Abstract
Maltreated and child welfare-involved youth are over-represented in juvenile justice systems. These youth are at a greater risk of serious offending and justice system entrenchment relative to their non-maltreated peers. Understanding gender differences in the pathways to justice involvement and the nature of offending among maltreated children is critical for informing policy and practice. Yet, this body of evidence is fragmented. This scoping review identified and narratively synthesized evidence from studies reporting on gender differences in the individual characteristics, maltreatment experiences, child protection involvement and offending profiles of maltreated youth who offend. A comprehensive search of four databases generated 11,568 publications, from which 180 met the review’s inclusion criteria. These primary studies included participants aged 8–21 years with a history of childhood maltreatment and youth offending and reported at least one gendered analysis. Some consistent findings were reported across studies. A greater level of child welfare involvement and maltreatment exposure (particularly sexual abuse and multi-type maltreatment) was found for justice involved girls, relative to boys. Maltreated and child welfare-involved boys appear more likely to offend than girls, but findings about how gender moderates the maltreatment-offending relationship were inconsistent. Child welfare systems involvement (particularly foster care and residential care) appeared to be an important moderator for girls, and school performance mediated outcomes for boys. Across this body of evidence, few studies accounted for under-reporting of abuse and neglect when using youth self-report measures of maltreatment. Future research is needed which explicitly explores how gender moderates the maltreatment-offending relationship.
Introduction
Rationale
A growing body of literature has examined the relationship between youth offending and childhood maltreatment (Malvaso, Delfabbro, & Day, 2018). Maltreated youth are over-represented in juvenile justice systems and face greater risks of early onset and more violent offending, as well as later involvement in the adult criminal justice system (Cho et al., 2019; Malvaso et al., 2017b). Reflecting these trajectories, youth involved in child welfare systems are also over-represented among juvenile justice populations, constituting a group termed ‘crossover’, ‘dually involved’ or ‘dually adjudicated’ children, in recognition of their traversal of these two statutory systems (Baidawi & Sheehan, 2020b).
Youth offending is not evenly distributed by gender, with males being more likely than females to offend and have contact with the justice system. Criminal justice statistics indicate that males account for 70–80% of all justice-involved youth in the United States, Canada and Australia (Australian Institute of Health and Welfare, 2020; Malakieh, 2020; Puzzanchera, 2020). Additionally, the offending repertoires of young males are generally more prolific, violent and persistent than their female counterparts (Fergusson & Horwood, 2002; Loeber et al., 2017). Due to their over-representation in justice systems, much of what is known about the origins of youth offending, risk factors for recidivism and effective interventions has been derived from research with male samples. However, the notion of gender distinction is becoming increasingly recognized and explored in the youth offending literature, as several jurisdictions grapple with increases in girls’ arrests (see Pusch & Holtfreter, 2018; Shepherd et al., 2013a). Understanding differences in the pathways, risks/needs and offending trajectories of justice-involved youth is necessary to inform whether and how approaches for preventing youth (re)offending need to be differentiated by gender.
Although research identifies overlap in important risk factors for youth offending across gender (Fergusson & Horwood, 2002), a growing body of literature on female delinquency highlights several differences. Justice-involved young females evidence higher rates of dysfunctional family dynamics, abuse and victimization than their male counterparts (Conrad et al., 2014; Hubbard & Pratt, 2002). These traumatic experiences are argued to be relevant drivers of female delinquency via pathways of truancy, emotion dysregulation, substance abuse, mental illness, poverty, peer rejection, and further victimization and exploitation (Blum et al., 2003; Shepherd et al., 2013b). Several studies show female offenders have higher rates of mental health issues than male offenders, including post-traumatic stress disorder, substance use disorders, depression, psychosis, self-harming behaviour and attempted suicide (e.g. Coid et al., 2009). In contrast, learning difficulties, school disconnectedness and disengagement, delinquent peer influences and antisocial personality traits have been shown to be more relevant for young male offenders (Blum et al., 2003; Piquero et al., 2005).
In recognition of these gendered pathways to youth offending and need for gender-specific responses, a growing number of studies exploring the relationship between child maltreatment and youth offending have undertaken gendered analyses. Other studies have highlighted the need to understand the unique experiences of males and females in child welfare systems, in relation to their risk and experiences of offending (Ryan & Testa, 2005). However, this research is fragmented and there is no existing review that consolidates gender-related findings for this population. This scoping review sought to consolidate the body of evidence in this area to provide researchers, practitioners and policy-makers with an up-to-date overview of evidence of reporting on gender differences related to the maltreatment-youth offending nexus. A synthesis of this research will support the development of gender-specific preventative and responsive measures to these trajectories, as well as informing future research to address the over-representation of maltreated and child protection-involved children in youth justice systems.
Objective
This scoping review aimed to identify and synthesize evidence from studies that report on gender differences in the individual characteristics, maltreatment experiences, child protection involvement and offending of maltreated youth who offend and/or experience contact with juvenile justice systems.
Methods
Scoping Reviews
As opposed to systematic reviews, scoping reviews are broader in scope and aim to map the extent and nature of research activity in a field, which enables the identification of research gaps, as well as areas where more narrowly focussed systematic reviews may be viable (Arskey & O’Malley, 2005). Scoping reviews can summarize and disseminate key research findings, but do not generally include an appraisal of the quality of the evidence reported.
