was obtained from the University of Michigan Institutional Review Board and the State of Michigan Office of Human Services. A Certificate of Confidentiality was obtained from the National Institutes of Health.

Consent was obtained for the study from parents or the court (for wards of the state) before asking the minors for assent to participate. Letters were sent to parents with consent documents attached. If there was no response to a second letter and the juveniles were covered under Delinquency Act 150, we asked that the State Department of Human Services give permission for the youth the participate in the study. Legal adults (those youth over 18 years old) were asked to sign participant consent. After parent/guardian, court or Human Services permission was granted, each youth was contacted individually to obtain assent. Researchers explained that the study was voluntary and that no consequence positive or negative would result from participation or refusal to participate. If necessary, assistance with reading the surveys was provided. Youth were not compensated directly for participation in the study. Instead, a donation was made to the facility for purchase of recreational equipment.

Participants

The sample consisted of 115 males that ranged in age from 13 to 20 with a mean age of 17 (SD=1.304). Ethnicities were approximately 1/3 African-American, 1/3 Caucasian and 1/3 self-defined as mixed ethnicity. Grade levels were between 7th and the first year of college (See Tables 1) Crimes committed prior to the study ranged in seriousness and in type (see Table 2).

Table 1.

Sample Family and Demographic Characteristics

Family of Origin N Percent Race/ethnicity N Percent
Two parents 37 36.6 Caucasian 34 39.5
Mom headed household 39 38.6 African American 30 34.9
Dad headed household 7 6.9 Multiracial 19 22.1
Relative 15 14.9 Hispanic, Asian or other 3 3.5
Foster care 3 3.0
Age N Percent Last Grade Completed N Percent
13–14 2 1.8 7th grade 3 2.8
15 5 4.5 8th grade 15 14.2
16 23 20.9 9th grade 10 9.4
17 33 30.0 10th grade 18 17.0
18 31 28.2 11th grade 28 26.4
19 10 9.1 12th grade 27 25.5
20 6 5.5 1 year college 5 4.7
Open in a new tab

Table 2.

Percent of Sample Offense Type and Class

N Percent
Violent class 1 8 9.0
Violent class 2 3 3.4
Violent class 3 13 14.6
Violent class 4 8 9.0
Sexual class 1 17 19.1
Sexual class 2 30 33.7
Sexual class 3 2 2.2
Property class 1 7 7.9
Property class 3 1 1.1
Total 89 100.0
Open in a new tab

Note: There were no property class 2 violations.

Measures

Intrafamilial violence was measured using an adapted version of the Conflict Tactics Scale (CTS) (Straus & Gelles, 1990) to assess how conflicts in the family are handled on a 1 to 5 Likert Scale where 1 is equal to “never” and 5 equals “every day.” Internal consistency of the CTS was determined as part of the National Family Violence Survey (n = 2143). This measure has been widely used in incarcerated populations (Reed, et al., 2009; Hosser, Raddatz, & Windzio, 2007; Haapasalo, & Virtanen, 1999). The CTS was amended to include violence between siblings. Participants were asked to state how often they experiences each of 18 conflict tactics during two time frames; from 6–12 and from 13–18. The 18 items were repeated within four familial relationship types; parents to participant, parent to parent, sibling to participant and participant to sibling. Subjects were instructed to include any woman that raised them (e.g. grandmothers or foster mothers) as mothers, and any significant male (e.g. boyfriend or foster father) as father. If subjects did not have experience with both figures, they were instructed to leave blank the corresponding section.

Although there is likely overlap between sibling to participant and participant we did not include participant to sibling in the present analysis as it most closely represents violence perpetration by the subject rather then experienced by the subject. There existed a high correlation between these subscales. The intrafamilial violence scale used in this study is a sum of total intrafamilial violence at both developmental stages and is defined as total violence from siblings, parents and intimate partner violence (IPV). Cronbach’s alphas were between .923 and .945.

Mental Health was assessed with the 118-item Youth Self-Report (YSR) (Achenbach, 1991) and with the Beck Depression Inventory (BDI). Three subscales were created from the BDI according to the factors derived by Shafer (2006). General Depression (GD), Performance Impairment (PI) and Somatic Complaints (SC). Reliability for the first two scales was moderate but was low for the third factor (GD - α = .79; PI - α = .73; SC - α = .18). Eight subscales of the YSR were derived according to the manual. In this study, reliability of the eight subscales, Withdrawn, Somatic Complaints, Anxious/Depressed, Mood Disorders, Thought Disorders, Attention Problems, Delinquency and Aggression, range from α = .71 to .97.

A computerized battery included the Wisconsin Card Sort (WCST), a test of strategic planning and cognitive set shifting (Kongs, Thompson, Iverson, & Heaton, 2000) and five cognitive processing tasks: Go – No Go (G-NG), Task Switching (TS) (Rogers & Monsell, 1995), Flanker Shape (FS), Shape Matching (SM) and Anti-Saccade Arrows (ASA) (Friedman & Miyake, 2004). The response and reaction time (RT) were recorded. Average RTs and percent correct were calculated for each condition. G-NG and ASA are response inhibition tasks. ASA requires oculomotor inhibition. TS required the participant to switch mental sets between two different tasks. The FS and SM tasks require resistance to distracter interference.

