Abstract
Current knowledge about the characteristics of adolescents involved in recidivist adolescent to parent violence offending remains limited. This study employed more than 50,000 linked administrative police (from birth) and health (from age five) data events to examine predictors of adolescent to parent violence recidivism in a geographically-distinct case series of 775 Australian adolescents. The predictive association between adverse childhood experiences, health and police involvement related characteristics and frequency of recidivism was found to vary by sex and level of exposure to parental intimate partner violence. Events occurring before an adolescent’s first offence, including sustained exposure to adverse childhood events and IPV exposure combined with sexual offence victimization, amplified the frequency of re-offending. Developmental life-course trajectories involving family violence verbal arguments, and other antisocial behaviors in mid to late adolescence, had a stronger predictive association with the frequency of re-offending. These results highlighted several key intervention points with adolescents and families across the life course.
Introduction
There is a global consensus that family violence, defined as violence between family members and within kinship ties, is a leading social, political, economic, and public health problem (AIHW, 2020). Violence or abusive behavior from an adolescent (aged 11–18 years) toward a parent or other unpaid caregiver is a specific form of family violence, referred to herein as adolescent to parent violence. Adolescent to parent violence is defined as a pattern of behavior intended to threaten, exert power or control, or cause damage and harm to a parent (Campbell et al., 2020; Cottrell & Monk, 2016; Holt & Retford, 2013). These acts can include physical violence, property damage, verbal abuse, coercive and controlling behaviors, and financial abuse (Campbell et al., 2020; Calvete et al., 2015; Fitz-Gibbon et al., 2018). Adolescent to parent violence is not uncommon and is thought to be underreported. Available data on prevalence rates range from 5 to 20% for physical violence (Crime Statistics Agency, 2019; Freeman, 2018; Moulds, Day, et al., 2019; Simmons et al., 2018), with rates of psychological violence reported to be much higher (33–93%; Simmons et al., 2018). This violence can have long term effects on the physical and emotional well-being of those involved and have a devastating impact on a parent-child relationship.
Prevalence and Patterns of Adolescent to Parent Recidivism
Despite evidence of strong links between childhood maltreatment and both adolescent to parent violence and juvenile offender recidivism (Barra et al., 2018), few studies have examined adolescent to parent violence recidivism. In Australia, Freeman's (2018) examination of the offending records of adolescents (boys: n = 688; girls n = 367) proceeded against by the police for a family violence related assault in 2014 found that 23% (n = 243) re-offended with another family violence assault offence within 12 months of the reference offence. Supporting these findings, a recent study of 4000 adolescents (aged 12–18 years) reported to police for their involvement in a domestic and/or family violence incident found approximately one in four adolescents were involved in a repeat offence within 6 months (Boxall & Morgan, 2020). The proportion of adolescents involved in recidivist offending was slightly higher for adolescents reported for a family violence offence (28%, n = 904) compared to an offence against a partner (domestic violence offence; 26%, n = 193). Spanish studies directly examining adolescent to parent family violence report recidivism rates around 35% (Cuervo et al., 2017; Maroto & Cortés, 2018). However, caution has been advised when comparing cross country rates of juvenile recidivism due to legislative differences relating to the age of criminal responsibility (Fazel & Wolf, 2015).
Studies of re-offending patterns amongst adult recidivist domestic and family violence offenders have linked the timing between offences to the frequency (Boxall & Morgan, 2020; Barnham et al., 2017; Morgan et al., 2018) and severity of re-offending (Kerr et al., 2017). For adolescents, Freeman (2018) found that almost half re-offended with another family violence offence within 90 days of the reference offence. Boxall and Morgan (2020) found that the risk of repeat violence for adolescents involved in family violence offences was highest between 19 and 31 days after the reference offence. Adolescents who re-offended in this period were at greater risk of a third offence than those recording a longer time period between repeat offences. This finding is consistent with prior studies examining the short-term re-offending patterns of adult domestic and family violence offenders (Morgan et al., 2018; Richards et al., 2014; Stansfield & Williams, 2014). These findings highlight the importance of early identification of at-risk repeat offenders and victims to prevent both short-term and longer-term repeat domestic and family violence (Boxall & Morgan, 2020).
Evidencing Risk Factors for Adolescent to Parent Violence Recidivism
Despite a substantial body of research examining risk factors and predictors of re-offending amongst adult domestic and family violence offenders (e.g. Coghlan & Millsteed, 2017; Dowling & Morgan, 2019; Millsteed & Coghlan, 2016), few studies have examined risk factors associated with adolescent to parent violence recidivism. Studies comparing repeat and one-time offending adolescents found involvement in the following risk factors significantly increased the probability of re-offending against a partner or family member (including parents); prior family violence and non-family violence related convictions (Uekert et al., 2006; Freeman, 2018, Boxall & Morgan, 2020), protection order breaches (Boxall & Morgan, 2020; Freeman, 2018; Uekert et al., 2006), committing offences against a female victim (Uekert et al., 2006), a history of abuse or exposure to parental intimate partner violence (IPV) as a child (Uekert et al., 2006) and exposure to a greater number of combined adverse childhood experiences (ACE’s; Craig & Zettler, 2021).