Protocol and Registration
All review methods were determined a priori and written into a protocol, before commencing literature searches and screening. While registration was attempted, PROSPERO was not accepting scoping review protocols at that time. The original protocol can be provided by the corresponding author on request. The review follows the reporting guidelines outlined in the PRISMA Extension for Scoping Reviews (PRISMA-ScR) Checklist (see Supplementary Appendix A).
Eligibility Criteria
The following selection criteria were used to determine whether to include or exclude studies identified by the search strategy.
Participants
Participants included children/youth with a history of child maltreatment and youth offending. Studies inclusive of offending between the ages of 8–21 years of age were included with a view to capturing research that examined juvenile justice involvement across diverse jurisdictions and offending behaviour among youth transitioning from out-of-home care (OOHC). Studies may have compared maltreated and offending youth with other groups, for instance, non-maltreated offending youth or maltreated non-offending youth. Studies included both male and female child/youth participants, while studies including only female or only male participants were excluded.
Because this review focused on youth offending, studies that included only adult participants were excluded. For example, studies were excluded if they only reported on the relationship between childhood maltreatment/child protection involvement and adult crime/criminal justice involvement. Where studies included both youth and adult participants, they were only included when the majority of the participants were under the age of 18 or where sub-analyses separating age cohorts were included.
Settings
Eligible settings included, but were not limited to, juvenile justice, child welfare, education, homelessness and health settings. Studies that used surveys or questionnaires, and relevant studies that analysed administrative or other data relating to the target participants were also included.
Study designs
Studies must have reported on an empirical primary research study relating to the intersection between child maltreatment (either self-report or as indicated by child protection involvement as a proxy) and youth offending (crime and/or youth justice system involvement). Studies that used either quantitative or qualitative methods were eligible for inclusion. Definitions of child maltreatment were inclusive of self-report measures (including adverse childhood experiences (ACE) scales and via standardized instruments such as the Childhood Trauma Questionnaire), as well as administrative data (e.g. child protection substantiations and court records). While child protection system involvement does not always relate to experiences of abuse and neglect, a large body of studies in the maltreatment-youth offending field have utilized protective services’ administrative data as a proxy measure of child maltreatment. Similarly, definitions of youth offending were inclusive of self-report as well as official records/indicators (e.g. youth convictions or sentences). Studies must also have included an assessment or analysis of sex/gender differences of relevant outcomes relating to children who experience child maltreatment and youth offending.
Measures of interest
This review aimed to capture studies that described gender/sex differences in measures relating to the relationship between child maltreatment and youth offending. Given the expectation that there would be few studies that report a gender analysis, this review included studies that reported on a range of measures. These included, but were not restricted to, the following: Child/youth characteristics: for example, age, race/ethnicity, mental health and disability diagnoses. Child/youth maltreatment experiences: for example, types of maltreatment, experiences of multi-type maltreatment and maltreatment recurrence/persistence. Child/youth child protection pathways: for example, age at first child protection notification/substantiation, age at first OOHC placement, placement stability and placement types. Child/youth offending: any offending/convictions, offence types and violent offending.
Studies solely reporting measures relating to aggression, antisocial or externalizing behaviour, and substance misuse, rather than specific offending among children and young people of the age of criminal responsibility, were not included. Additionally, as this review is primarily interested in examining gender differences in the childhood maltreatment (by caregivers or other adults)-youth offending nexus, studies solely including victimization in the context of romantic relationships (or ‘dating violence’) were excluded.
Sources of evidence
The review included evidence from both published and non-published sources. There were no limits placed on the language or year of publication.
Literature Search Strategy
An electronic search strategy was designed to identify studies that reported on gender differences in the characteristics, child protection pathways and offending of child protection-involved youth. Keywords relating to “youth” (e.g. child, minor and adolescent), “maltreatment” (e.g. abuse, neglect and trauma), “child protection” (e.g. child welfare, looked after and foster care), “offending” (e.g. delinquency, crime and recidivism) and “gender” (e.g. male, female, sexes) were used to search the following electronic databases in November 2020): (1) Ovid MEDLINE(R), Ovid MEDLINE(R) In-Process & Other Non-Indexed Citations, Ovid MEDLINE(R) Daily and Ovid OLDMEDLINE(R) (1946 to present); (2) Criminal Justice Abstracts via Ebsco (1980 to present);(3) ProQuest: Social Sciences Premium Collection (1914 to present) and (4) CINCH—Australian Criminology Database (1882 to present). Text word searches were mapped verbatim into each database, excepting adjustments made for database specific syntax. The reference lists of systematic reviews identified by the search strategy were also hand searched for additional eligible primary studies.
Data Management and Software
Reference management software EndNote X8 (Clarivate Analytics, 2018) was used to compile all titles and abstracts derived from the search strategy, and duplicates were removed. All citations were then transferred to systematic review software Covidence to undertake title/abstract and full text screening of studies and to identify, track and resolve discrepancies across reviewers.
Study Selection
Prior to study selection, all review authors underwent training to ensure a comparable understanding of the purpose of the review and the selection criteria. Titles and abstracts retrieved from the electronic searches were screened to exclude publications that did not meet the selection criteria. This stage of the screening process was highly inclusive, and a full text review was undertaken when the information provided in the titles and abstracts was unclear or insufficient. All 11,568 titles and abstracts were screened by one author (A1) and 26.1% in total were screened by the other two authors (A2 and A3). Consensus was reached by a discussion between authors in cases of conflict.
For the full text review, each study was assessed by one review author (A1), and 24.7% duplicate assessments were performed independently by the other two reviewers (A2 and A3). There was an 86% agreement rate between screeners for duplicate screened text. Discrepancies were again discussed within the research team until a consensus was reached.