Self-regulation was measured using the Behavior Rating Inventory of Executive Function (BRIEF) (Guy, Isquith, & Gioia, 2004). In this study, reliability of the eight subscales, Withdrawn, Somatic Complaints, Anxious/Depressed, Mood Disorders, Thought Disorders, Attention Problems, Delinquency and Aggression, range from α = .71 to .97.

Verbal and non-verbal (or performance) ability were assessed using the Kaufman Brief Intelligence Test Second Edition (K-Bit-2) (Kaufman & Kaufman, 2004). Reading, spelling and arithmetic achievement were assess with the Wide Range Achievement Test (WRAT-3) (Wilkinson, 1993).

Analyses

Missing Data

Thirty-eight participants had missing data on some of the measures. Participants with missing data were compared to participants with complete data across demographic and other key variables. Group differences were not significant. To avoid excessive missing data, means were imputed using an iterative procedure. Missing data were imputed using parameter estimates that were obtained by performing a linear regression of main variables with the variable of interest as the outcome measure. Data had to be present in two thirds of other variables in order to be imputed. These parameter estimates were then used to impute missing data. For each iteration a new regression was run to obtain new parameter estimates using imputed data.

Structural Equation Modeling

Structural Equation Modeling (SEM) is similar to multiple regression but takes into account latent constructs measured by multiple indicators in order to model interaction, measurement error, correlation of error and interactions. In the current study a model specification process was used to determine mediation (Kenny, 2006). In order to test the hypotheses above, a model was tested positing that cognitive and learning problems mediate or partially mediate the known relation between experience of family violence and subsequent mental health function. In order to assess the adequacy of power to detect differences in with or sample size and the number of variables in the SEM, we conducted a power analysis. With 230 df and a sample size of 115, in order to reject a model with an RMSEA of .10 at a significance level of .05 we have power of 0.9999912 (Preacher & Coffman, 2006).

Results

Descriptive Statistics

Sample Characteristics

There was a high prevalence of both mental health problems and out-of-home placement in this sample (See Table 3). From the educational records, it was found that 70% had a current Individual Education Plan (IEP) with a lifetime incidence of special education diagnosis of 75.5%. Attention deficit disorder was the most common diagnosis affecting 47% of the population, followed by cognitive impairment and speech problems (approximately 12%).

Table 3.

Percent of Sample in Prior Treatment

Yes No
N Percent N Percent
Foster care 38 37.6 63 62.4
Group home 18 18.0 82 82.0
Medicine prescribed by a psychiatrist 50 49.5 51 50.5
Sex offender treatment 39 38.6 62 61.4
Psychological Treatment 87 85.3 15 14.7
Substance Abuse Treatment 15 15.0 85 85.0
Open in a new tab

Intrafamilial violence

Mean levels of intrafamilial violence from the CTS ranged from 1.64 to 1.95 (.74–.81 SD) on the four subscales, which is equivalent to a mean level of less than once a year. Approximately 25% of the participants reported IPV of at least a couple of times a year. Thirty-four percent of the sample reported mean maltreatment by parents at least a couple of times a year or more and 46% reported this level of maltreatment by their sibling. Intrafamilial violence was negatively related to non-verbal (or performance) intelligence and positively related to three subscales of the BDI and the withdrawn subscale of the YSR (see Table 5).

Table 5.

Intercorrelations Between Violence, Cognitive Processing and Mental Health Subscales Controlling for Age and Psychiatric Treatment

