Abstract
Currently, there is limited empirically published Australian studies on effective rehabilitative programs in youth justice. This study used a mixed-methods approach to evaluate the Healthy Relationships program which was designed to challenge attitudes relating to intimate partner violence for male adolescents in detention in Western Australia. Paired group analyses compared pre and post measures of attitudes towards intimate partner violence and traditional gender roles and stereotypes for the total sample (N = 65). Given the significant overrepresentation of First Nations youth in detention, additional analyses were also run separately for this group (n = 51). Participation in the Healthy Relationships program was expected to reduce participant endorsement of attitudes supporting intimate partner violence and endorsement of attitudes supporting traditional gender roles and stereotypes. Results supported our expectations for the total sample and the First Nations youth sample, indicating significant improvements across all outcomes following program participation. Qualitative analysis of participant program feedback further supported the quantitative results and identified the importance of the therapeutic alliance and incentives-based learning approaches. The findings contribute to the “what works” literature and provide insights into factors that improve positive treatment outcomes for youth in detention. Suggestions for program improvements and future research directions are discussed.
Plain language summary
The purpose of this study was to evaluate the effectiveness of the Healthy Relationships program developed for adolescent male offenders in custody. Participants attitudes towards intimate partner violence and traditional gender roles and stereotypes were compared pre and post program completion, with a separate analysis conducted for First Nations youth given the significant overrepresentation of this population in custody and the need for culturally-appropriate programs. Results indicated significant reductions in attitudes supportive of intimate partner violence and attitudes endorsing gender roles and stereotypes for the total sample and First Nations youth sample, providing support for the effectiveness of the program. Participant feedback provided additional insights into the components of the program that made it more effective, including facilitator style of delivery and resources used. Our study contributes to the “what works” literature for youth offenders and offers unique insights into to what makes criminogenic programs effective from the perspective of participants. Though our sample was limited to Western Australian youth, the findings could be used to assist in the design of culturally safe and effective programs for male adolescents involved in the justice system. Suggestions for program improvements and future research directions are discussed.
A Note on Terminology
Within this article the term “First Nations” is used to refer to Aboriginal and Torres Strait Islander peoples. This is in recognition that Aboriginal and Torres Strait Islander peoples were the First people of Australia and are the oldest living culture in the world (Malaspinas et al., 2016).
Violence in intimate partner relationships is a significant social and public health concern that affects adolescents worldwide. The types of relationships individuals experience during adolescent years can have implications for development and functioning in adulthood. Healthy relationships, inclusive of intimacy and support, are associated with positive outcomes for life functioning and wellbeing (Benham-Clarke et al., 2023; Collins et al., 2009). Collins et al. (2009) found that healthy romantic relationships in adolescence (e.g., those that involved dating someone for longer than a month) can lead to commitment and higher quality relationships in early adulthood and quality romantic relationships were associated with positive self-worth. Kanksy and Allen (2018) conducted a longitudinal study on the impact of romantic relationships during adolescent years. They found that the quality of the relationship, as opposed to duration, predicted future mental health functioning. Specifically, higher support in the romantic relationships predicted fewer negative externalizing behaviors directed towards their partner, by age 26 (Kanksy & Allen, 2018). Similarly, Lambert et al. (2014) conducted a study using 8,679 high school students and found that even when accounting for negative factors, having strong social relationships—with peers, family, and at school, were positively correlated with adolescent happiness.
In contrast to healthy relationships are those characterized by severe conflict, leading to violence within relationships. Unhealthy relationships inclusive of low-quality interactions (e.g., antagonism), lower sense of control and lack of “authenticity” can lead to increased conflict and controlling behaviors (Benham-Clarke et al., 2023; Collins et al., 2009). Family violence, which includes any act of physical, psychological, or sexual violence, was reported as one of the most prevalent crime types in a study conducted in Australia (State of Victoria, 2016). The effects of family violence can extend indirectly to children, which can result in intergenerational trauma (Boxall et al., 2020). Adolescent family violence significantly impacts family members (who are the victims) and can increase risk of violent behavior in adulthood for the perpetrators (Boxall et al., 2020). Risk factors for adolescents perpetrating interpersonal violence include witnessing parent-to-parent violence, exposure to violence in the community, suffering abuse as a child, and experiencing mental health challenges (Boxall & Morgan, 2020; Ehrensaft et al., 2003). There are additional factors influencing antisocial behaviors and violence that are linked to an individual’s exosystem including socioeconomic status, ethnicity due to racism, being a child in care, and the neighborhood one grows up in (Benham-Clarke et al., 2023; Taquette & Monteiro, 2019).