Data Charting Process and Items
A data extraction form was developed a priori, and one reviewer extracted the data (A1). The following data items were extracted: study characteristics (e.g., authors, year published, country and publication type); study purpose/aims; methods; study design; participant information (including population, sample size, % male); maltreatment measures and offending measures. A summary of key analyses by gender was also recorded. Publications that drew data from the same study were extracted (and have been reported) separately, in order to capture the different analyses presented.
Synthesis of Results
Following data extraction, studies were characterized by sample type (juvenile justice, child welfare, crossover, other, community and mixed samples). The key results of the scoping review were then narratively synthesized.
Results
Selection of Evidence Sources
A flow diagram outlining the study selection process is presented in Figure 1. Initial title and abstract screening of the 11,568 publications identified 11,014 irrelevant studies, with 530 studies deemed eligible for full text screening. Of these, 180 met the inclusion criteria. PRISMA flow diagram.
Sample Characteristics
Sample Characteristics (n = 180).
a‘Other’ includes a book chapter and a bulletin.
b‘Other’ includes the Netherlands, Serbia, Spain, Portugal and the UK (each n = 1).
c‘Community’ samples include general community samples of youth, school children or samples recruited from service settings (e.g. homelessness and mental health services).
dSample includes more than one of the above sample types.
eRandomized control trial.
fIncludes caregiver report, other administrative data (e.g. mental health) or triangulation of multiple data sources.
A minority of studies utilized diverse community samples (10.0%) including general population samples, school children and samples of homeless youth. Study designs were predominantly cross-sectional (48.9%) or longitudinal (48.3%), with a small number of randomized control trials and case–control studies (n = 5 studies). Studies with juvenile justice samples were predominantly cross-sectional in nature (75.9%), analysing the maltreatment histories of these children, while those drawing on child welfare or ‘multiple’ samples were mostly longitudinal in design (81.1% and 71.4%, respectively), examining the trajectories of some children into offending or justice system involvement.
Finally, in relation to maltreatment measures, around one half of studies (51.1%) utilized child welfare (child protection system and/or court) data, 40.6% drew on some form of youth self-report (e.g. historical assessments or current interview/survey data), and the remaining 8.3% of studies utilized multiple data sources to gather maltreatment data (11/15 of these publications included youth self-report maltreatment data in addition to informant and/or administrative maltreatment data). Youth self-report measures of maltreatment tended to be utilized more in studies with juvenile justice (74.7%) and community samples (66.7%) compared with studies that drew on child welfare samples (13.2%).
A Note on Terminology, Diversity and Intersectionality
In seeking to provide an overview of studies that examine gender differences in the child maltreatment-youth offending nexus, the review itself addresses a key area relating to diversity in this field. This scoping review utilizes the term ‘gender’ to denote biological sex of study participants/samples, as described by the included studies, yet it is important to acknowledge this terminology in the broader social sciences typically references gender identity. Also of note is that only one of the included studies included analyses or findings relating to sexual or gender identity (Narendorf et al., 2020), representing a gap in the research base. On the other hand, a number of studies further probed gender differences by incorporating gender by race analyses, as highlighted in the results below. Findings of these studies demonstrate the value of looking beyond biological sex to consider intersectionality in the child maltreatment-youth offending relationship.
Synthesis of Results
Summary Table of Critical Findings (N = 180).
Gender differences in the maltreatment experiences of offending youth
A total of 59 publications (32.8%) reported on gender differences in the maltreatment experiences of offending youth. These often found a greater prevalence of maltreatment among female youth justice samples compared with their male counterparts (n = 27) (see, e.g. Kenny et al., 2007; King et al., 2011; Malvaso, Delfabbro, Day, & Nobles, 2018; Moore et al., 2013; Protic et al., 2020; van der Put et al., 2014; Vitopoulos et al., 2019) and appeared particularly pronounced in studies analysing gender differences in sexual abuse and multi-type maltreatment (n = 32) (Baglivio & Epps, 2016; Conrad et al., 2014; Farina et al., 2018; Kenny et al., 2007; King et al., 2011; Kowalski, 2020; Nowakowski-Sims & Rowe, 2017; Vitopoulos et al., 2019; Wolff et al., 2020). Some studies contrasted maltreatment prevalence among justice-involved youth by gender and offence type. For instance, Levenson et al. (2017) reported greater prevalence of maltreatment among female juvenile sex offenders (JSOs) compared with female juvenile non-sex offenders (JNSOs), but inconsistent findings among males, where some maltreatment types were higher in JSOs (e.g. physical and sexual abuse, physical neglect) and others in JNSOs (emotional abuse and emotional neglect). Such findings demonstrate the importance of examining gender differences in maltreatment prevalence among subtypes of justice-involved youth. Another notable feature of some studies that utilized self-report of maltreatment was the inclusion of measures to detect under-reporting of child maltreatment or idealization of childhood; these studies identified justice-involved males as more inclined to under-report or minimize experiences of abuse and neglect compared with females (Kenny et al., 2007; Moore et al., 2013) and more likely to idealize childhood experiences relative to non-convicted males (Protic et al., 2020). Variations in maltreatment under-reporting by race/ethnicity were also identified.
