Child Abuse & Neglect 42 (2015) 40–53

Contents lists available at ScienceDirect

Child Abuse & Neglect

Research article

The role of positive youth development practices in building resilience and enhancing wellbeing for at-risk youth夽 Jackie Sanders a,∗ , Robyn Munford a , Tewaporn Thimasarn-Anwar b , Linda Liebenberg c , Michael Ungar c a School of Social Work, Practice Research and Professional Development Hub, Massey University, Private Bag 11 222, Palmerston North, New Zealand b Practice Research and Professional Development Hub, School of Work, Massey University, New Zealand c Resilience Research Center, Dalhousie University, PO Box 15000, Halifax, NS B3H 4R2, Canada

a r t i c l e

i n f o

Article history: Received 10 November 2014 Received in revised form 3 February 2015 Accepted 9 February 2015 Available online 12 March 2015 Keywords: Positive youth development Resilience Multi-system youth Service use Service satisfaction Adolescent Risk

a b s t r a c t Services that utilise positive youth development practices (PYD) are thought to improve the quality of the service experience leading to better outcomes for at-risk youth. This article reports on a study of 605 adolescents (aged 12–17 years) who were concurrent clients of two or more service systems (child welfare, juvenile justice, additional education, mental health). It was hypothesised that services adopting PYD approaches would be related to increases in youth resilience and better wellbeing outcomes. It was also hypothesised that risks, resilience, service experiences and wellbeing outcomes would differ by age, gender and ethnicity. Youth completed a self-report questionnaire administered individually. Path analysis was used to determine the relationship between risk, service use, resilience and a wellbeing outcome measure. MANOVA was then used to determine patterns of risk, service use, resilience and wellbeing among participants based on their demographic characteristics. Services using PYD approaches were significantly related to higher levels of youth resilience. Similarly, increased resilience was related to increased indicators of wellbeing, suggesting the mediating role of resilience between risk factors and wellbeing outcomes. When professionals adopt PYD practices and work with the positive resources around youth (their own resilience processes) interventions can make a significant contribution to wellbeing outcomes for at-risk youth. © 2015 Elsevier Ltd. All rights reserved.

Introduction There is growing interest in defining the characteristics of psychosocial services that are most effective in improving outcomes for young people who are exposed to high levels of enduring individual, family and contextual risk such as abuse, neglect, educational disengagement and neighborhood distress (Berzin, 2010). Young people who face the most risk are often clients of multiple service systems (Haapasalo, 2000; Hazen, Hough, Landsverk, & Wood, 2004; Kroll et al., 2002); their exposure to these risks brings them into contact with the child welfare, juvenile justice, specialist educational and mental health systems. In principle, this repeated contact means that there are numerous opportunities for services to have a positive effect. It is not clear, however, that these multiple opportunities to intervene do result in improved outcomes for

夽 The authors thank Ministry of Business, Innovation and Employment for funding this research program. ∗ Corresponding author. http://dx.doi.org/10.1016/j.chiabu.2015.02.006 0145-2134/© 2015 Elsevier Ltd. All rights reserved.