Adolescent family violence significantly overlaps with adolescent dating violence (violence perpetrated against an intimate partner), and victimization. Global estimates indicate around one in five adolescents reported experiences of dating violence (Gracia-Leiva et al., 2019; Wincentak et al., 2017). Common forms of dating violence are physical, sexual, psychological, and verbal abuse (Daff et al., 2021; Wincentak et al., 2017). The type of violence can differ, with girls more likely to use physical violence, and boys more likely to use sexual violence, against their partners (Daff et al., 2021). Although dating violence perpetrated by both adolescent partners is more common than “unidirectional” dating violence; male-on-female violence continues to have a higher prevalence than female-on-male violence and female victims are nearly always more severely affected than their male counterparts (e.g., feeling fearful and trapped; Daff et al., 2021). Ackard et al. (2007) found adolescent dating violence to be positively correlated with binge eating and suicide ideation for male youths, and smoking marijuana and high depressive symptoms for female youths. Studies have identified other increased risks including substance abuse, mental health issues, risky sexual behaviors, poorer academic results, and isolation (Daff et al., 2021; Taquette & Monteiro, 2019; Wincentak et al., 2017). Consequently, engagement in adolescent dating violence is associated with numerous negative outcomes for both partners.
Patriarchal principles can lead to gender inequalities within relationships. Specifically, male dominance has been found to contribute to adolescent dating violence. Barros and Schraiber (2017) found that women report more occurrences of violence against them than their male partners acknowledge having perpetrated, indicating that men may trivialize or minimize violence, believing the behavior is warranted in some situations. Gender studies further explain that this trivialization is likely due to an historically constructed and valued way to solve conflict and acculturate women to a subordinate role in the relationship (Barros & Schraiber, 2017). In contrast, Collins et al. (2009) found that adolescent dating violence is less frequent in cultures with more egalitarian or gender-equal attitudes. In relationships that are viewed as egalitarian, couples perceive the support to be more than a mother provides, this is even when the relationship experiences conflict (Collins et al., 2009). Conversely, perceptions of inequality within relationships have been repeatedly found to negatively affect emotional functioning (Collins et al., 2009).
In Australia, First Nations women are at the greatest risk of experiencing violence at the hands of their partners (both black and white). First Nations women are 35 times more likely to be hospitalized due to family violence (up to 80× in remote areas) and 11 times more likely to die from sustained injury, compared to non-First Nations women (Hill, 2019). Furthermore, Stolen Generation survivors (children who were forcibly removed from their families), tend to stay longer in violent relationships due to not wanting their children to grow up in a fractured family (Aboriginal and Torres Strait Islander Healing Foundation, 2017). First Nations youth are also significantly overrepresented in the Australian justice system, accounting for approximately 50% of young people in detention or under community supervision on any given day (Australian Institute of Health and Welfare, 2021). Despite well-intentioned efforts and significant investments, the gap between First Nations and non-First Nations youth continues to widen. Prior to colonization, First Nations peoples lived in an egalitarian hegemony society, where the rights of men and women were balanced with specific and valued gender roles and responsibilities. Hill (2019) postulates that this negative impact on First Nations women is a continuing effect of colonization as the acculturation effects resulted in First Nations women being dependent on men for basic needs, due to the patriarchal rules of the western society.
A group at risk of perpetrating adolescent dating violence in the future is youth in juvenile detention. Male youth in detention centers tend to present with multiple risk factors which make them susceptible to perpetrating violent offences, including family and intimate partner violence (both are closely linked; Daff et al., 2021). There are also higher rates of youth with cognitive impairments in detention, which may further compound the risk factors due to difficulties with learning, impulsivity, and consequential thinking. A study by Dellar et al. (2022) reported 29% of youth detainees in Western Australia had a diagnosed cognitive impairment, with Fetal Alcohol Spectrum Disorder (FASD) being the most common. Male youth in detention may not have experienced seeing or having healthy relationships and may lack the motivation and skills to engage in healthy relationships. A treatment program delivered in detention centers may assist in reducing the likelihood of violence, once a young person returns to the community (Boxall & Morgan, 2020). Additionally, Collins et al. (2009) described a phenomenon called “selective partnering,” postulating that at-risk youths tend to choose partners with similar psychological and physical aggression to oneself. Therefore, skills development and challenging attitudes may lead to better choices in partners and adolescence may be the most effective time to provide support, as it is a critical time for establishing norms, values, and behaviors (Armytage & Ogloff, 2017).