Studies of community samples have similarly found a higher prevalence of maltreatment among females who offended relative to males who offended, including those examining likelihood of general offending, as well as specific offence types (Duke et al., 2010; He et al., 2019; Herrera & McCloskey, 2001). Studies examining gender differences in maltreatment among crossover samples involved in both the youth justice and child protection systems have produced more mixed findings (Baidawi & Sheehan, 2020a; Dirig, 2016. One study found a greater prevalence of emotional and sexual abuse among female crossover children, relative to male crossover children, but no gender differences in the prevalence of other maltreatment types (e.g. physical abuse, neglect and exposure to family violence) (Baidawi & Sheehan, 2019, 2020a).
Gender differences in the child welfare system experiences of offending youth
Of the total, 24 studies (13.3%) reported on gender differences in the child welfare system experiences of offending youth. Several of these studies (n = 10) found that justice-involved females experienced greater child welfare system involvement than their male counterparts (e.g. lifetime contact, number of notifications, placement in OOHC and placement in residential care) (Fitzgerald et al., 2012; Malvaso, Delfabbro, Day, et al., 2018; Ringland et al., 2015; Rodway et al., 2011; Shrifter, 2012; Taylor-Kindrick, 2011). Similarly, several studies (n = 11) demonstrated that dually involved or dually adjudicated individuals in the juvenile justice system are more likely to be female (Baglivio et al., 2016; Baidawi & Sheehan, 2020a; Dierkhising et al., 2019; Huang et al., 2012). Studies conducting other relevant analyses did not find gender differences in age of initial child protection involvement among crossover children (Sentencing Advisory Council [SAC], 2020). However, some studies found gender differences in the grounds for child protection involvement among crossover children (Shrifter, 2012) and in maltreatment recurrence following arrest (Huang et al., 2012).
Greater maltreatment exposure and child welfare system involvement among racial/ethnic minority females and males who offend, relative to their race/ethnic majority counterparts, is also noted across several studies (AIHW, 2016, Sentencing Advisory Council, 2020; Baidawi & Sheehan, 2020a; King et al., 2011). For instance, the Australian Institute of Health and Welfare (2012) identified that 31% of Indigenous females under youth justice supervision in Australia had one or more child protection notifications, compared with 19% of Indigenous males, 17% of non-Indigenous females and 8% of non-Indigenous males, demonstrating the importance of intersectional perspectives in understanding these relationships.
Gender differences in offending among maltreated/child welfare-involved youth
A total of 62 publications (34.4%) examined gender differences in offending of maltreated or child welfare-involved youth, for example, in relation to general offending, violent offending, age of onset and recidivism (e.g., Fox, 2019; Song et al., 1998). Most prominently, longitudinal and cross-sectional analyses with child welfare and community samples identify that maltreated or child welfare-involved males are more likely to exhibit lifetime or current juvenile offending and to experience justice system contact (convictions, detention, etc.), compared with maltreated or child welfare-involved females (n = 37 studies) (e.g., Goodkind et al., 2013; Halemba et al., 2004; Jonson-Reid & Barth, 2000; Malvaso et al., 2017b; Ryan et al., 2010; Vidal et al., 2017). These findings are also consistent across samples of youth transitioning from OOHC (e.g., Cusick et al., 2012), and others that examine offending violence or seriousness (Bjorkenstam et al., 2018; Garrido et al., 2018; George, 2010; Malvaso et al., 2017b), or specific offence types (Mota et al., 2016; Wall, 2004; Widom & Maxfield, 1996). Even after accounting for the impact of maltreatment, child welfare involvement, and individual variables in multivariate models, several studies (n = 5) indicate that males’ greater likelihood of offending and/or justice system contact persists among maltreated and child welfare-involved youth (Malvaso et al., 2017b; Vidal et al., 2017; Yampolskaya & Chuang, 2012).
At the same time, some findings suggest that the impact of childhood maltreatment exposure on the risk of youth offending and juvenile justice involvement may be more significant for females compared with males (Widom & Maxfield, 1996). For example, Maxfield and Widom (1996) found that maltreatment increased the risk of arrest in both males and females, but the impact was slightly more pronounced among females, both in relation to any arrest and violent arrest. However such findings remain inconsistent, with other studies reporting that sex does not appear to moderate the relation between maltreatment and youth offending outcomes (e.g., Bender, 2008). Finally, some studies have sought to analyse gender differences in the impact of specific maltreatment types. Several such studies have highlighted the more detrimental impact of sexual abuse on youth offending among males, relative to females (Duke et al., 2010; Kim, 2005; Matta Oshima et al., 2014; Morrow et al., 2019).
Gender x race differences in the maltreatment-offending relationship were examined in studies drawing on both child welfare and youth justice samples (e.g., AIHW, 2012; Baidawi & Sheehan, 2019; Bright & Jonson-Reid, 2008; George, 2010; Goodkind et al., 2013; Ryan et al., 2010). These studies at times identified a greater risk of justice system involvement among maltreated racial minority youth by gender. For instance, an Australian birth cohort study found that 63% of Indigenous males with a substantiated notification had also been under juvenile justice supervision in the study state, compared with 14% of non-Indigenous males, 17% of Indigenous females and 5% of non-Indigenous females (AIHW, 2012).