J. Sanders et al. / Child Abuse & Neglect 42 (2015) 40–53

41

children facing high levels of risk. Indeed, there is some evidence that the involvement of more services is associated with increased rather than reduced risks (Ungar, Liebenberg, Dudding, Armstrong, & Van de Vijver, 2003; Garland, Aarons, Brown, Wood, & Hough, 2003; Hazen et al., 2004; Kroll et al., 2002; Loeber, Farrington, Stouthamer-Loeber, & Van Kammen, 1998;) and with poorer outcomes for very vulnerable young people (Berzin, 2010). The study reported on in this paper examines the possible mediating influence of resilience on wellbeing outcomes, accounting for the quantity and quality of services by a cohort of youth facing high individual and contextual risks. It also considers the way in which youth perceive the risks they face, the quality of the services received, the presence of resilience resources and the wellbeing of youth by age, gender and ethnicity. Of central concern is understanding whether or not service delivery that takes a Positive Youth Development approach (PYD) has the potential to boost the resilience resources of youth making them more able to achieve positive outcomes when they confront high levels of risk. A Positive Youth Development Approach (PYD) to Supporting At-Risk Youth Despite having its origins in the development of generic programs for youth in community settings, positive youth development theory (PYD) is relevant to the delivery of programs to young people who face enduring risks and vulnerabilities and who become clients of multiple service systems. PYD has redefined adolescence so that, as Lerner notes (2005), rather than being seen as . . . broken, in need of psychosocial repair, or [as] problems to be managed (Roth, Brookes-Gunn, Murray and Foster, 1998). . . all youth are seen as resources to be developed. This positive orientation to understanding adolescence is important to the study of at-risk youth who have often been defined primarily in terms of their exposure to risk factors such as abuse and neglect, the difficulties they face in adapting positively and in the troubling behaviors they exhibit (Case, 2006). Key components of PYD programs include the encouragement of personal agency in youth, respectful approaches to youth and their families, and a focus on young people’s strengths and competencies alongside the risks and challenges they may confront. PYD argues that when these positive components inform program delivery the likelihood that all young people will develop into citizens who thrive and contribute to society increases (Lerner, 2005, 2006; Lerner, Almerigi, Theokas, & Lerner, 2005; Pittman, 1999). Because a key feature of adolescence is growing autonomy, the PYD emphasis upon the active involvement of youth in decision making is important. Positive relationships are a critical dimension of adolescent development (Lerner, 2005) and they are central to PYD definitions of effective programs; when young people have mutually beneficial relationships with the people and institutions in their social world they will thrive and contribute (Heinze, 2013). Others have also observed when services are respectful and provide opportunities for at-risk youth to exercise agency that better outcomes are achieved than when these characteristics are not features of professional practice (Duncan, Miller, & Sparks, 2004). Relationships can open up new networks and provide opportunities for emotional connection and attachment; factors that have a powerful influence on outcomes in adulthood (Schofield & Beek, 2009). These factors are also likely to boost resilience processes – those relational and personal resources youth can draw on as they cope with adversity and stresses and challenges. Positive youth development occurs when opportunities are made available to youth in meaningful ways and when the people around youth support them to develop their own unique capacities and abilities. The components of PYD identified above have relevance for work with youth who have high risk burdens both in terms of their own behaviors and in terms of levels of exposure to neighborhood and family risks. Evidence is beginning to suggest that positive gains can be made with at-risk youth when a PYD approach guides practice (Eccles, Barber, Stone, & Hunt, 2003; Heinze, 2013; Scales, Benson, & Mannes, 2006). The PYD focus on growth and development is a valuable characteristic of practice with such youth because it emphasises that change is achievable even for youth who confront significant adversity, who may engage in very harmful behaviors and who are clients of multiple services. There are three reasons why PYD approaches work well with at-risk youth. First, young people who have been exposed to abuse and neglect and who face enduring individual and contextual risks experience compressed childhoods and accelerated transitions to autonomy (Rogers, 2011; Stein, Ward, & Courtney, 2011); their childhoods end early and often abruptly, and they take on adult responsibilities at a young age. Programs that encourage personal agency and participation in decision making provide opportunities for young people who have had compressed childhoods and accelerated autonomy experiences to be supported to learn how to use their autonomy in pro-social ways; these approaches have the potential to build the resilience resources available to these youth. Such programs also avoid the risk of infantilising youth for whom childhood ended early. Second, the strengths emphasis provides an antidote to the predominant focus upon risks, problems and deficits in work with this population of youth (Case, 2006). It recognises that at-risk youth are more than the sum total of the challenges they face and that they have capacities, competence and resilience resources that can be used to good effect in the treatment process (Cheon, 2008). Third, the ecological emphasis of PYD approaches increases the chances that interventions will be respectful of and responsive to the particularities of the individual youth and their family. This means that interventions are more likely to take account of and adapt to the realities of youth circumstances thereby increasing the chances that interventions will be meaningful and relevant. Furthermore, such approaches are more likely to make a positive contribution to the resilience resources already around youth (Ungar et al., 2013; Bottrell, 2009). The service quality measure used in the current study assesses the presence of these PYD dimensions of practice in order to ascertain whether or not they are linked to better resilience and to positive outcomes.