Systematic reviews, meta-analyses, and research syntheses have identified several key factors of effective treatment programs for addressing risk factors in youth. A systematic review conducted by Pooley (2020) found that the most effective youth offending programs were based on theory, had high fidelity (replication), and considered cultural backgrounds. For example, First Nations youth perceive a support program as credible, when someone with shared cultural factors (place, language, history, or beliefs) implements these programs (Fazal, 2014). Effective programs also incorporate the Risk-Need-Responsivity (RNR) model (Bonta & Andrews, 2017), with collaborative inter- and intra-agency coordination (referrals and services), and encourage good practitioner-client relationships (Pooley, 2020). Evidence suggests that youths want to work with practitioners who have faith in them, empathize with them, and recognize their strengths (Pooley, 2020). Cognitive-behavioral approaches that provide psychoeducation, teach social skills, and include various techniques such as modelling, role-playing, performance feedback and transfer training, are effective in addressing risk factors associated with adolescent dating violence (Armytage & Ogloff, 2017; Lipsey, 2009). The use of incentives or contingency management as part of treatment approaches has been found to enhance positive outcomes (Stewart et al., 2014; Wodahl et al., 2017). Finally, ongoing evaluation of programs has been found to improve effectiveness, due to adaptations from findings (Armytage & Ogloff, 2017; Pooley, 2020).
Healthy Relationships Program
The Healthy Relationships program was developed to address adolescent dating violence and is grounded in the RNR framework (Bonta & Andrews, 2017). The program has been delivered to male youths in custody in Western Australia since 2016 and was evaluated in the current study (DoJ, 2021). Commensurate with evidence on best-practice, Healthy Relationships is a Cognitive Behavioral Therapy (CBT) based program that includes psychoeducation, skills practice, and healthy relationship modelling. The program explicitly aims to challenge attitudes relating to acceptance of couples’ violence, male role norms, and attitudes towards women, to improve interpersonal interactions and reduce violence within relationships (DoJ, 2021). An additional overarching aim of the program is to reduce recidivism. The program was developed in consultation with First Nations Welfare Officers working in custody and uses First Nations Australian resources within its delivery activities. Psychologists and Social Workers employed in the DoJ’s Youth Justice Psychological Clinical Programs team facilitate the program delivery.
Treatment programs delivered in detention centers were found to be effective across a wide range of countries when they adhered to the RNR principles (Bonta & Andrews, 2017; Vitopoulos et al., 2012). The RNR model is founded on three key principles that guide effective assessment and rehabilitation of offenders (Bonta & Andrews, 2017; Taxman et al., 2006). The risk principle assumes that criminal behavior can be predicted. The level of treatment intensity should match the level of risk; the higher the risk, the more intensive and extensive the program; with low-risk individuals needing minimal support (Bonta & Andrews, 2017; Vitopoulos et al., 2012). The Healthy Relationship Program used the Youth Level of Service/Case Management Inventory (YLS/CMI; Hoge & Andrews, 2003) tool to assess risk and participants needed to score Moderate or higher to be accepted into the program. The need principle relates to programs targeting criminogenic risk factors (treatment needs) associated with recidivism such as antisocial attitudes (Bonta & Andrews, 2017; Vitopoulos et al., 2012). The need component was achieved by addressing criminogenic factors known to increase risk of youth intimate partner violence, such as attitudes supportive of violence and traditional gender roles and stereotypes. The final principle, responsivity, relates to the type of program and tailoring the delivery to the needs and learning styles of the participants (Bonta & Andrews, 2017). Responsivity was addressed through program flexibility and responsiveness of facilitators to adjust to participant’s specific needs, for example, utilizing more visual tools (videos) to accommodate for participants with FASD and other cognitive impairments.
Evaluating the effectiveness of criminal justice programs is essential to ensure that resources are being directed to the programs delivering the most culturally sound and effective outcomes. Given the significant overrepresentation of First Nations youth in the Australian justice system, there is also an urgent need to evaluate and identify evidence-based and culturally appropriate programs. The current study aimed to evaluate the effectiveness of the Healthy Relationships Program in reducing attitudes endorsing intimate partner violence and gender roles and stereotypes. It was expected that participation in the program would lead to a decrease in attitudes supportive of intimate partner violence for both First Nations and non-First Nations participants. Participation in the program was also expected to lead to a decrease in attitudes endorsing traditional gender roles and stereotypes for both groups.