Mixed findings are reported when examining the predictive association between mental illness and family violence re-offending. Uekert et al. (2006) found high rates (60%, n = 245) of mental illness diagnoses among adolescents who attended court in response to committing a domestic or family violence offence in the United States of America in 2002. However, a significant predictive association between a mental illness diagnosis and family violence re-offending was not evidenced. When examining the association between returning to juvenile detention in response to adolescent to parent family violence re-offending, Méndez and Tomás (2018) found that adolescents with a mental disorder diagnosis were more likely to receive another sentence than adolescents without a diagnosis. However, neither study reported the type of mental disorders experienced by adolescents.
Limitations within the Current Body of Adolescent to Parent Recidivism Risk Factor Research
An important limitation of the current family violence recidivism research is the absence of a specific focus upon adolescent to parent violence (e.g. Freeman, 2018; Uekert et al., 2006) and the failure to define the nature of domestic violence offences included (e.g. Craig & Zettler, 2021). Research evidence suggests that the risk factor profiles of adolescents involved in adolescent to parent violence may differ from adolescents involved in other forms of family violence and non-family violence offending behaviors (Contreras & Cano, 2015; Ibabe et al., 2014; Kennedy et al., 2010; Nowakowski-Sims, 2019; Palanques et al., 2021). Nowakowski-Sims (2019) explored childhood adversity and delinquency in two cohorts of adolescents, one involved in adolescent to parent violence and the other sibling-to-sibling violence. Adolescents involved in adolescent to parent violence reported a higher prevalence of childhood adversity, particularly maltreatment and witnessing family conflict, and cumulative exposure to ACE’s than adolescents involved in sibling-to-sibling violence. In contrast, proportionately more adolescents involved in sibling-to-sibling violence had prior arrests. In addition, they reported higher rates of involvement in substance use and fighting at school than adolescents involved in adolescent to parent violence (Nowakowski-Sims, 2019).
Furthermore, research evidence suggests that the psychological profiles of adolescents involved in violence towards a parent differ from adolescents involved in other types of offending behavior and also non-offenders (Contreras & Cano, 2015; Ibabe et al., 2014; Kennedy et al., 2010). Adolescents who are violent towards a parent have been found to have higher rates of clinical diagnoses (Contreras & Cano, 2015), including depressive symptoms and hyperactivity (Ibabe et al., 2014). In addition, they were more likely to be psychiatrically hospitalized, have attempted suicide, and be medicated (Kennedy et al., 2010) than adolescents involved in other non-family violence offences. These findings suggest that to develop the adolescent to parent violence evidence base, studies are needed that focus specifically on adolescent to parent violence.
Within the field of criminology, an extensive body of research has compared cohorts sourced from within the same populations based on offending behavior, offence types, frequency of offending and individual characteristics of the offender (Boxall & Sabol, 2021; Moffitt, 1993). Studies conducting comparisons of adolescents involved in adolescent to parent violence offending are less common, despite evidence suggesting that adolescents who are violent or abusive towards a parent form a heterogeneous group (Ibabe, 2020; Contreras & Cano, 2014). Studies conducted to date have predominantly compared cohorts of adolescent to parent violence offenders based on sex (Simmons et al., 2018). Several studies employing justice samples have evidenced differences in the age, health, behavioral and prior offending characteristics of boys and girls involved in adolescent to parent violence (Armstrong et al., 2018; Kennedy et al., 2010; Walsh & Krienert, 2007). Another smaller body of work has explored differences associated with the type of offence committed and the nature of offending (Beckmann et al., 2021; Calvete et al., 2013; Ibabe, 2020; Moulds, Mayshak, et al., 2019). Building the evidence base through conducting comparisons of adolescents involved in adolescent to parent violence may assist in developing a more nuanced understanding of the factors that place adolescents and parents at risk of reoccurring violence. This research has important implications concerning early intervention and prevention of re-offending, including developing or enhancing adolescent to parent violence specific risk assessment tools to assist in identifying both adolescents and families at risk of ongoing violence (Boxall & Morgan, 2020; McManus et al., 2017).