Gender differences in violent and serious offending among maltreated/child welfare-involved youth
Twenty publications (11.1%) specifically examined gender differences in the likelihood of more violent or serious offending among maltreated youth (e.g., Baidawi & Sheehan, 2019; Fitzgerald et al., 2012; Fore, 2018; Fox, 2019; Malvaso, Delfabbro, Day, et al., 2018; Perez, 2017). Whether conducted with juvenile justice or community samples, these studies generally found associations between increased violence and greater levels of maltreatment (Clark, 2010; Garrido et al., 2018; O’Connor, 1994). Some studies found further gender differences in these relationships while others did not (Fitzgerald et al., 2012; Fore, 2018), depending on the variables being examined. For example, two studies found that child welfare-involved female offenders had fewer charges and less violent charges relative to child welfare-involved male offenders (see, e.g. Baidawi & Sheehan, 2019). Kaufman (2003) found that for females, involvement in more serious delinquency was associated with childhood physical abuse and sexual abuse, while only neglect was found to be associated with multiple problem behaviours for males. The findings in relation to the impact of physical child abuse and girls’ violent offending were echoed in another study (Herrera & McCloskey, 2001). Several studies of justice-involved youth have reported that after accounting for maltreatment and adversity, males were more likely to violently offend compared with females (Clark, 2010; Malvaso, Delfabbro, Day, et al., 2018; Perez, 2017). Some studies used separate path models to examine gender differences in factors leading to violence, including maltreatment. For example, O’Connor (1994) found that for girls, abuse and placement in a foster home did not directly affect violence, but influenced the use of violence through other factors, such as math achievement and withdrawal. For males, abuse appeared to have a major impact, relating directly to lower verbal intelligence (among lower socioeconomic (SES) status males) and higher rebelliousness (among higher SES males), and culminating in poor school achievement, low self-esteem and high withdrawal. Other variables (e.g. SES) were important predictors in both the male and female models.
Gender differences in age of offending onset of maltreated/child welfare-involved youth
Though younger justice-involved children are known to have experienced greater maltreatment and child protection involvement (SAC, 2020), only two studies (1.1%) examined the impact of maltreatment on the age of onset of offending by gender. For example, one study reported gender differences in the age of delinquency onset among maltreated youth, with males exhibiting younger age of onset as compared with females (Cho et al., 2019).
Gender differences in the relationship between maltreatment/child welfare involvement and recidivism
Another significant area of exploration, examined by 25 studies (13.9%), was gender differences in the impact of maltreatment and/or child welfare involvement on youth recidivism. Several studies employing multivariate models identify both maltreatment and male gender as significant predictors of recidivism (Benda & Tollett, 1999; Dembo et al., 1996). For example, in a logistic regression predicting 12-month return to juvenile justice, Benda and Tollett (1999) reported that male gender, and being subject to abuse/neglect, was each associated with significantly greater odds of recidivism. However, such findings are not unequivocal, with other studies finding that solely maltreatment, solely gender, or neither, predict youth recidivism (see, e.g. Craig et al., 2020). For example, in a moderated logistic regression with matched samples of male and female incarcerated youth, Vitopoulos et al. (2019) found that criminal history and maltreatment exposures were both significant predictors of recidivism and that gender did not moderate the relationship between number of maltreatment types and recidivism. Conversely, Baglivio et al. (2016) reported that among males, cumulative ACE exposures were a significant predictor of recidivism, but once criminal history and individual risk factors were controlled for, the effect of cumulative ACEs was insignificant. For females in this study, cumulative ACE exposures were not significant predictors of recidivism. Conversely, Conrad and colleagues (2014) found no significant relationship between overall maltreatment and recidivism for either gender; however childhood sexual abuse was a significant predictor of recidivism among females but not males. The apparent divergence in these findings may partly reflect differences in samples (including the prevalence of maltreatment across specific samples), or differences in the impact of maltreatment specifically, compared with ACEs, which encompass both maltreatment and items relating to household/family adversity. Studies investigating the impact of total ACE scores on the odds of recidivism report diverse findings. Depending on the population and recidivism measures examined, studies generally found ACE scores to be associated with greater odds of recidivism among males, but having no or less impact on female recidivism (e.g., Kowalski, 2019), or as associated with equal odds of reoffending among both males and females (Craig et al., 2017). One study also determined that both male and female children who reported a greater number of ACEs were likely to be rearrested sooner (Wolff et al., 2018).
Other studies have analysed gender differences in recidivism among crossover youth who experience involvement with both child welfare and youth justice systems (Halemba et al., 2004; Huang et al., 2012). Halemba et al. (2004) reported that justice-involved females with a child welfare history were more likely to recidivate over a 12-month period than their male counterparts (65% vs 61%), though it was unclear if this difference was statistically significant. On the other hand, Huang et al. (2012) found that dually involved male youth were 1.33 times more likely to have a new offence over a 6 year period, compared with their female counterparts. While these findings apparently diverge, the difference in either instance is small relative to gender differences in recidivism among the broader population of justice-involved youth.
Other differences arose depending on the definition of recidivism utilized. For example, Taylor-Kindrick (2011), in mixed gender logistic regression models predicting rearrest and readjudication and probation violations, reported neither gender nor physical or sexual abuse were significant predictors in either instance. In single gender models, however, abuse history was a significant predictor of several examined recidivism outcomes (rearrest, readjudication and probation violation) and was a better predictor than risk assessment scores in the case of institutional commitment for girls, but not for boys (Taylor-Kindrick, 2011).