Method
Research Design
This study employed a mixed methods approach to evaluate the Healthy Relationships program, with quantitative data as the primary source with qualitative data used to provide supplementary information about the program. The quantitative component used a within-participants design to compare the results on pre-program questionnaires to results on the questionnaires following program participation. The first paired-group analysis compared pre- and post-program scores on the Acceptance of Couples Violence Scale and the Attitudes towards Women Scale. The second analysis included a bivariate regression with treatment readiness as the independent variable, which was used to predict the criterion variable, treatment engagement. Post-program scores on Acceptance of Couples Violence and Attitudes Toward Women scales were included as additional covariates. Qualitative data was used to further explain the level of treatment engagement, as this data relates to participants’ perceptions of the program. Our research team included both First Nations (RW) and non-First Nations (LR and KD) researchers.
Sample
The data was provided by the Western Australian Department of Justice (DoJ) and included pre and post program assessment data for 65 male adolescents aged 14 to 18 years (M = 17.1, SD = 1.09) who participated in the Healthy Relationships program between November 2016 and June 2021. Program facilitators are responsible for collecting the data during the pre- and post-group interviews and entering the data into the Departmental database. Not all participants who complete a pre-group interview are deemed suitable for participation in the program for a variety of reasons, the most common of which include low treatment readiness and motivation, cognitive impairments limiting ability to engage, or not enough time left in custody prior to release to be able to complete the program. The final sample in the current study therefore only included participants with full pre (Time 1) and post (Time 2) data. A sensitivity analysis conducted in G*Power indicated the sample had sufficient power to detect small to medium effects sizes for planned analyses.
Measures
Acceptance of Couples Violence (ACV) Scale
The ACV (Foshee et al., 1998) measures the level of acceptance for violence within relationships and includes three attitudinal subscales (i) male-on-female violence, (ii) female-on-male violence and (iii) general relationship violence. The ACV is an 11 item self-report measure that is scored on a 4-point Likert scale with 1 = strongly disagree to 4 = strongly agree. An example item is “violence between dating partners can improve the relationship.” The ACV is administered pre and post program participation to assess the level of acceptance of violence within intimate partner relationships. Possible scores for the full scale range from 11 to 44, with higher scores indicating more acceptance of violence within intimate relationships. Previous studies have found good internal consistency (full scale α = .80). In our study, the internal consistency was strong for Time 1 (α = .88) and Time 2 (α = .94).
Attitudes Toward Women (ATW) Scale
The ATW Adolescent version (Galambos et al., 1985) measures participants’ attitudes regarding women’s roles and rights in society or relationships. The 12 item self-report measure is scored on a 4-point Likert scale with 1 = disagree strongly to 4 = agree strongly. An example item is “boys are better leaders than girls.” Scores range from 12 to 48 with higher scores indicating less traditional attitudes toward women. The ATW scale is administered at both pre and post program participation stage, to assess participants endorsement of attitudes supporting traditional gender roles and stereotypes. Previous studies have found good internal consistency (female α = .72 and male α = .78; Galambos et al., 1985). Within our study, following an exploratory factor analysis (EFA; Promax rotation) only four items were extracted with an Eigenvalue greater than 1.00. A second EFA was conducted, forcing a 1-factor solution. Five items were removed to increase the internal consistency of the scale, resulting in Cronbach’s alpha at Time 1 (α = .81) and Time 2 (α = .84).
Treatment Readiness Questionnaire (TRQ)
The TRQ (Casey et al., 2007) measures participants’ readiness for change. There are four components (i) attitudes and motivation, (ii) emotional reaction, (iii) offending beliefs, and (iv) efficacy. The scale is a 20 item self-report measure, using a 5-point Likert scale with 1 = strongly disagree to 5 = strongly agree. An example item is “generally I can trust other people.” Possible scores range from 20 to 100, with higher scores indicating a higher degree of readiness to participate in a treatment program. For our study, internal consistency was high (α = .82). The TRQ is administered prior to program participation to assess level of treatment readiness and motivation. In line with the RNR model, caution is applied to young people who score low on the TRQ (< 60), and consideration is given to preparatory work (i.e., motivational interviewing) that may be required to increase treatment readiness prior to inclusion.