The Current Study
This study employed data linkage methods, and life-course data sourced from two Australian Government agencies using a geographically-distinct sample to investigate the characteristics of adolescents proceeded against by the police for adolescent to parent violence offending and re-offending. Within the body of research examining adolescent to parent violence to date, only a few studies have utilized longitudinal data to investigate risk factors and predictors of adolescent to parent violence (see Arias Rivera & Hidalgo García, 2020 for a review). An even smaller number of studies have employed data from early childhood (see Pagani et al., 2003, 2004, 2009 for notable exceptions). Through the linkage of police and health record data, from birth to age 18 years, and the analysis of a range of risk factors, this study builds on the findings of prior studies investigating recidivist family violence offending (Boxall & Morgan, 2020; Craig & Zettler, 2021; Uekert et al., 2006) and provides a significant advancement to the current body of research evidence.
This study also builds on the findings of previous risk factor investigations (e.g. Armstrong et al., 2018; Beckmann et al., 2021; Calvete et al., 2020) by comparing childhood risk factors associated with adolescent to parent violence re-offending among adolescents exposed to police reported parental IPV in childhood to those with no recorded parental IPV exposure. A large body of research has explored the link between childhood exposure to parental IPV and negative outcomes in adolescence and early adulthood (Vu et al., 2016). Additionally, evidence suggests that many adolescents involved in adolescent to parent violence have witnessed violence between parents (Beckmann et al., 2021). However, to date, only one study (Aguilar Ruiz & Pereda, 2021) has explored differences in the risk factor profiles of males exposed to parental IPV to those with no recorded exposure and the association with non-offence specific adolescent recidivism.
Overall, the current study aimed to investigate life course predictors of adolescent to parent violence re-offending in a case series of adolescents issued a legal action by the New South Wales Police Force (NSWPF) for an adolescent to parent violence offence. It also aimed to examine whether these predictors varied by sex and level of exposure to police reported parental IPV.
Method
A retrospective case series study design was employed to investigate childhood life-course predictors associated with adolescent to parent violence recidivism. This type of case study is an observational study that includes only subjects known to have experienced a common outcome (Dekkers et al., 2012). Seven hundred and seventy-five adolescents (535 boys and 240 girls) issued a legal action by the NSWPF for an adolescent to parent violence offence from July 2008 to June 2018 were included in the case series. These adolescents were born between 1994 and 2006 and were aged between 11 and 18 years at the time of their first adolescent to parent violence offence. They resided in a large geographical area of NSW in Australia.
Definitions
Adolescent to parent violence offences included all police recorded domestic violence related assault, malicious damage, offence against another person, or sexual assault incidents, with a legal action issued, in which a parent or responsible caregiver was recorded as being an involved victim.
Data sources and Data Linkage
Data were drawn from the NSWPF Computerized Operational Policing System (COPS; 1995–2018) as well as three NSW Ministry of Health Data Collections: (1) the Admitted Patients Data Collection (APDC; 2001–2018), (2) the Emergency Department Data Collection (EDDC; 2005–2018), and (3) the Mental Health Ambulatory Data Collection (MH-AMB; 2006–2018). The data requested was determined by reviewing the research evidence on risk factors for adolescent to parent violence (see Peck et al., 2021). The earliest available records were sourced depending on data availability.
The Centre for Health Record Linkage (CHeReL) provided the data linkage service for this study. Probabilistic linkage methods using matching variables of name, date of birth, residential address and sex were used to link the NSWPF identified case series to NSW Ministry of Health data holdings. The CHeReL conducts data linkage with an estimated false positive linkage rate of <0.5% (CHeReL, 2021). In accordance with nationally legislated privacy protocols, all identifying data are removed and replaced with a unique project person number (PPN) upon dissemination. This study was conducted with ethical approval from the Southern Cross University Human Research Ethics Committee (ECN-18–198) and the NSW Population and Health Services Research Ethics Committee (PHSREC 2019ETH00173). Approvals were granted by the NSWPF and NSW Ministry of Health data custodians.
Measures
Re-offending Outcome Variable
An adolescent’s first recorded adolescent to parent violence offence with a legal action issued was the index offence. For each adolescent the number of adolescent to parent violence related offences, committed in the 4 years following the index offence, was calculated as a measure of recidivism.
Predictor Variables
In addition to sex and age, the relationship between 18 predictor variables and adolescent to parent violence re-offending was investigated at Time 1 (before the first adolescent to parent violence offence) and Time 2 (before age 19 years).
Adverse Childhood Experiences
Five predictor variables formed the ACE category. Two factor variables were created. Factor 1 consisted of three items: NSWPF recorded parental alcohol or drug use, NSWPF or NSW Health recorded neglect and NSWPF recorded criminal environment child at-risk. Factor 2 consisted of four items: NSWPF reported parental IPV before age 13 years; victim of a family violence related assault; victim of a family violence related verbal argument; and NSWPF recorded physical abuse child at-risk. Two other variables in this category were dichotomous measures coded as 0 for no involvement or one for one or more involvements. They included being the victim of a sexual offence or being the victim of a non-family violence related assault incident. The final variable, duration of ACE exposure, was a count of the number of years between an adolescent’s first and last recorded victimization or child-at-risk involvement before age 19.