The above findings regarding the maltreatment-recidivism relationship by gender are not necessarily reflected in child welfare samples. For example, a cohort study of children in foster care found that a second arrest was more common among females than males (Huang et al., 2016), while another identified gender differences in recidivism by placement type (Ryan et al., 2010). Gender differences in other child welfare factors impacting on recidivism have also been examined, identifying that certain factors appear more influential for males (e.g. running away and school-related variables) and females, respectively (e.g. OOHC) (Ringland et al., 2015; van der Put et al., 2014). Finally, other studies have investigated the differential impacts of child welfare involvement on recidivism for race/ethnic- and sex-specific subsamples (Baglivio & Epps, 2016; Zettler et al., 2018), or examined the effect of maltreatment or ACEs and gender on recidivism while accounting for other variables of importance including future orientation (Craig, 2019), mental health and substance use (Craig et al., 2019), substance abuse and mental health treatment (Kowalski, 2020), resilience (Fox, 2019) and social bonds (Craig et al., 2017). For instance, Craig et al. (2019), reported on gender differences in path analyses of the ACE-recidivism relationship that considered direct and indirect effects via mental health and substance use variables, finding an indirect effect of ACEs on the likelihood of recidivism that operated through child welfare placement for male, but not female youth.
Gender differences in the relationship between maltreatment/child welfare involvement and juvenile justice outcomes and experiences
Four studies (2.2%) analysed relationship between maltreatment/ACEs, gender and receiving custodial sanctions (Pasko & Mayeda, 2011; Taylor-Kindrick, 2011). For example, Zettler and colleagues (2018) found that ACE scores and male gender were both predictive of residential juvenile justice placement among their sample, even when considering other factors (e.g. mental health history and race). Child welfare placement was not predictive of residential juvenile justice placement in this model, or in other models that examined gender and race specific subsamples
Gender differences in the child welfare factors that relate to offending
Twenty-two publications (12.2%) examined a range of child welfare factors as moderators/mediators of the child maltreatment-youth offending relationship, including maltreatment type, substantiation status and recurrence, age at out-of-home placement, number of removals, placement stability, reasons for placement, age of case closure, service/intervention type (e.g. in home vs out-of-home intervention) and placement type (Bright & Jonson-Reid, 2008; Cho et al., 2019; George, 2010; Goodkind et al., 2013; Huang et al., 2016; Ryan et al., 2010; Vidal et al., 2017; Yampolskaya & Chuang, 2012). Of note were several studies reporting gendered impacts of child welfare involvement, including OOHC placement, on children’s risk of offending or justice system involvement (E. Bjorkenstam et al., 2018; Bright & Jonson-Reid, 2008; Jonson-Reid & Barth, 2000). These studies have in many cases found that child welfare intervention more greatly increases the likelihood of conviction among females relative to males (E. Bjorkenstam et al., 2018; Bright & Jonson-Reid, 2008; Jonson-Reid & Barth, 2000).
Furthermore, several studies (n = 8) found that OOHC placement in particular more greatly increases the risk of youth offending among females, as compared with males (Doyle, 2007; Goodkind et al., 2013; Jonson-Reid & Barth, 2000; Malvaso et al., 2017a; Ryan et al., 2010; Ryan & Testa, 2005). For instance, Goodkind et al. (2013) found that out-of-home placement more than doubled the predicted probability of juvenile justice involvement for girls, while only increasing the predicted probability for boys by 23%. Ryan and Testa (2005) similarly found that any OOHC placement increased the risk of delinquency for girls, while for boys, it was placement instability (particularly having 3 or more placements) that was associated with increased risk of delinquency, rather than any initial OOHC placement.
A greater effect of foster and residential care (also termed congregate or group home care) on juvenile justice involvement of girls as compared with boys was also evident in multiple studies (Goodkind et al., 2013; Malvaso et al., 2017a). Malvaso and colleagues (2017a) found that experiencing one or more residential care placements resulted in 7.78 times greater odds of convictions for females compared to 4.99 times greater odds of convictions for males.
Gender differences in non-child welfare moderators and outcomes
There were 41 publications (22.8%) that examined gender differences in non-child welfare moderators and outcomes among maltreated children who offended. For example, descriptive studies identified various gender differences in the support needs of crossover children, including a greater prevalence of self-harm, suicidal ideation and attempts, running away, and risk of sexual exploitation among crossover females, and greater prevalence of neurodevelopmental and intellectual disabilities, ‘challenging behaviour’, and school exclusion among crossover males (Baidawi & Piquero, 2021; Baidawi & Sheehan, 2019, 2020a; Halemba et al., 2004). Other studies utilizing both child welfare and community samples have examined whether gender differences in these and other non-child welfare factors may moderate the association between child maltreatment/child welfare involvement and youth offending. Examined factors include socioeconomic status, school and educational factors, substance abuse and treatment, emotional or behavioural difficulties, intelligence, caregiver closeness, parent characteristics and resources and early parenthood (see, e.g. Bright & Jonson-Reid, 2008; Crooks et al., 2007; Goodkind et al., 2013; He et al., 2019; Kim, 2005; van der Put et al., 2014; Zingraff et al., 1994).
One interesting area concerns the relative importance of school-related factors in moderating the maltreatment-delinquency association for girls and boys. Several studies have found that school and education-related factors, including school performance or achievement (O’Connor, 1994; van der Put et al., 2014; Zingraff et al., 1994), and attendance (He et al., 2019), more greatly impacted the offending of maltreated boys, than that of girls. Conversely, He et al. (2019) also reported that school enrolment in year 7 significantly decreased the hazard of offending for girls, but not for boys, while other research found no gender differences in the impact of school safety (Crooks et al., 2007), or school disengagement (Bender, 2008), on offending among maltreated youth.
The impacts of gender, offending and maltreatment have also been examined in relation to other outcomes including future orientation (Craig, 2019), self-control and impulsivity (Meldrum et al., 2020), ‘adolescent problem behaviours’ such as school difficulties and drop out (Perez, 2017), gang involvement (Wolff et al., 2020), running away (Byrne, 2014) and reflective functioning (Protic et al., 2020). In each study, gender differences persisted in the examined outcome after accounting for the influence of exposure to maltreatment and/or ACEs.