Treatment Engagement Questionnaire (TEQ)
The TEQ (Casey et al., 2007) is a self-report scale measuring participant perception of the treatment program, including treatment setting and therapy context. It consists of three factors (i) alliance with facilitators, (ii) group process/dynamics, and (iii) confidence—participant’s self-confidence to change offending behavior. The TEQ is administered to participants after program participation to assess their level of engagement in the program and identify if any of the three factors were particularly influential in level of engagement. The self-report measure is scored on a 5-point Likert scale with 1 = strongly disagree and 5 = strongly agree and total score ranges from 17 to 85, with higher scores indicating higher perceived engagement with the treatment program. An example item is “I have gained a better understanding of my problem with offending.” Internal consistency for the full TEQ in our study was α = .76.
Participant Feedback Questionnaire
The participant feedback questionnaire comprised the qualitative component of the study. The questionnaire includes seven open-ended questions which were used to illuminate the participants’ perceptions of the program, including facilitator alliance. An example question was “what did you learn from the program?.” In the final session of the program, participants are given the opportunity to provide their feedback on the program and facilitators. This process is completed prior to the post-group assessment interviews and therefore only includes participants who have completed the entire program.
Procedure
The Healthy Relationships Program was available to male youths in Western Australia aged 14 years and above who were in custody and were rated as Moderate or higher on the YLS/CMI risk assessment tool (Hoge & Andrews, 2003). Participants were referred to the program by their custodial case manager if they had enough time remaining in custody to complete the program (i.e., minimum 6 weeks). Following the referral, facilitators of the Healthy Relationships Program completed a pre-group assessment interview with the youth to determine their suitability and motivation to participate. Participation in the program was on a voluntary basis and in line with the RNR model, youths were prioritized if they had higher level of risk or if they had not engaged in any prior treatment programs. Facilitators made the final decision for program inclusion based on (i) group dynamics, (ii) security risk, (iii) level of cognitive functioning, (iv) scores on pre-assessment measures, and (v) treatment need.
Program delivery consisted of eight closed-group sessions (two per week; 1.5-hours each). The program was delivered over approximately 5 weeks, inclusive of pre and post assessment and graduation. Qualitative data were collected in the final session and post quantitative data were collected approximately 1 week after program participation. A post-program interview was scheduled where youths were returned the work completed during the sessions (e.g., accomplishments and strategies learned) and provided with feedback on how well they progressed through the program.
Results
All data was analyzed using SPSS version 22.0 software. Assumptions testing was conducted prior to each analysis to ensure appropriate conclusions could be drawn from the data. Ninety-six participants commenced the program, of which 65 completed the ACV at Time 1 and Time 2, with 40 completing the Attitudes Towards Women measure at Time 1 and Time 2. It is noted that another measure was used previously but discontinued due to participants experiencing difficulties understanding the items. Due to the small sample sizes, full scales rather than subscales were used in all analyses. Eleven participants did not complete the Treatment Readiness Questionnaire and were therefore excluded from the regression analysis.
Pre and Post Program Comparisons
Two one-tailed paired samples t-tests were conducted to compare pre- and post-program scores on the Acceptance of Couple Violence and Attitudes Towards Women scales (see Table 1). On average, participants post-program Acceptance of Couple Violence scores were a significant 1.32 points lower than pre-program scores (95% CI [0.03, 2.61]); a small effect size. On average, participants post-program Attitudes Towards Women scores were a significant 1.75 points lower than pre-program scores (95% CI [0.82, 2.67]); a medium effect size. As expected, the results indicated that participation in the program resulted in significantly lower attitudes endorsing intimate partner violence and attitudes endorsing traditional gender roles and stereotypes.
Paired Groups Results for Full Sample and First Nations Only Sample on Pre and Post Assessment Measures.
Note. N = number of participants; M = mean; SD = standard deviation; t = t-test; p = significant value; d = Cohen’s d; CI = confidence interval.
First Nations Youth
An exploratory analysis examining differences between First Nations (n = 51) and non-First Nations (n = 14) youths was not feasible due to varying group sizes resulting in low power. Instead, paired samples t-tests were deemed feasible to analyze only First Nations participant data. To do this, three one-tailed boot-strapped paired samples t-tests were conducted to compare pre- and post-program scores on ACV and ATW scales. On average, participants post-program ACV scores were 1.21 points lower than pre-program scores (95% CI [−0.11, 2.58]). This difference showed a small effect size but was not statistically significant. On average, participants post-program ATW scores were 1.40 points lower than pre-program scores (95% CI [0.36, 2.39]). This difference showed a medium effect size and was statistically significant (see Table 1 for full results).