Health Related
Adolescents involvement in drug use, alcohol use or receiving a head trauma diagnosis formed the health related category. All three variables were dichotomous measures coded as 0 for no involvement or one for one or more involvements. Drug use included being involved in a drug related NSWPF incident as a person of interest (POI) 1 , being issued a legal action for a drug related offence or receiving a NSW Health drug related diagnosis. Alcohol use involved being recorded as a person of interest in a NSWPF alcohol related incident 2 or receiving a NSW Health alcohol related diagnosis.
Mental Health Related
Three variables formed the mental health/behavioral disorder category. All three variables were dichotomous measures coded as 0 for no diagnosis or one for receiving a diagnosis. Diagnostic categories relating to anxiety, mood, personality, psychotic and substance use related disorders were derived from hospital admission and outpatient presentations and combined into an anxiety/mood/other mental health disorder variable. Diagnostic categories relating to neurodevelopmental, behavioral or conduct related disorder diagnoses were combined into a developmental/conduct disorder variable. An attempt suicide/self-harm variable included receiving a NSW Health diagnosis for attempt suicide, self-harming behaviors and suicidal ideation, as well as being recorded in an attempt suicide incident by the NSWPF.
Adverse Police Involvement
Adverse police involvement was defined by five dichotomous variables (coded as 0 for no involvement or one for one or more involvements). The National Offence Index (NOI) ranking (Australian Bureau of Statistics, 2018) was used to classify and rank all family violence and non-family violence related legal actions in the dataset scale of 1 (most serious) to 185 (least serious). Property offence and violent offence variables included being issued a legal action by the NSWPF for a property crime related offence (NOI rankings 43, 49, 64, 70, 77, 79–86, 96, 97, 146, 150) or a non-family violence offence involving an act of violence (NOI ranking 1–18, 24–33, 139, 145 and 157). The person search/move on variable and family violence verbal argument POI variables included being recorded as a POI or person named in a NSWPF person search or move on incident or domestic violence episode (verbal argument). The apprehended violence order (AVO) variable included being recorded as a defendant of a family violence related AVO.
Statistical Analysis
A series of bivariate chi-square tests explored differences in the proportion of adolescents involved in each predictor variable by re-offending status, sex and level of exposure to NSWPF reported parental IPV. Negative binomial logistic regression was then used to model predictors of adolescent to parent violence re-offending for each group of adolescents. Negative binomial regression analysis techniques are suited to the analysis of outcome variables that include a count of incidents or events occurring within a certain period (Gardner et al., 1995). As not all adolescents had 4 years of data available after their index offence (36%, n = 128; <3 years of data: 23%, n = 46; <2 years of data: 10%, n = 25; <1 year of data: 3%, n = 10), an offset variable was created to control for differences in the years of available data in which re-offending could occur (Gardner et al., 1995; Osborne, 2017). An offset variable was created by calculating the natural log of the number of recidivist offences for each adolescent and included in each negative binomial regression model as an additional predictor variable (Crowson, 2019), with the rate of re-offending entered as the outcome variable. Assumptions of regression analysis were tested before conducting the analysis. For the final set of predictor variables, the ratio of adolescents to variables was found to be adequate, and tests for multicollinearity did not find any variables that were highly correlated (Tabachnick, Fidell, & Ullman, 2013). Furthermore, no unusually large parameter estimates or standard errors were observed across regression models (Tabachnick et al., 2013). As no specific hypotheses were formulated regarding the order or importance of the predictor variables, all variables were entered into each regression model simultaneously (Tabachnick et al., 2013). Data analysis was conducted using IBM SPSS version 27.0 (IBM Corp, 2020).
Results
Case Series Characteristics
Number of Adolescents by Sex, Level of Parental IPV Exposure and Frequency of Adolescent to Parent Violence Offending.
Note. % = percentage of the row total.
When examining the type of offences committed by adolescents involved in recidivist offending, over half of all adolescents (boys: 56%, n = 144; girls 64%, n = 60; adolescents exposed to parental IPV: 58%, n = 304; adolescents with no IPV exposure: 57%, n = 144) were issued legal actions for offences involving violence or serious violence. Adolescents involved in recidivist offending predominantly committed offences against a parent as well as other family members (65%, n = 229). For sole victim offenders, more mothers (27%, n = 95) were victims of offending from a child than fathers (8%, n = 27).
Proportion of Adolescents with One or More Recorded Involvements in Each Predictor Variable Before Age 19 by Re-offending Status, Sex and Level of Parental IPV Exposure.
Note: % = percentage of column n
Multivariate Results
Negative Binomial Logistic Regression Model Results for Rate of Adolescent to Parent Violence Recidivism for Boys and Girls.
Note: IRR = incident rate ratio; * p<.05, ** p <.01; CI = confidence interval; - variable not included in model.