VE Verbal NV Read Spell Arith PS1 PS2 TSW FRT SMRT AS
Violence 1.000
Verbal IQ .070 1.000
Nonverbal IQ −.298** .534*** 1.000
Reading −.046 .608*** .396*** 1.000
Spelling −.038 .478*** .318** .758*** 1.000
Arithmetic −.055 .633*** .555*** .480*** .495*** 1.000
WCST 1 −.044 .445*** .314** .273* .208 .405*** 1.000
WCST 2 −.074 .468*** .433*** .181 .131 .488*** .793*** 1.000
Task switch .117 .187 .183 .086 .033 .231* .207 .249* 1.000
Flanker RT .151 .234 .096 .208 .258* .157 .194 .162 −.023 1.000
Match RT −.092 .054 .060 .084 .060 .050 .084 .154 .022 .337** 1.000
Anti-saccade −.039 .341** .230 .258* .408** .472*** .330** .271* .086 .173 .149 1.000
Go No-Go −.044 .335** .368** .152 .260* .264* .159 .218 .041 .193 .068 .369**
Behavior .064 .005 −.102 −.071 −.076 −.029 −.071 −.108 .100 −.148 −.087 −.156
Metacog .139 .058 −.184 −.106 −.207 −.127 −.050 −.120 .094 −.065 .035 −.207
Depression .244** −.072 .006 −.026 .079 .002 −.111 −.122 .097 .115 .146 −.080
Performance .218* −.144 −.129 −.055 .005 −.026 −.063 −.076 .048 .025 .111 −.092
Somatic BDI .276** .178 −.034 .164 .026 −.030 −.028 −.022 .039 −.025 .121 −.020
Withdrawn .219* .042 −.006 −.086 −.046 −.001 .082 .044 .135 .008 .097 .082
Somatic YSR .067 .027 −.105 .016 −.003 −.057 .042 .030 .069 .178 .020 .073
Anxious .131 −.042 −.047 −.034 .119 −.053 −.075 −.117 .092 .066 .049 .051
Social prob .181 .085 −.052 −.046 −.015 −.056 −.026 −.033 .079 −.017 .017 .001
Delinquent .035 −.070 −.228 −.172 −.119 −.144 .088 .042 .039 .068 .089 −.056
Aggressive .118 .151 −.069 .073 .102 −.018 .075 .003 .011 −.002 .033 −.010
GNG BI MRI GD PI SB WD SY AD SP DEL AG
Go No-Go 1.000
Behavior −.082 1.000
Metacog −.163 .753*** 1.000
Depression .011 .293** .371*** 1.000
Performance −.095 .375*** .447*** .724*** 1.000
Somatic BDI −.195 .258* .261* .305** .372*** 1.000
Withdrawn .032 .468*** .584*** .411*** .489*** .343** 1.000
Somatic YSR −.025 .441*** .498*** .237* .413*** .387*** .610*** 1.000
Anxious −.013 .551*** .576*** .453*** .471*** .337** .789*** .625*** 1.000
Social prob −.138 .487*** .634*** .408*** .451*** .370*** .743*** .604*** .760*** 1.000
Delinquent −.085 .499*** .449*** .115 .254* .242* .435*** .521*** .423*** .383*** 1.000
Aggressive −.115 .639*** .572*** .232* .305** .390*** .607*** .621*** .637*** .603*** .710*** 1.000
Open in a new tab

Note:

***

= .000,

**

< .010,

*

< .050,

VE = Violence exposure, PS1 = Wisconsin Card Sort Task factor 1, PS2 = Wisconsin Card Sort Task factor 2, TSW = Task switching, FRT = Flanker RT, SMRT = Shape match RT, AS = Anti-Saccade, GNG = Go/No Go, BI = Behavioral Index, MRI = Metacogntive Regulation Index, GD = BDI General Depression, PI = BDI Performance Impairment, SB = BDI Somatic Complaints, WD = YSR Withdrawn, SY = YSR Somatic Complaints, AD = YSR Anxious/Depressed, SP = YSR Social problems, DEL = YSR Delinquency, AG = YSR Aggression

Cognitive Processing

KBIT means for verbal, non-verbal and composite IQ were all within the average range. WRAT mean scores for reading and arithmetic were in the low average range. Twelve percent of the population qualified for specific math disability, 7% for reading and 5.2% for spelling disability. IQ and achievement scores were highly correlated (see Table 5).

Self-regulation, as measured by the BRIEF, was highly variable in the population. Clinical t-scores (70 or more) were reported by 16.7% of the population for self-regulation. In terms of cognitive processing, WCST means were all within the normal range but clinically significant t-scores were common (35% - 1 measure, 47% - 2 or more measures). TS mean RTs were significantly higher during switching than during non-switching tasks (t = 18.220, df = 98, p = .000). During interference, RT differed between the congruent and the incongruent trials (FS − t = 28.621, df = 97, p = .000; SM − t = 27.2348, df = 95, p = .000). Mean percent correct for AS was 89% and mean correct hits in Go/No-Go trials was 55%. Academic achievement and IQ correlated with measures of cognitive processing including problem solving (WCST), task switching, interference and response inhibition (see Table 5).

Mental Health Results

Means for the three subscales from the BDI were less than 1, which represents between no problem and minor problems on the 0–3 Likert scale. Mean scores on the YSR subscales ranged from .43 to .74 on a 0 to 2 Likert scale. The majority of participants scored in the normative range on the YSR on all eight subscales. Clinical levels of T-scores on internalizing subscales of the YSR ranged from 9.6–16.3% of the population, and on externalizing problems subscales ranged from 16.3–27.9%. Almost half of the sample (47.3%) reported clinically significant scores on more than one measure across the WCST, BRIEF, or the YSR. Mental health measures were highly intercorrelated and also correlated with low self-regulation (see Table 5).

Results of Hypothesis Testing

To test Hypothesis 1, that incarcerated adolescents will show a relation between intrafamilial violence and level of mental health function, correlations were conducted. Intrafamilial violence was examined based on both the level of intrafamilial violence (psychological, mild, severe and total) and the type of intrafamilial violence (sibling, parent and interpersonal violence). Intercorrelations between the intrafamilial violence measures were high as were intercorrelations among the mental health variables.

Total mean violence was related to the YSR withdrawn subscale (see Table 5). Psychological violence was associated with BDI Somatic Complaints. Mild violence was associated with BDI General Depression, YSR Withdrawn, Social Problems and Thought Problems. Severe violence was not significantly correlated with any mental health variable. Sibling psychological violence was significantly associated with all three BDI subscales and the Withdrawn scale from the YSR. Sibling mild violence was related to YSR Social Problems. All three types of sibling violence correlated with YSR attention. Parental psychological violence correlated with Somatic Complaints from the BDI and the YSR Withdrawn scale. Parent mild violence was related to three factors of the YSR Internalizing scale; Withdrawn, Anxious/depressed and Social Problems. Mild IPV was associated with BDI General Depression.