Regression Analysis
A bivariate regression was conducted to estimate the proportion of variance in treatment engagement that was accounted for by treatment readiness. Assumptions of normality, linearity, multicollinearity and homoscedasticity were met, and multivariate outliers were not of concern as Mahalanobis distance did not exceed the critical value for any case in the data file. Contrary to expectations, results showed that TRQ scores accounted for a non-significant 2.7% of the variability in post-program scores on TEQ (R2 = .027, adjusted R2 = −.038, F(1, 15) = .409, p = .532). To evaluate whether treatment readiness could account for a significant portion of the variance of attitude change, beyond that already accounted for by pre-program scores, two multiple regressions analyses were performed.
Acceptance of Couple Violence
On step 1 of the hierarchical multiple regression analyses, pre-programs scores on Acceptance of Couple Violence accounted for a significant 47.2% of the variance in post-program scores, R2 = .472, F(1, 52) = 46.54, p < .001. On step 2, Treatment Readiness was added and accounted for an additional non-significant 2.9% of the variance in post-program scores, R2 = .482, F(1, 51) = 2.96, p = .091. In combination, Treatment Readiness and pre-program scores on Acceptance of Couple Violence accounted for 50.1% of variability in Acceptance of Couple Violence post-program scores, R2 = .50, adjusted R2 = .48, F(2, 51) = 2.96, p < .001. The regression coefficients can be seen in Table 2.
Hierarchical Multiple Regression Predicting Post-Test Scores on the Acceptance of Couples Violence Scale.
Note. b = unstandardized regression coefficient; CI = confidence interval; β = standardized regression coefficient; sr2 = squared semi-partial correlations; p = significance.
Attitudes Towards Women
On step 1 of the second hierarchical multiple regression analyses, pre-program scores on Attitudes Towards Women accounted for a significant 57.9% of the variance in post-program scores, R2 = .579, F(1, 29) = 39.84, p < .001. On step 2, Treatment Readiness was added and accounted for an additional non-significant 1.1% of the variance in post-program scores, R2 = .56, F(1, 28) = .73, p = .40. In combination, treatment readiness and pre-program scores on Attitudes Towards Women accounted for 59% of variability in Attitudes Towards Women post-program scores, R2 = .59, adjusted R2 = .56, F(2, 28) = 20.10, p < .001. The regression coefficients can be seen in Table 3.
Hierarchical Multiple Regression Predicting Post-Test Scores on the Attitudes Towards Women Scale.
Note. b = unstandardized regression coefficient; CI = confidence interval; β = standardized regression coefficient; sr2 = squared semi-partial correlations; p = significance.
Qualitative Analysis
To see how participants perceived the Health Relationships Program, a conventional content analysis (Hsieh & Shannon, 2005) was used to inductively code participants responses into categories from the seven open-ended questions. Familiarization occurred by reading through the responses, units were condensed, then organized into meaningful clusters that lead to the categories used. Data was limited, with 20 participants who responded to the seven questions. Thirteen categories were developed, and Table 4 provides an overview of questions with multiple categories of responses. One of the question’s responses were cross coded by two researchers. Cohen’s kappa was calculated and reflects good inter-rater reliability (K = .65; Fleiss et al., 1979).
Results from Content Analysis of Participant Feedback Questionnaire.
Note. Responses could be coded into multiple categories.
The feedback from program participants further supports the quantitative data in demonstrating the benefits of participating in the Healthy Relationships program. Most participants reported a positive experience of the program in terms of learning about and gaining skills to develop and maintain healthy intimate partner relationships. Consistent with the quantitative data and prior research (see De Boer et al., 2023), the feedback highlighted the impact that the therapeutic alliance has on participants experience and engagement in the program. As one participant noted, “the facilitators were very respectful and understanding at all times, thank you.” Consistent with the literature on incentive-based learning systems (Stewart et al., 2014; Wodahl et al., 2017), participants frequently commented on the positive benefits of incorporating incentives into the program. Program duration emerged as a theme as participants expressed a desire for the program to run over a longer duration of time. Participants also expressed a desire for the program to be more intensive to allow more opportunity to learn about the concepts covered. In addition, when asked for “Any other comments/feedback?,” participants thanked the facilitators and expressed how much they enjoyed the program, for example, one participant wrote “this program was deadly (cool), I appreciate everything, cheers.”