Negative Binomial Logistic Regression Model Results for Rate of Adolescent to Parent Violence Recidivism for Adolescents Exposed and Not Exposed to NSWPF Reported Parental IPV.
Note: IRR = incident rate ratio; * p<.05, ** p <.01; CI = confidence interval; - variable not included in model.
Tables 3 and 4 show the unstandardized beta co-efficient, standard error, incident rate ratio and 95% confidence interval results for each predictor variable in the negative binomial logistic regression models for boys and girls (see Table 3) and adolescents exposed and not exposed to NSWPF reported parental IPV (see Table 4) at Time 1 and Time 2. When interpreting incident rate ratios (IRR) in the negative binomial regression models, incident rate ratios greater than one indicate that as the values on the predictor variable increased (i.e. from 0, no recorded involvement, to 1, one or more recorded involvements), there was one unit increase in the number adolescent to parent violence offences recorded in the 4 years following the initial offence (Gardner et al., 1995; Osborne, 2017).
Sex specific findings: Sex differences are evident across each regression model. For boys, being younger at the time of their first adolescent to parent violence offence (IRR = .87, p = <.001, 95% CI [.8–.9]) and having a longer duration between their first and last recorded ACE exposure (IRR = 1.04, p = .004, 95% CI [1.0–1.1]) at Time 1, were found to have a significant predictive association with the rate of re-offending. For girls, additional variables including: • involvement in property crime offending (IRR = 2.12, p = .005, 95% CI [1.3–3.6]), or • being a victim of a sexual offence (IRR = 1.75, p = .021, 95% CI [1.1–2.8])
before their first family violence offence were found to have a significant association with re-offending at a higher rate. Consequently, for girls being the victim of a non-family violence offence was associated with adolescent to parent violence re-offending at a lower rate (IRR = .59, p = .021, 95% CI [1.1–2.8]) when compared to girls who were not the victim of a non-family violence offence.
NSWPF reported parental IPV exposure related findings: For adolescents with no recorded NSWPF reported parental IPV exposure, receiving a diagnosis for anxiety, mood or other mental health disorder (IRR = 1.8, p = .044, 95% CI [1.1–3.2]) at Time 1 and also Time 2 (IRR = 1.97, p = .017, 95% CI [1.1–3.4]) almost doubled the odds of re-offending at a higher rate when compared to adolescents without a mental health related diagnosis. For adolescents exposed to NSWPF reported parental IPV, a reliable association was found between duration of exposure to ACE’s (IRR = 1.04, p = .016, 95% CI [1.0–1.1]) and being the victim of a sexual offence (IRR = 1.40, p = .044, 95% CI [1.0–1.9]) prior their first adolescent to parent violence offence and re-offending at a higher rate.
Age related findings: When examining involvement at Time 1, boys (both exposed and not exposed to NSWPF reported parental IPV) were found to re-offend at a higher rate than girls (see Table 4). For adolescents exposed to NSWPF reported IPV, gender was not a significant predictor of re-offending at a higher rate when examining involvement at Time 2.
A predictive association between drug use and the rate of adolescent to parent violence recidivism was evidenced for boys and adolescents exposed to parental IPV. Involvement in drug use prior to age 19 years increased the odds of re-offending at a higher rate (Boys: IRR = 1.60, p = .007, 95% CI [1.1–2.3]; IPV Exposure: IRR = 1.67, p = .003, 95% CI [1.2–2.3]) when compared to adolescents not recorded for drug use.
A consistent pattern of offending was evident when examining the association between adolescents' involvement in violent offences against non-family members before age 19 and the rate of re-offending against parents. This predictive association was strongest for adolescents not exposed to NSWPF reported parental IPV. For these adolescents, involvement in a violent offence against a non-family member before age 19 years almost doubled the odds of re-offending against a parent at a higher rate (IRR = 1.89, p = .031, 95% CI [1.1–3.3]) when compared to adolescents who did not commit a violent offence against a non-family member. A reliable predictive association was also found between violent offending against non-family members and an increased rate of family violence recidivism for all boys (IRR = 1.38, p = .024, 95% CI [1.0–1.8]).
Ongoing involvement as a POI in family violence related verbal arguments up to age 19 years had a strong predictive association with an increased rate of adolescent to parent violence recidivism for all groups in the case series. The relationship was strongest for adolescents not exposed to NSWPF reported IPV, with involvement in family violence related verbal arguments tripling the odds of re-offending at a higher rate (IRR = 3.27, p <.001, 95% CI [2.0–5.4]) when compared to cases not recorded for involvement in verbal arguments.