Model Testing

Hypothesis 2, that the above relation will be partially mediated by cognitive processing, learning or executive functioning problems, was tested using SEM with LISREL 8.52. The model (Table 5) had one latent independent variable (IV), seven mediating variables (MV) and four dependent variables (DV). In the measurement model, all 24 indicators were indicators for exactly one construct; within each construct indicator residuals were allowed to correlate when significant (see Tables 5 and 6). Although the χ2 difference test was significant (Kline, 2005) (see Table 7), the RMSEA and NNFI suggest a reasonably good fit of the model. NNFI suggests a good fit when the index is close to 1. A more liberal cut-off for a small sample is .80. An RMR of close to 0 suggests a good fit of the model. The RMR of this model was adequate. We therefore adopt the mediation model as the more parsimonious model with reasonable fit.

Table 6.

Measures used in Structural Equation Model

Independent Measure Mediation Variables Dependent Measures
Indicators Latent Constructs Indicators Latent Constructs Indicators
Total Intrafamilial violence Verbal Ability Verbal IQ Internalizing Withdrawn YSR
Reading Ability Anxious/depressed YSR
Spelling Ability Social Problems YSR
Nonverbal Ability Nonverbal IQ Externalizing Delinquent YSR
Arithmetic Ability Aggression YSR
Interference Processing Flanker RT Depression General depression BDI
Shape Match RT Performance Impairment BDI
Response Inhibition Go - No Go Rejections Somatic Complaints YSR Somatic
Anti-Saccade RT BDI Somatic
Switching Ability Switch cost
Frontal Lobe WCST factor 1
WCST factor 2
Self-Regulation Behavior Rating
Metacognition
Open in a new tab

Table 7.

SEM Models Indices

Model Description χ2 df RMSEA NNFI
1 Measurement Model (PSI Matrix Free) 434.81 237 0.073 0.88
2 Final Mediation Model 346.30 230 0.057 0.92
Model Test X2diff dfdiff p
2-1 88.52 7 0.000
Open in a new tab

In the final model (Figure 11), intrafamilial violence was negatively associated with non-verbal (performance) ability and self-regulation. Non-verbal ability was positively associated with response inhibition but did not have any relation with mental health adjustment. Self-regulation was associated with mental health function, including Internalizing, Externalizing, Depression, and Somatic Complaints such that higher self-regulation predicted better mental health adjustment. Mediation was shown where self-regulation mediated the relation between intrafamilial violence and mental health adjustment. The SEM analysis used pairwise deletion to remove missing variables from the covariance matrix for those subjects with too much missing data to be imputed.

Figure 1.

Figure 1

Open in a new tab

Full mediation model - Current self-regulation mediates relation between past violence exposure and current mental health functioning

Discussion

This study represents a significant contribution to the understanding of the role of intrafamilial violence in the development of mental health problems in incarcerated youth. There are very few studies that have attempted to study youth during incarceration and have overcome the significant barriers to receiving permission from research institutional review boards, states, corrective institutions, youth and parents to conduct research (Mulcahy, Krezmien, Leone, Houchins, & Baltodano, 2008). In addition, the present study incorporates measurement of past violence history, mental health, and cognitive processing including direct measurement of academic functioning and intelligence, self-regulatory capacity, response inhibition, and interference from distracters. This combined approach of psychological experimental research, records review and survey research allows for a comprehensive examination of the complexities of mental health functioning in incarcerated youth.

Mediation of Cognitive Processing

Childhood intrafamilial violence is a major public health problem that has devastating effects. An often-overlooked outcome of exposure to violence is in the realm of cognitive and academic competence, where intrafamilial violence during development can lead to debilitating problems (Jonson-Reid et al., 2004). The hypothesis that cognitive processing and related self-regulatory skills partially mediate the relation between intrafamilial violence and mental health function is partially supported in this population of incarcerated adolescent boys. Self-regulation measured at the time of incarceration, mediates the relation between past history of exposure to violence and present mental health functioning but other cognitive processing measures do not mediate this relation. The current study supports the findings of other researchers that intrafamilial violence has a negative influence on the development of self-regulation (Merwe & Dawes, 2000).

Our results suggest that self-regulatory capacity may be integral to explaining high rates of metal health problems in incarcerated populations. Dodge and Crick (1990) have suggested that the mechanism for this negative relation is through the over-salience of threat in the environment. Our finding that self-regulation capacity is a mediator suggests a failure to process social cues appropriately in children exposed to violence. Other research suggests that this may be due to a neuroanatomical failure of top-down control over emotionally salient stimuli, interpreted as threat. This provides a mechanism by which history of exposure to violence causes self-regulatory deficits, necessary for school competence and achievement (Eccles & Wigfield, 2002). We fail to find mediation of other cognitive processing measures, which suggests that self-regulation may separately mediate the relation between intrafamilial violence and cognitive processing or mental health.