Discussion
The current research sought to evaluate the effectiveness of the Healthy Relationships program in reducing participants endorsement of attitudes supporting violence in relationships and traditional gender stereotypes. As expected, comparisons of pre- and post-program data showed statistically significant reductions in participants attitudes endorsing intimate partner violence and traditional gender roles and stereotypes. First Nations youth benefited from participation in the program, with significant reductions in attitudes supportive of traditional gender stereotypes. Although non-significant, the results also showed a reduction in endorsement of attitudes supporting intimate partner violence for First Nations youth, indicating promising results. Contrary to expectations, results showed that treatment readiness did not significantly predict treatment engagement and was not found to account for a significant variance of change, beyond that already accounted for by pre-program scores. This indicates that within our sample, participants were found to benefit from the Healthy Relationships Program irrespective of their score on the Treatment Readiness Questionnaire, which supports Casey et al.’s (2007) recommendation that treatment readiness scores should not be used to exclude participants from treatment programs. Collectively, the results provide empirical support for the effectiveness of the Healthy Relationships program in reducing risk factors associated with adolescent dating violence, namely, attitudes supportive of intimate partner violence and attitudes endorsing traditional gender roles and stereotypes.
Qualitative analysis of participant feedback supported the quantitative findings, indicating that participants found the program to be beneficial. Participant feedback highlighted the importance of the therapeutic alliance and the use of incentive-based systems for positive program outcomes. These findings are consistent with prior research that identified the importance of practitioner-youth relationships which are grounded in understanding, recognition, and trust (see De Boer et al., 2023). This is particularly relevant for First Nations youth and may provide some insight into the effectiveness of the Healthy Relationships program for First Nations youth. Prior research with First Nations youth has emphasized the importance of developing relationships with program facilitators (Stewart et al., 2014). Many participants commented on how respectful and skillful the facilitators were and how helpful they found the program. Although the facilitators of the Healthy Relationships Program may not have shared any cultural factors (as they were white women), the participants’ positive feedback, as seen in qualitative data, is likely linked to Pooley’s (2020) findings that youths want to work with practitioners who have faith in them, empathize with them, and recognize their strengths; this also links to the responsivity aspect of the RNR model.
The delivery of culturally appropriate programs is paramount. Considering the over representation of First Nation youths in detention centers, particularly in Australia, the positive results from our study are promising. A review of the qualitative data showed that there were no negative comments about the program or the facilitators. Although participants provided positive feedback, some adjustments may improve outcomes. One inclusion could be to have at least one First Nation facilitator delivering the program, as First Nation youths may perceive a program to be more credible if the facilitator shares some cultural factors (Fazal, 2014); and if possible, a male facilitator might be appropriate to balance the gender aspect for First Nation youths.
Following participation in the Healthy Relationships program, youth reported less acceptance for violence in relationships. This suggests that they may be less likely to get involved, or stay with, a partner who accepts violence within relationships, as evidenced by Collins et al.’s (2009) theory that at-risk youth select partners with similar psychological and physical aggression to oneself. Participation in the Healthy Relationships program was found to increase participant attitudes towards women as being more equal, in terms of their rights and responsibilities within society and relationships. The real-world implications may mean less frequency of violence within relationships, as evidenced by Collins et al. (2009) findings that adolescent dating violence was less frequent in relationships when couples viewed the relationships as more egalitarian.
Strengths and Limitations
The major strength of our study is that it evaluated a specific identifiable program, within the area of youth justice, in Australia. As identified by Armytage and Ogloff (2017), very few specific program evaluations have been published in this field to date which limits the ability to effectively review, compare, and assess available treatment options. Our study can inform future evaluations of youth justice programs and the development of new measures or methodologies for collecting rich data from First Nations and detention center youth. We assessed a population that can be difficult to access, for research purposes, due to their high-risk nature from an ethical standpoint. As the program was seen to benefit youths in this sample, the Healthy Relationships Program could be recommended for use in other youth detention centers, and with some adjustment (around gender considerations) the program could be transferrable to female youth in detention. The program could also be delivered to youth in the community as there are no detention-specific elements of the program that could not be replicated in community settings. This would allow the program to reach a larger range of justice-involved youth and may assist in reducing the likelihood of ending up in detention for intimate-partner related offending.