Discussion
This study investigated ACE, health, and police involvement related predictors of adolescent to parent violence re-offending. It examined whether these predictors varied by sex and level of exposure to police reported parental IPV. In contrast to previous studies that used an adolescent’s second or third family violence offence as the reference offence (Boxall & Morgan, 2020; Uekert et al., 2006), this study used each adolescent’s first offence. As prior family violence offending was found to be a significant risk factor for re-offending in previous studies (Boxall & Morgan, 2020; Uekert et al., 2006), this approach allowed for the association between risk factors occurring before adolescent’s involvement in family violence offending to be investigated without the influence of prior adolescent to parent violence offending. Employing this approach provided valuable and unique insight into the association between specific types of childhood adversity, emerging mental health issues, adolescent antisocial behaviors, and increased rates of adolescent to parent violence offending. To our knowledge, this is the first study to compare life-course predictors of adolescent to parent violence recidivism among adolescents exposed and not exposed to parental IPV.
In the current study, almost 70% (n = 523, 67%) of adolescents had been exposed to police reported parental IPV before age 13 years (with 28% (n = 220) the victim of a family violence related verbal argument and 17% (n = 135) the victim of a family violence assault offence prior to their first adolescent to parent violence offence). Across the body of adolescent to parent violence research evidence, exposure to violence at home (see reviews by Armstrong et al., 2018 and Simmons et al., 2018) and difficult family circumstances (Cuervo, 2021; Maroto & Cortés, 2018; Moulds, Mayshak, et al., 2019) are factors that are consistently related to adolescent to parent family violence. Higher rates of exposure to parental IPV and family violence related victimization have also been found amongst adolescent to parent violence offending groups when compared to adolescents involved in other types of offending (Contreras & Cano, 2015; Cuervo, 2021). The findings of the current study offer additional support for theoretical explanations suggesting an intergenerational association between children witnessing or experiencing family violence and then using violence against parents, other family members and in future intimate partner relationship (Harbin & Madden, 1979; Holt & Retford, 2013; Kaufman-Parks et al., 2018; Mahoney & Donnelly, 2000; Widom, 1989).
Interestingly, however, the current study found that exposure to parental IPV and other family violence related victimization was not a significant predictor of the frequency of adolescent to parent violence re-offending. By examining the combined frequency of IPV exposure and family violence victimization, across a multi-year period, the current study aimed to advance evidence within the IPV exposure research field, which to date has relied heavily on the use of self-report data and binary yes/no measures to elucidate the prevalence of parental IPV exposure in early childhood (Holmes et al., 2022; Margolin, 2005). The failure to find a significant association between parental IPV exposure and adolescent to parent violence recidivism is consistent with several studies exploring the predictive association between family violence exposure and general juvenile recidivism (Aguilar Ruiz & Pereda, 2021; Ferguson et al., 2009; Shaw, 2019) but not supportive of others where a predictive association was established (García-España et al., 2011; Holmes et al., 2015). The current study’s findings, combined with the findings from previous research, suggest that despite an established association between IPV exposure, general delinquency and using violence against a parent, the relationship between IPV exposure and recidivism remains less clear (Aguilar Ruiz & Pereda, 2021; Shaw, 2019). It is possible that the predictive association between childhood experiences of parental IPV and adolescent to parent recidivist offending (and other adverse outcomes) may depend on the nature, severity and timing of parental IPV exposure experienced. This association warrants further exploration as it is yet to be investigated in detail across the body of work examining exposure to parental IPV (Aguilar Ruiz & Pereda, 2021; Holmes et al., 2015; Vu et al., 2016).
Childhood exposure to parental IPV has been found to affect children differently depending on the accumulation of other risk and/or protective factors (Farrington et al., 2017; Jolliffe et al., 2017; Noble-Carr et al., 2020). Uekert et al. (2006) found a significant predictive association between re-arrest for a domestic or family violence offence and exposure to parental IPV when IPV exposure was combined with the experience of any abuse as a child (OR = 2.1, p = 0.032). In the current study, adolescents experiencing both parental IPV exposure and sexual offence victimization were 40% more likely to re-offend at a higher rate than adolescents recorded for IPV exposure alone. A recent study conducted by (Baglivio et al., 2021) compared the association between cumulative and individual exposure to ACE’s and the types of victims targeted by a group of 28,579 violent juvenile offenders. This study found a predictive association between sexual abuse, experiencing family violence and household mental health and committing a violent offence against a family member. However, these variables did not predict offending against strangers, acquaintances, intimate partners or authority figures. Considering (Baglivio et al., 2021) findings and those of the current study, children exposed to parental IPV, who are also the victim of a sexual offence, should be considered a high risk group for family violence offending and re-offending.