This study does find evidence that intrafamilial violence is related to cognitive processing through non-verbal skills. Deficits in non-verbal skills are in turn related to a decreased ability to inhibit a prepotent response. This supports the more general idea that intrafamilial violence has detrimental effects on cognitive and academic skills. A relation between cognitive processing, academic skills and mental health function was found. As Nigg (2000) has suggested executive function may be central to this relation and has a particularly strong negative association in the present study as has been seen in previous studies of at-risk populations (Campbell, Sameroff, Lewis, & Miller, 2000). Finally, there is high correlation between the different types of family violence, which suggests that violence in the home is interrelated, a finding that is supported by past research (Straus & Gelles, 1990).

Disability

We exhibit further evidence that learning disabilities and behavioral problems are very common in incarcerated populations. Estimated national rates of mental health disorders in children using the Achenbach System of Empirically Based Assessment are between three and eleven percent (Achenbach, Dumenci & Rescorla 2003). For individual mental health we find clinical rates from between 11 and 28%, well over national norms but below those found in other studies of incarcerated youth (Colins, Vermeiren, Shuyten, & Broekaert, 2009; Quinn, Rutherford, Leone, Osher, & Poirier, 2005). This difference may be due to underreporting when using self-report measures.

Rates of cognitive processing problems and learning disabilities are similarly high. Seventy percent of the sample has current IEPs. This is at the higher end of prevalence rates reported in youth incarceration nationally (Quinn, Rutherford, Leone, Osher, & Poirier, 2005; Krezmien, Mulcahy, & Leone, 2008) and may reflect differences in state special education classification systems or differences in rates of youth incarceration versus service in residential placement. Mean IQ and achievement scores are 10 points below national norms at about 90, in line with the findings of other studies of incarcerated youth (Krezmien, Mulcahy, & Leone, 2008).

Comorbidity is also common. Over 70% of subjects have at least one measure of mental health or cognitive function on which they perform above the clinical cut-off. Of youth who meet the study criteria for clinical level mental health or academic problems, two-thirds of the sample present with more than one clinical level of academic and mental health functioning. Other studies have failed to adequately quantify comorbidity in the diagnosis of cognitive and learning disabilities and mental health problems in incarcerated populations. A more nuanced understanding of the subpopulations within incarcerated groups is necessary to further treatment and prevention programs (Perkins, Smith-Darden, & Graham-Bermann, 2011). Mental health and academic problems tend to co-occur and the compound effect is often more devastating. Approximately one-third of the sample has reached the age of maturity (18) suggesting that adults with IEPs are being treated in juvenile facilities rather than in adult facilities. High rates of cognitive and learning disabilities in prison populations argue for increased attention to the special needs of incarcerated youth (Mulcahy, Krezmien, Leone, Houchins, & Baltodano, 2008).

Limitations

The study has several limitations that include generalizability, diagnosis of cognitive and learning disabilities, missing data and data collection. The sample represents a particular population that is more constricted in certain ways that might have influenced the findings. Close to forty percent of the population had some experience with prior foster home placement, which suggests that many of the youth come from homes with a complex set of problems in the family of origin and our measures are limited in disentangling all of those complexities. This may also suggest that we have underestimated the rates of violence exposure. We asked about how conflicts were handled in the families. We intentionally left that open ended because we were aware that many of the youth might have had multiple home placements. Several boys specifically asked if they grew up with a relative or in a foster home how they should answer. We asked them to answer for the time period independent of the type of family but this may have lead to an underestimation.

The population has particularly high levels of mental health problems and low incidence behaviors (such as crime and delinquency). This might have created a ceiling effect on some measures. Every attempt was made to use measures that had strong reliability and validity but none of the norming samples included incarcerated adolescents. It is difficult for a large-scale study to accurately capture the complexity of diagnosis of disabilities. The inherent difficulties of collecting data in locked facilities made it extremely difficult to collect complete records on all subjects. Subjects were often released without notice, which meant that many subjects had incomplete data files. A total of 38 subjects were released with partial data collected. Missing data were replaced using an iterative process that maximized each subject’s empirical data. The conditions for collecting both survey and computerized data were not an ideal, e.g., group of adolescents were given the survey at the same time. The conditions may have compromised the integrity of the data that were collected, particularly in regard to reaction times on cognitive tasks.

Implications

There are approximately 100,000 juveniles in residential placement nationwide. In this study 70% of the sample had current IEPs. Juvenile detention facilities clearly service large populations of children with disabilities and academic challenges. The model tested emphasizes the development of cognitive and learning skills as central to the known relation between intrafamilial violence and mental health function. Targeted interventions directed at cognitive processing and remediation of academic problems would seem to be essential in children with significant exposure to violence regardless of whether or not their functioning on any of those dimensions rises to the level of a diagnosable disability.

The finding that intrafamilial violence and self-regulation are significantly related to mental health functioning in general suggests the need to assess self-regulatory skills, anxiety, depression and withdrawal and their relation to learning disability in this population. Prevention and intervention programs often focus on externalizing behavior problems of aggression and delinquency. This is likely due to the societal cost of these behaviors but the current study finds clinical levels of anxiety, depression or withdrawn behavior in our population. This suggests that underlying or preceding these behaviors may be deficits in self-regulatory behavior and that mental health must be approached holistically for at-risk youth. .