While the results of our study provide empirical support for the effectiveness of the Healthy Relationships program, there are limitations that are important to discuss in contextualizing the results and providing recommendations for future research. There was no control group in our sample, which limits the ability to assert that the program was solely responsible for the observed improvements. Although a control group would be useful for comparison purposes, it is acknowledged that this is not currently a feasible option in the detention center where the Healthy Relationships program is delivered. The unique demographics and relatively small sample size may limit the generalizability of our results to the wider population. However, the responsivity principle of the RNR model suggests that the program may be effective in other jurisdictions if the delivery of the content is adapted for the learning styles and needs of the target population. For example, using localized terminology and population-specific resources. Longitudinal research is needed to assess the long-term impact of participation in the program and observed changes over time. This is an endeavor that should be undertaken, as healthy relationships are linked to positive self-worth and general competency in self (Collins et al., 2009). Additionally, skills developed through forming healthy relationships may mitigate risk factors for other antisocial behaviors, further reducing recidivism (Miller et al., 2015). The Healthy Relationships program was initially designed to challenge attitudes, rather than change attitudes. The risk aspect of the RNR framework denotes that the higher the level of risk, the higher the dose needed for change. To sustain attitude change, considerations of dose effect may be needed when developing future programs, by increasing their intensity (number of sessions in the same timeframe) or extensity (extend the duration of the program), or scaffolding programs. This could be done once the youth return to their community or as a consecutive program whilst still in the detention center.
The measures used to collect data require further consideration. The Attitudes Towards Women scale’s construct validity for this sample was compromised, as five items had to be removed from the scale which only initially contained 12 items. The Attitudes Towards Women scale was developed with a sample of adolescents from the United States mid-west rural and suburban areas. Participants were mostly Caucasian from lower to middle socioeconomic status; age range 11 to 17 years (Galambos et al., 1985). American scales may not translate well in Australia, particularly with the high number of First Nations participants, with a large number from regional areas. Moreover, gender stereotypes may not be recognizable for First Nations youths, as prior to colonization, First Nations culture was an egalitarian society and had specific and valued gendered roles. Although many First Nations cultural understandings have been fragmented, due to colonization, the underlining mechanisms of the culture still exist within communities, though they may be invisible through a western cultural lens (Hill, 2019). There is a need to direct research funding into developing appropriate measures, to ensure program assessment is grounded in cultural understandings that lead to improved outcomes for First Nations youths and youths in detention centers. Resourcing a more appropriate measure for this sample is recommended and if one does not exist a new measure that is ecologically balanced for the identified group (First Nations youth and youths in detention centers) needs to be developed. Consideration should be given to design a scale that assesses differences in traditional gender roles, or a better theoretical or philosophical reason for violence in relationships for First Nations youths.
Qualitative data provided valuable insights from the perspectives of program participants into what they identified as beneficial for their engagement in the program. Further qualitative research, potentially utilizing First Nation research methodologies (Martin, 2017), could further increase understanding of participants’ experiences of the program and assist in identifying specific program characteristics that contribute to positive outcomes. Yarning sessions, semi-structured or open interviews would be culturally appropriate, as First Nation culture relies on oral communication for information sharing and society functioning (Hill, 2019). It would also be beneficial to conduct interviews with participants who opt to withdraw from the program, to identify potential barriers to program completion and improvements that could be implemented to improve program participation and engagement.
Conclusion
The findings from our study support the continuation of the Healthy Relationships Program and contribute to the “what works” literature on effective risk reduction approaches for justice-involved youth. However, longitudinal research is needed to assess the long-term impacts of participation in the program. Longitudinal program evaluations should be prioritized, as effective and culturally appropriate programs are needed, particularly to address the overrepresentation of First Nations youth in detention centers. However, a program that can challenge attitudes relating to violence in relationships, and which has the potential to increase relationship skills and reduce recidivism, is a promising step in the right direction.
Footnotes
Acknowledgements
We acknowledge the participation and assistance of the Department of Justice in the conduct of our research. The research reported cannot be considered as either endorsed by the Department of Justice or an expression of the policies or view of the Department. Any errors of omission or commission are the responsibility of the researchers.
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) received no financial support for the research, authorship, and/or publication of this article.
Ethical Approval
Ethical approval for this research was obtained from the Department of Justice Research Application and Advisory Committee (RAAC; Project ID 480) and Curtin University Human Research Ethics Committee (Approval Number HRE2021-0240).
Data Availability Statement
Raw data were generated at the Department of Justice in Western Australia. Derived data supporting the findings of this study are available from the corresponding author KD on request