A growing body of evidence has reported an association between a higher number of cumulative ACE’s and the increased likelihood of adolescents being re-arrested for violent, domestic violence or sexual assault offences (Craig & Zettler, 2021), re-offending within a short time frame (Wolff et al., 2015), being classified as serious, violent or chronic juvenile offenders (Fox, Perez, Cass et al., 2015) or having an early onset, chronic offending trajectory (Baglivio et al., 2015). Building on these findings, the current study used life-course data to calculate the duration of ACE exposure to explore the sustained effect of ACE’s (in contrast to cumulative exposure) on the rate of adolescent to parent violence recidivism. It was found that exposure to ACE’s over a longer period (from birth to an adolescent’s first adolescent to parent violence offence) was a reliable predictor of re-offending at a higher rate for boys, girls and adolescents exposed to parental IPV. These findings highlight the need to not only consider the number of ACE’s but also the duration over which they occur when attempting to identify children and families at risk of ongoing adolescent to parent violence.
The findings of this study stress the importance of early intervention with at-risk children and families. The current study found that 43% of adolescents in the case series were reported as being victims of a police incident or a child-at-risk before age three. Ongoing exposure to adverse experiences and abuse in childhood has been found to affect a child’s psychological functions and relational skills, including empathy and self-control (Basto-Pereira & Da Maia, 2019; Narvey et al., 2021) and the ability to form secure attachments with others (Sousa et al., 2011). Low empathy, in particular, has been linked to violent behavior and dysfunctional personal interactions (Basto-Pereira & Da Maia, 2019; Jolliffe & Farrington, 2004; Narvey et al., 2021). As such, this early life contact with police, especially for children exposed to parental IPV and sexual victimization, represents a crucial point of intervention to divert the path towards offending and other adverse outcomes in later childhood and adolescence.
Comparing predictors by the level of exposure to IPV highlighted that not all adolescents involved in adolescent to parent violence are from known violent households or experience adversity during childhood. A significantly higher proportion of adolescents involved in recidivist offending had been diagnosed with an anxiety, mood or other mental health disorder or a developmental or conduct disorder before age 19 than one-time offenders. However, receiving a diagnosis for an anxiety, mood or other mental health disorder was only predictive of re-offending at a higher rate for adolescents not exposed to parental IPV. Prior studies have found that depression (Kennedy et al., 2010), having a depressed state (Cuervo, 2021) or an antisocial personality (Aguilar Ruiz & Pereda, 2021; Cuervo, 2021) were predictive of adolescent to parent violence. However, no prior studies have explored the association between mental health diagnoses and adolescent to parent family violence offending amongst adolescents exposed and not exposed to parental IPV. The findings from this study have important implications for mental health service providers. Many adolescents, with no recorded exposure to parental IPV, were found to have contact with a mental health service well before committing their first adolescent to parent violence offence. Considering this, screening for adolescent to parent family violence at child and adolescent mental health services may assist in identifying at risk adolescents and families without prior police or child protective services involvement. This early identification and subsequent intervention efforts with these at-risk families and adolescents may help prevent future involvement in the criminal justice system.
The current study found that factors predictive of the frequency of adolescent to parent violence recidivism changed across the life-course. Maltreatment related factors (sexual abuse victimization and duration of ACE exposure), found to have a significant predictive association with adolescent to parent violence recidivism prior to an adolescents’ first offence, lost their predictive strength during mid to late adolescence in favor other antisocial behavior related factors. This finding is consistent with prior research examining life-course risk factors associated with adolescent re-offending in general (Aguilar Ruiz & Pereda, 2021; Blasco et al., 2014; Hoge et al., 2015) and a recent meta-analysis examining the association between child maltreatment (physical, sexual or psychological abuse) and adult IPV perpetration (Li et al., 2020). This meta-analysis of 63 studies found a significant association between child maltreatment and IPV perpetration. However, the effect size of this association was found to be relatively small, suggesting that the association between childhood maltreatment and adult IPV perpetration was not conclusive and influenced by other factors.
Previous studies have found mixed results when examining the association between substance use and adolescent to parent violence (see reviews conducted by Peck et al., 2021 and Simmons et al., 2018). The results of these previous studies were largely dependent on the methodology employed and whether substance use was captured via official records or adolescent or parent self-report data. By using official police and health records of drug use, this study established a predictive association between drug use and frequency of adolescent to parent violence offending for boys and adolescents exposed to NSWPF reported parental IPV. These findings suggest that the failure to find a significant association between drug use and family violence recidivism in the study conducted by Uekert et al. (2006) may have resulted from their use of a combined cohort of boys and girls, both exposed and not exposed to parental IPV. Craig et al. (2019) found a strong relationship between ACE’s and recidivism among adolescents involved in moderate to high levels of substance use. In the current study, drug-using adolescents exposed to parental IPV were 67% more likely to commit a higher number of adolescent to parent family violence offences than non-drug using IPV exposed adolescents. These findings highlight the need for multi-agency holistic, trauma informed early intervention efforts with drug-using youth, especially those with a history of parental IPV exposure in childhood.