Social control theory would benefit from incorporating our findings on the mediation of self-regulation in the prediction of criminal behavior. Currently researchers in this vein have seen self-control as fixed at the age of ten (Winfree, Taylor, He, & Esbensen, 2006). Recently researchers have attempted to incorporate the development of cognitive processing into social control theory (Beaver, Wright, & Delisi, 2007). Further expansion of social control theory to incorporate self-regulation as a developmental construct, would add a mechanism by which violence may be connected to later criminal behavior.

One future direction for research might focus on evaluating the efficacy of intervention and rehabilitation programs that can target self-regulation skills and cognitive deficits. It may be that struggling with language and information processing contributes to youth who have difficulty with maintaining behavior that conforms to societal norms. Dodge and others described deficits in social information processing as central to misinterpreting social cues as threatening or aggressive. It may also be that cognitive processing deficits, which are often seen as separate from behavior, contributes to youth misunderstanding of interpersonal interactions. Thus, the ability to self-regulate appears to be central in understanding how family violence effects youth behavior and should be a target of both future research study and intervention.

Table 4.

Cognitive Processing and Mental Health Measure Means

Measure N Minimum Maximum Mean SD α
KBIT Verbal 88 62 130 91.69 14.03 0.91
KBIT Non-verbal 88 48 130 96.77 14.43 0.85
KBIT Composite 88 54 128 93.35 14.54 0.67
WRAT Reading 93 12 118 89.24 15.78 0.90
WRAT Spelling 95 49 123 90.23 15.99 0.80
WRAT Math 96 45 120 84.53 15.27 0.83
WCST Total Errors* 89 19.00 65.00 44.82 11.95 N/A
WCST Perseverative Responses* 89 20.00 72.00 46.67 11.29 N/A
BRIEF Global Composite Percent 95 1.00 99.00 59.83 30.89 93.39
AS Percent Correct 78 0.06 0.98 0.54 0.29 N/A
G-NG Correct Rejections Percent 93 0.32 1.00 0.89 0.16 N/A
TS - Switch Cost 93 −819.72 1588.48 248.41 464.21 N/A
FS difference in RT† 98 −1062.25 948.14 350.31 323.69 N/A
SM difference in RT† 96 −308.05 647.67 178.81 647.67 N/A
BDI General Depression 98 0.00 2.00 0.63 0.46 0.79
BDI Performance Impairment 98 0.00 1.86 0.49 0.46 0.73
BDI Somatic Complaints 98 0.00 2.33 0.53 0.51 0.18
YSR Withdrawn 103 0.00 2.00 0.64 0.44 0.75
YSR Somatic Complaints 100 0.00 2.00 0.43 0.41 0.84
YSR Anxious/Depressed 101 0.00 2.00 0.58 0.44 0.91
YSR Social Problems 101 0.00 1.88 0.58 0.41 0.75
YSR Thought Problems 102 0.00 2.00 0.51 0.74 0.75
YSR Attention Problems 101 0.00 1.78 0.74 0.44 0.80
YSR Delinquent 102 0.00 2.00 0.70 0.41 0.81
YSR Aggression 101 0.00 2.00 0.74 0.41 0.90
YSR Internalizing 101 0.00 2.00 0.55 0.38 0.94
YSR Externalizing 104 0.00 2.00 0.73 0.38 0.92
YSR Total 104 2.00 203.00 63.22 34.88 0.97
Open in a new tab

Note:

*

= t score,

†

= milliseconds.

Footnotes

1

Residual correlations between latent dependent variables were omitted for the sake of simplicity of the figure.