Gender differences in mental health and substance abuse among maltreated youth in the juvenile justice system have also been analysed. Bassett (2013) found a correlation between ACE history and substance abuse disorder among juvenile justice-involved females, but not males, while Armstrong and colleagues (2018) theorized that higher levels of substance abuse among females who perpetrated child-to-parent violence were reflective of their more extensive victimization histories relative to males. This is supported by research by Fore (2018) who found that gender was not a significant predictor of illicit drug abuse referrals after accounting for ACEs in a juvenile justice sample. Yet positive associations have been observed between ACE or maltreatment exposure, and mental health problems among justice-involved youth of both genders (Baglivio et al., 2017; Kowalski, 2019). Some studies found gender differences in psychopathology among justice-involved youth following maltreatment exposure and/or child protection involvement (Baglivio et al., 2017; Bhatta et al., 2014; Farina et al., 2018; King et al., 2011; Wareham & Dembo, 2007). For instance, Wareham and Dembo (2007) reported that among male youths, physical abuse was a more salient factor in the long-term experience of psychological problems, whereas family member alcohol, other drug use, mental health problems and sexual victimization had a more adverse effect on the psychological functioning of the female youths. Studies have also examined specific mental health diagnoses by gender and maltreatment among juvenile justice samples, including psychopathy (Farina et al., 2018) (more strongly associated with physical and emotional abuse among girls), anxiety, affective disorders and ADHD (King et al., 2011), and anxiety and depression (Nowakowski-Sims & Rowe, 2017).
Suicidality and self-harm among maltreated and justice-involved children have also been examined across several studies, with findings highlighting the greater odds of suicidal ideation (Kretschmar & Flannery, 2011; Logan-Greene et al., 2017), and suicide attempts (e.g., Perez, 2017) among juvenile justice-involved females, as well as increased odds of suicide attempts among those with higher ACE scores. Perez (2017) theorized that the odds of both serious delinquency and suicidal behaviour are increased by higher ACE scores; however males tend to externalize the impacts of such exposures via serious and violent delinquency, while females tended to internalize via suicidal behaviour. Among a child welfare sample, Bjorkenstam et al. (2013) found that more convictions predicted a higher prevalence of suicidal behaviour, which was also more evident among females, after controlling for a range of variables. Finally, gender differences in the nature of service system contact among crossover youth was identified in some research, for example, in relation to homelessness services (AIHW, 2016), as well as drug and alcohol services, police sexual abuse services and secure welfare services (Baidawi & Sheehan, 2019, 2020a).
Discussion
This scoping review aimed to provide an overview of research concerning gender differences in the maltreatment-youth offending relationship. Across four databases, 180 publications meeting inclusion criteria were evenly divided between those drawing on juvenile justice and child welfare samples, and those which employed cross-sectional and longitudinal designs. The included studies were disproportionately from the US, and the majority were conducted since 2010, likely reflecting increased attention to gender differences in the context of rising juvenile justice involvement of girls (Pusch & Holtfreter, 2018; Shepherd et al., 2013a). Somewhat surprising were the small number of UK-based studies, and that no qualitative studies were identified for inclusion. These characteristics may reflect the smaller number of justice-involved females in some regions, and associated challenges undertaking quantitative gendered analyses in these areas, as outlined by Rodway et al. (2011).
Summary Implications for Future Research.
The review findings highlight several gender differences and other pertinent findings concerning the characteristics, maltreatment experiences, child protection pathways and juvenile offending of maltreated boys and girls. First, a greater prevalence of child welfare involvement and maltreatment exposure was evident among justice-involved girls relative to boys, most notably in relation to sexual abuse and multi-type maltreatment (e.g., Baglivio & Epps, 2016; Malvaso et al., 2019). Yet, the few studies that included relevant validity measures consistently identified a greater likelihood of under-reporting of maltreatment by justice-involved males (Kenny et al., 2007; Moore et al., 2013; Protic et al., 2020). Such findings reflect clinical and research evidence regarding the lower rates of maltreatment disclosure by males, particularly in relation to sexual abuse (Lev-Wiesel et al., 2019; O’Leary & Barber, 2008). Given these findings, it remains unclear how under-reporting of maltreatment (particularly by males) may have influenced the conclusions drawn by the body of research examining the maltreatment-youth offending relationship. Beyond the impact on research evidence, the current findings point to the policy and practice importance of including measures of under-reporting in child maltreatment screening and assessment instruments, for instance, those utilized by juvenile justice and child welfare practitioners. Key implications of the review for future research are summarized in Table 3, including the usefulness of future research that explores justice-involved youths’ willingness and/or reluctance to disclose childhood maltreatment. Any such research should pay attention to differences across dimensions known to be related to under-reporting and non-disclosure, such as gender, race/ethnicity, age, maltreatment dimensions and other individual characteristics (e.g. mental health and cognitive impairment status) (Lev-Wiesel et al., 2019). Additionally, future research should consider triangulation of multiple maltreatment data sources (e.g. self-report and official data) given limitations associated with each, alongside research suggesting that studies based on prospective/official data versus retrospective self-report measures may in fact be identifying different underlying populations (Baldwin et al., 2019). Aside from addressing under-reporting of maltreatment, the review also draws attention to the importance of considering intersectionality, particularly in terms of race and ethnicity differences in the maltreatment-offending association. The review found that gender x race differences are recognizable in several areas, including maltreatment prevalence and differential mechanisms linking maltreatment, child welfare involvement and youth offending (see, e.g. Goodkind et al., 2013; Zettler et al., 2018).