Consistent with Moulds, Mayshak, et al. (2019) findings, the current study found that adolescent to parent violence offending in isolation was rare amongst adolescents who had committed a family violence offence against a parent. Building on the findings of previous studies (Uekert et al., 2006; Freeman, 2018; Boxall & Morgan, 2020), involvement in other non-family violence related offending was predictive of a higher rate of adolescent to parent family violence offending for girls (property offences), boys (violent offences) and adolescents not exposed to parental IPV (property and violent offences). Early involvement in violent offences was also found to predict re-offending for adolescents not exposed to parental IPV in the study conducted by Aguilar Ruiz and Pereda (2021). Interestingly, involvement in other offending was not predictive of re-offending at a higher rate for adolescents exposed to parental IPV. For these adolescent’s involvement in a police person search or move on incident was found to have a stronger predictive association with re-offending. Ongoing family conflict in the form of verbal arguments was predictive of re-offending at a higher rate across all groups of adolescents. The current study is the first known study to investigate NSWPF POI data for a case series of adolescents involved in violent offending against a parent. Identifying that almost half of all adolescents were involved in a family violence verbal argument incident as a POI prior to their first adolescent to parent family violence offence has important preventative implications. This contact with police represents a crucial point of early family violence specific targeted intervention with adolescents and their families to prevent ongoing and escalating incidents of violence.
Limitations
Several limitations should be recognized when interpreting the findings of this study. Sourcing data relating to ACE's and other victimization involvements from official police and health records has limitations. The detail and accuracy of the data in these databases is reliant on the officer or health official entering the data into the database (Stewart et al., 2015). Whether a police incident is classified as being drug or alcohol related is reliant on the discretion of the police officer recording the incident data. This needs to be kept in mind when interpreting the findings of studies using NSWPF data.
As many crimes are known to be underreported, the data recorded in police datasets are restricted to what a parent or adolescent reports to authorities (Langton et al., 2012; Piquero, et al., 2014). As such, it is possible that adolescents in the current study were more likely to be involved in frequent or serious offending against a parent than non-justice system involved adolescents. Additionally, adolescents involved in the justice system are also known to be an at-risk sample (Connolly, 2020). This may have limited the generalizability of the studies findings to samples in which lower rates childhood adversity are reported. The current study was also unable to examine the association between familial factors such as parental criminal history, incarceration, parenting styles or family structure and adolescent to parent violence recidivism. A recent study conducted by Connolly (2020) provides evidence to suggest that familial factors may contribute to an overestimation of the association between ACE’s and anti-social behavior. Considering these findings, further research is needed within the adolescent to parent violence field to explore the interaction between familial factors, ACE’s, anti-social behavior and adolescent to parent violence recidivism.
Longitudinal data collection in the current study was also limited by the data capture period for each database. Due to the time period in which electronic data was available across the NSW Ministry of Health data collections, only a small number of cases had available data from birth to age 5 years. As such, to maintain consistency in the findings reported across all cases, NSW Health data from birth to age five was excluded from this study. Furthermore, the data used in this study was limited to NSW based data holdings. Any interstate contact with police or health services would not be captured in this study. As all adolescents in the case series had a reported last known address in the NSW area, it is likely that the amount of missing data due to interstate police or health system involvement is minimal. Despite these limitations, the value of using linked multi-agency administrative data to investigate victimization and offending behavior across the life course is high and overcomes a number of methodological issues evident in traditional longitudinal studies (Stewart et al., 2015).
Conclusion
The current study found that many adolescents involved in repeat adolescent to parent violence offending have complex needs. Many of these adolescents have experienced adversity in childhood that manifests in challenging behaviors in adolescence (Johns et al., 2017). These findings provide further evidence to support the benefits of holistic multi-agency early intervention efforts with at risk families and the need for trauma-specific prevention programs (Fox et al., 2015). The findings from this study have important implications for policy and practice and highlight the need for adolescent to parent violence screening at child at adolescent mental health services. These screening and risk assessment tools need to be sex specific, age appropriate and consider family circumstances when attempting to identify at risk adolescents and families. The findings from this study also support the need for a response to early reported adolescent to parent verbal argument incidents. These responses need to take a whole family approach and address other issues such as childhood trauma, drug use, and the mental health needs of the adolescents and families involved to prevent both short-term and long-term re-offending.
Footnotes
Acknowledgments
The authors would like to acknowledge the NSW Ministry of Health and New South Wales Police Force for the provision of the data employed in this study and also the Centre for Health Record Linkage for their data linkage services.
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: The study was conducted under an industry funded scholarship. The finding bodies included the Department of Regional New South Wales, Mid North Coast Local Health District, the North Coast Local Health District and Southern Cross University.
Ethical approval
Ethical approval was received from Southern Cross University Human Research Ethics Committee (ECN-18-198) and the New South Wales Population and Health Research Ethics Committee (2019ETH00173).
Notes
Author biographies
Dr