References

  1. Achenbach TM. Manual for the youth self-report and 1991 profile. Burlington, VT: University Associates in Psychiatry; 1991. [Google Scholar]
  2. Achenbach TM, Dumenci L, Rescorla L. Are American children’s problems still getting worse? A 23-year comparison. Journal of Abnormal Child Psychology. 2003;31(1):1–11. doi: 10.1023/a:1021700430364. [DOI] [PubMed] [Google Scholar]
  3. Bandura A. Social foundations of thought and action: A social cognitive theory. Englewood Cliffs, NJ US: Prentice-Hall; 1986. Prentice-Hall series in social learning theory. [Google Scholar]
  4. Beaver KM, Wright JP, Delisi M. Self-control as an executive function: Reformulating Gottfredson and Hirschi’s parental socialization thesis. Criminal Justice and Behavior. 2007;34(10):1345. [Google Scholar]
  5. Bernier A, Carlson SM, Whipple N. From external regulation to self-regulation: early parenting precursors of young children’s executive functioning. Child development. 2010;81(1):326–39. doi: 10.1111/j.1467-8624.2009.01397.x. [DOI] [PubMed] [Google Scholar]
  6. Buckner JC, Mezzacappa E, Beardslee WR. Self-regulation and its relations to adaptive functioning in low income youths. The American Journal of Orthopsychiatry. 2009;79(1):19–30. doi: 10.1037/a0014796. [DOI] [PubMed] [Google Scholar]
  7. Buckner JC, Beardslee WR, Bassuk EL. Exposure to violence and low-income children’s mental health: Direct, moderated, and mediated relations. American Journal of Orthopsychiatry. 2004;74(4):413–23. doi: 10.1037/0002-9432.74.4.413. [DOI] [PubMed] [Google Scholar]
  8. Campbell SM, Sameroff AJ, Lewis M, Miller SM. Handbook of Developmental Psychopathology. 2. Kluwer Academic Publishers; 2000. Attention- deficit/hyperactivity disorder: A developmental view; pp. 383–401. [Google Scholar]
  9. Carlson BE, Shafer SM. Traumatic Histories and Stressful Life Events of Incarcerated Parents: Childhood and Adult Trauma Histories. The Prison Journal. 2010;90(4):475–493. [Google Scholar]
  10. Casey BJ, Jones RM, Hare TA. The adolescent brain. Annals of the New York Academy of Sciences. 2008;1124:111–26. doi: 10.1196/annals.1440.010. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Colins O, Vermeiren R, Shuyten G, Broekaert E. Psychiatric disorders in property, violent, and versatile offending detained male adolescents. American Journal of Orthopsychiatry. 2009;79(1):31–8. doi: 10.1037/a0015337. [DOI] [PubMed] [Google Scholar]
  12. Curtis RL, Jr, Leung P, Sullivan E, Eschbach K, Stinson M. Outcomes of child sexual abuse contacts: Patterns of incarcerations from a national sample. Child Abuse & Neglect. 2001;25(5):719–736. doi: 10.1016/s0145-2134(01)00233-2. [DOI] [PubMed] [Google Scholar]
  13. Dodge KA, Crick NR. Personality & Social Psychology Bulletin: Special Issue: Illustrating the Value of Basic Research. 1. Vol. 16. US Sage Publications, Inc; 1990. Social information-processing bases of aggressive behavior in children; pp. 8–22. [Google Scholar]
  14. Eccles JS, Wigfield A. Motivational beliefs, values, and goals. Annual Review of Psychology. 2002;53(1):109–132. doi: 10.1146/annurev.psych.53.100901.135153. [DOI] [PubMed] [Google Scholar]
  15. Efklides A. Metacognition: Defining its facets and levels of functioning in relation to self-regulation and co-regulation. European Psychologist. 2008;13(4):277–287. [Google Scholar]
  16. Friedman NP, Miyake A. The relations among inhibition and interference control functions: A latent-variable Analysis. Journal of Experimental Psychology. 2004;133(1):101–135. doi: 10.1037/0096-3445.133.1.101. [DOI] [PubMed] [Google Scholar]
  17. Gottfredson M, Hirschi T. A General Theory of Crime. Stanford, CA: Stanford University Press; 1990. [Google Scholar]
  18. Graham-Bermann SA, Gruber G, Girz L, Howell KH. Factors discriminating among profiles of resilient coping and psychopathology in children exposed to domestic violence. Child Abuse & Neglect. 2009;33(9):648–660. doi: 10.1016/j.chiabu.2009.01.002. [DOI] [PubMed] [Google Scholar]
  19. Graham-Bermann SA, Howell KH, Miller LE, Kwek J, Lilly M. Traumatic events and maternal education as predictors of verbal ability for preschool children exposed to intimate partner violence (IPV) Journal of Family Violence. 2010;25(4):383–392. [Google Scholar]
  20. Guy SC, Isquith PK, Gioia GA. Behavior rating of executive function – self-report version professional manual. Lutz, Florida: Psychological Assessment Resources; 2004. [Google Scholar]
  21. Hanson RF, Borntrager C, Self-Brown S, Kilpatrick DG, Saunders BE, Resnick HS, et al. Relations among gender, violence exposure, and mental health: The national survey of adolescents. American Journal of Orthopsychiatry. 2008;78(3):313–21. doi: 10.1037/a0014056. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Haapasalo J, Virtanen T. Paths between childhood emotional and other maltreatment and psychiatric problems in criminal offenders. Journal of Emotional Abuse. 1999;1(4):15–35. [Google Scholar]
  23. Hosser D, Raddatz S, Windzio M. Child maltreatment, revictimization, and violent behavior. Violence and Victims. 2007;22(3):318–33. doi: 10.1891/088667007780842829. [DOI] [PubMed] [Google Scholar]
  24. Jonson-Reid M, Drake B, Kim J, Porterfield S, Han L. A prospective analysis of the relationship between reported child maltreatment and special education eligibility among poor children. Child Maltreatment. 2004;9(4):382–394. doi: 10.1177/1077559504269192. [DOI] [PubMed] [Google Scholar]