Reflecting general trends in the broader youth population, the findings consistently identified that males who are maltreated were more likely to offend as youth and to exhibit violent offending and recidivism, compared with females who are maltreated (e.g., Cho et al., 2019). At the same time, there were mixed findings regarding the extent to which gender moderated the maltreatment-offending relationship, that is, whether or not child abuse and neglect exert different impacts on the likelihood of juvenile offending (including violent offending and recidivism) among girls and boys. This forms another important area for future examination via meta-analyses and/or well-designed studies that can overcome the aforementioned issue of maltreatment under-reporting. Such studies should address the question of differential impact of maltreatment by gender, undertaking analyses based on maltreatment type and severity, as well as in relation to diverse juvenile offending outcomes (e.g. any offending and offending severity/chronicity, including recidivism).
In contrast to studies which examine the impact of maltreatment, one fairly consistent finding was that child welfare intervention, particularly OOHC placement in foster or residential care, more greatly increased the likelihood of youth convictions among females relative to males (Jonson-Reid & Barth, 2000; Malvaso et al., 2017a). Likewise, the apparently greater importance of school performance in moderating the maltreatment to delinquency pathway among boys also deserves further research attention (e.g. Zingraff et al., 1994). While also requiring more detailed examination in future, such findings should be of the utmost interest to child welfare policy-makers and practitioners. Each points to the potential usefulness of gender-based strategies, underpinned by research evidence, which may reduce offending and juvenile justice system involvement among maltreated girls and boys. While the required evidence to systematically inform gender-responsive interventions remains absent in the youth space, gender-based approaches for reducing offending and justice system involvement do exist for adult populations. One example is the Beyond Violence manualized curriculum for justice-involved women with histories of violence, which gives attention to both violent victimization and perpetration in aiming to prevent future violent experiences among this cohort, with promising results (Covington, 2013; Kubiak et al., 2016).
Finally, the review described the support needs of justice-involved children who are maltreated or child welfare-involved, including the greater prevalence of neurodisability among males (Baidawi & Piquero, 2021), and generally greater mental health needs, particularly in relation to suicidality and self-harm, among females (Logan-Greene et al., 2017; Perez, 2017). Understanding these differences can support targeted service planning across both child welfare and juvenile justice systems.
Despite its usefulness, the review has three key limitations, which should be borne in mind when interpreting its conclusions. First, several of the included publications have drawn on the same or similar datasets. As these publications often addressed slightly different research questions, or utilized amended forms of the dataset (e.g. additional waves of longitudinal data and examination of different data subsets), studies were not pooled for the presentation of data in the supplementary tables, nor in the quantification of specific findings presented in the results. A second and related limitation is the absence of a quality appraisal, which while not performed due to resource limitations, may be more easily undertaken for smaller-scale systematic reviews deriving from these analyses. Finally, the restriction of gender to a binary in the majority of available research means that a key research gap exists concerning the experiences and trajectories of gender diverse children and young people who experience maltreatment and justice system involvement.
Conclusion
Due to the seriousness of youth offending for both children and the broader community, identifying strategies to prevent and respond to offending and justice system involvement of maltreated youths remains a key research and policy priority. Drawing together a fragmented body of evidence in these areas, this scoping review highlighted studies that have examined gender differences in the characteristics, maltreatment experiences, child protection pathways and offending of maltreated youth. The findings highlight a greater prevalence of child welfare involvement and maltreatment exposure among justice-involved girls relative to boys, alongside the greater propensity of maltreated boys to offend relative to maltreated girls. While questions remain as to the extent to which gender moderates the maltreatment-offending relationship, it appears that child welfare intervention, particularly OOHC placement, more greatly increases the likelihood of youth convictions among girls, while school performance appears be a more important factor in the maltreatment-offending pathway among boys. Future research should aim to overcome limitations brought about by youth under-reporting experiences of maltreatment, to examine the maltreamtent-offending trajectories of gender diverse youth and to assess whether the quality of available evidence is sufficient to support gender-informed policy and practice recommendations addressing the maltreatment-youth offending relationship.
Supplemental Material
sj-pdf-1-tva-10.1177_15248380211052106 – Supplemental Material for Gender Differences in the Maltreatment-Youth Offending Relationship: A Scoping Review
Supplemental Material, sj-pdf-1-tva-10.1177_15248380211052106 for Gender Differences in the Maltreatment-Youth Offending Relationship: A Scoping Review by Susan Baidawi, Nina Papalia, and Rebecca Featherston in Trauma, Violence, & Abuse
Supplemental Material
sj-pdf-2-tva-10.1177_15248380211052106 – Supplemental Material for Gender Differences in the Maltreatment-Youth Offending Relationship: A Scoping Review
Supplemental Material, sj-pdf-2-tva-10.1177_15248380211052106 for Gender Differences in the Maltreatment-Youth Offending Relationship: A Scoping Review by Susan Baidawi, Nina Papalia, and Rebecca Featherston in Trauma, Violence, & Abuse
Footnotes
Acknowledgments
The authors wish to thank Gabby Lamb, Social Work Librarian at Monash University for assistance in developing database search translations.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This review was supported by a Discovery Early Career Researcher Award [DE190101104] Fellowship from the Australian Research Council.
Supplementary Material
Supplementary material for this article is available online.
Author Biographies
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