  25. Kaufman AS, Kaufman NL. Kaufman Brief Intelligence Test. 2. Circle Pines, Minnesota: AGS Publishing; 2004. [Google Scholar]
  26. Kenny DA. Mediation. 2006 Feb 7; Retrieved February 8, 2007, from http://davidakenny.net/cm/mediate.htm.
  27. Kline RB. Principles and practice of structural equation modeling. 2. New York: The Guilford Press; 2005. [Google Scholar]
  28. Kongs SK, Thompson LL, Iverson GL, Heaton RK. Wisconsin Card Sorting Test-64 Card Version Professional Manual. Lutz, Florida: Psychological Assessment Resources; 2000. [Google Scholar]
  29. Krezmien MP, Mulcahy CA, Leone PE. Detained and committed youth: Examining differences in achievement, mental health needs, and special education status. Education and Treatment of Children. 2008;31(4):445–64. [Google Scholar]
  30. Margolin G, Vickerman K, Oliver P, Gordis E. Violence Exposure in Multiple Interpersonal Domains: Cumulative and Differential Effects. Journal of Adolescent Health. 2010;47(2):198–205. doi: 10.1016/j.jadohealth.2010.01.020. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Merwe AVD, Dawes A. Prosocial and antisocial tendencies in children exposed to community violence. Journal of Child and Adolescent Mental Health. 2000;12(1):19–37. [Google Scholar]
  32. Mischel, et al. ‘Willpower’ over the life span: decomposing self-regulation. Social Cognitive and Affective Neuroscience. 2010;2010:1–5. doi: 10.1093/scan/nsq081. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Mulcahy CA, Krezmien MP, Leone PE, Houchins DE, Baltodano H. Lessons learned: Barriers and solutions for conducting reading investigations in juvenile corrections settings. Reading & Writing Quarterly. 2008;24(2):239–252. [Google Scholar]
  34. Nigg JT. On inhibition/disinhibition in developmental psychopathology: Views from cognitive and personality psychology and a working inhibition taxonomy. Psychological Bulletin. 2000;126(2):220–246. doi: 10.1037/0033-2909.126.2.220. [DOI] [PubMed] [Google Scholar]
  35. Perkins SC, Smith-Darden J, Graham-Bermann S. Typologies of Achievement and Violence Exposure in Incarcerated Adolescents. 2011. Manuscript submitted for publication. [Google Scholar]
  36. Preacher KJ, Coffman DL. Computing power and minimum sample size for RMSEA [Computer software] 2006 May; Available from http://www.quantpsy.org/
  37. Quinn MM, Rutherford RB, Leone PE, Osher DM, Poirier JM. Youth with Disabilities in Juvenile Corrections: A National Survey. Exceptional Children. 2005;71(3):339–346. [Google Scholar]
  38. Reed E, Raj A, Falbo G, Caminha F, Decker MR, Kaliel DC, Missmer SA, Molnar BE, Silverman JG. The prevalence of violence and relation to depression and illicit drug use among incarcerated women in Recife, Brazil. International Journal of Law and Psychiatry. 2009;32(5):323–328. doi: 10.1016/j.ijlp.2009.06.006. [DOI] [PubMed] [Google Scholar]
  39. Rogers RD, Monsell S. Costs of a predictable switch between simple cognitive tasks. Journal of Experimental Psychology: General. 1995;124(2):207–231. [Google Scholar]
  40. Ross T, Fontao MI. Self-Regulation in violent and non-violent offenders: A preliminary report. Criminal Behaviour and Mental Health. 2007;17(3):171–178. doi: 10.1002/cbm.651. [DOI] [PubMed] [Google Scholar]
  41. Ruchkin V, Henrich CC, Jones SM, Vermeiren R, Schwab-Stone M. Violence exposure and psychopathology in urban youth: The mediating role of posttraumatic stress. Journal Abnormal Child Psychology. 2007;35(4):578–593. doi: 10.1007/s10802-007-9114-7. [DOI] [PubMed] [Google Scholar]
  42. Shafer AB. Meta-analysis of the factor structures of four depression questionnaires: Beck, CES-D, Hamilton, and Zung. Journal of Clinical Psychology. 2006;62(1):123–146. doi: 10.1002/jclp.20213. [DOI] [PubMed] [Google Scholar]
  43. Straus MA, Gelles RJ. Physical Violence in American Families. New Brunswick, New Jersey: Transaction Publishers; 1990. pp. 29–47. [Google Scholar]
  44. Schwartz D, Gorman AH. Community violence exposure and children’s academic functioning. Journal of Educational Psychology. 2003;95(1):163–173. [Google Scholar]
  45. Tang, et al. Short-term meditation training improves attention and self-regulation. Proceedings of the National Academy of Sciences of the United States of America. 2007;104(43):17152–6. doi: 10.1073/pnas.0707678104. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Thompson RA, Wyatt JM. Current research on child maltreatment: Implications for educators. Educational Psychology Review. 1999;11(3):173–201. [Google Scholar]
  47. U.S. Department of Education, Office of Special Education Programs. Annual Report to Congress on the Implementation of the Individuals with Disabilities Education Act. retrieved April 14, 2009, from http://www.ideadata.org/PartBdata.asp.
  48. Widom CS, White HR. Problem behaviours in abused and neglected children grown up: Prevalence and co-occurrence of substance abuse, crime and violence. Criminal Behaviour & Mental Health. 1997;7(4):287–310. [Google Scholar]
  49. Wilkinson GS. The Wide Range Achievement Test Administration Manual. Wilmington, Delaware: Wide Range, Inc; 1993. [Google Scholar]
  50. Winfree LT, Taylor TJ, He N, Esbensen FA. Self-control and variability over time: Multivariate results using a 5-year, multisite panel of youths. Crime & Delinquency. 2006;52(2):253–286. [Google Scholar]

RESOURCES