42

J. Sanders et al. / Child Abuse & Neglect 42 (2015) 40–53

Resilience as a Mediator of Outcomes for At-Risk Youth Berzin (2010, p. 493) suggests that, on their own, services will not produce better outcomes for vulnerable, multi-service system youth. She contends that positive outcomes are achieved when interventions work in conjunction with the resilience resources around youth. This observation is consistent with ecological theories of resilience, and also with PYD approaches, which argue that healthy development occurs within facilitative contexts that are shaped by families and neighborhoods as well as by wider political, social and economic systems. Service delivery systems form part of this wider ecology. What this means is that the resilience of individual children, that is; their capacity to cope well with adversity depends upon the resources around them (including services and informal support networks) as well as upon their innate characteristics (Ungar, 2011; Bottrell, 2009). In this study resilience is defined as the capacity of young people to navigate their way to the resources they need during crises, and their ability to negotiate for these resources to be provided in meaningful ways (Ungar & Liebenberg, 2011). Resilience is both individual processes that increase survival and the protective processes instigated by larger systems to provide opportunities for individuals to cope under stress (Lerner, 2006; Ungar & Liebenberg, 2011; Zautra, Hall, & Murray, 2008) (Ungar et al., 2013). This definition is important to the current study because it emphasises interactional dimensions of resilience and its potential as a mediator that sits between the risks youth confront, the services they receive, and the wellbeing outcomes they are able to achieve. In this regard, resilience is not an outcome in itself, but rather a cluster of positive resources upon which youth can draw as they strive to achieve positive outcomes. When resilience is viewed ecologically, services potentially become assets that can contribute to resilience for at-risk youth. Services can provide psycho-social, practical and emotional resources that make it easier for youth cope with crises themselves, and services can also act directly to mitigate risks. This understanding proposes that interventions by professionals lead to better outcomes by adding to young people’s resilience (Ungar et al., 2013; Gilligan, 2004; Mitchell, 2011). Research into resilience as an ecological construct is still in its infancy and there are relatively few studies that examine service provision as a component of resilience. What evidence there is suggests that well-co-ordinated and accessible interventions can increase the resilience of young people with complex needs (Garland, Hough, Landsverk, & Brown, 2001; Stathis, Letters, Doolan, & Whittingham, 2006). Further, particular types of services, such as supported independent living and transition planning have been found to promote the resilience of multiple-system involved youth resulting in them doing better than similar youth who do not have the benefit of such services (Berzin, 2010). An ecological understanding of resilience has a good fit with interventions that adopt PYD practices because both emphasise the capacities and competence of youth and both seek to build on young people’s own ways of coping with and responding to risks and crises rather than supplanting them. Services, Risks, Resilience and Outcomes The relationship between service delivery, risk, resilience and outcomes is complex. When risks accumulate around young people they become particularly vulnerable to poor outcomes and service system engagement can exacerbate this unless interventions are able to address the causes of the risks that youth confront (Berzin, 2010). Risks compound normal developmental challenges rendering young people particularly vulnerable to enduring negative consequences across the life span (Beale Spencer et al., 2006). Berzin (2010) suggests that understanding of the ways in which contextual risks, involvement in more than one service system, and other vulnerabilities, combine with youth resilience to influence outcomes is still in its infancy. Specifically, it is not clear how ‘multiple layers of risk’ (Berzin, 2010, p. 493) interact with service engagement and resilience to make multi-system youth more or less vulnerable to poor outcomes. Contextual risks, such as poverty, exposure to domestic violence, abuse and neglect, substance abuse, familial offending, exclusion from education, neighborhood distress as well as physical and mental health issues, all reduce the capacity of multi-system youth to achieve successful outcomes (Berzin, 2008, 2010; Bullis & Yovanoff, 2002; Courtney & Dworsky, 2006; Fagan & Freeman, 1999; Goerge et al., 2002; Mechanic and Tanner, 2007; Wagner et al., 2005). Current scholarship provides only limited guidance concerning how these complex and diverse risk factors influence the pathways of vulnerable youth through adolescence and of the ways in which services can best assist these youth. In fact, it has been suggested that rather than improving outcomes, involvement in more than one service system may intensify the challenge of making a successful transition to young adulthood (Berzin, 2010). This raises the possibility that service involvement might in some cases actually constitute an additional risk factor. It is, therefore, important to understand more about the ways in which outcomes are influenced by the interactions between service delivery, risk, and resilience for vulnerable young people. Demographic Factors, Risks, Resilience, Service Delivery and Outcomes The influence that demographic characteristics have upon patterns of risk, resilience, service delivery and outcomes appears to be complicated. For instance, the developmental pathways of male and female youth appear to be influenced by different combinations of risks and protective factors (Twenge & Nolen-Hoeksema, 2002); males report more risks than females (Capaldi, 1992) while females report elevated rates of depression (Botticello, 2009). Increased delinquency and

J. Sanders et al. / Child Abuse & Neglect 42 (2015) 40–53

43

Table 1 Recruitment of youth by service sector. Sector

n

Youth justice Education Child welfare Mental health

214 112 157 122

Total

605

% 35.3 18.5 26.1 20.2 100

subsequent increases in depressive symptoms have been observed in males but not females (Wiesner, 2003). Elsewhere, being female has been observed to provide protection in relation to externalising but not internalising risks, while being male appears to act in the reverse manner (Fergusson & Horwood, 2003). Females have been observed to have slightly higher resilience than males (Ungar et al., 2013). While youth from minority ethnic groups tend to be overrepresented in risk statistics and underrepresented in those relating to positive outcomes (Beale Spencer et al., 2006; Berzin, 2010; Department of Labour, 2012; Sveinsson, 2012). There is also evidence that minority ethnic identity in general and, indigenous identity in particular, may confer resilience advantages (Sanders & Munford, 2015; Mossakowski, 2003; Sellers & Shelton, 2003; Sellers, Caldwell, Schmeelk-Cone, & Zimmerman, 2003; Wong, Eccles, & Sameroff, 2003). Some differences have also been observed in relation to service utilisation, although as with risk and resilience, the picture appears complex; while some studies identify differences in service use related to demographic factors others do not. Hazen et al. (2004) in a comprehensive study of lifetime rates of mental health service use in a representative sample of multiple service-involved youth found that older youth were more likely to have had involvement with the justice system, that males reported higher utilisation of school and justice-related services and that females had greater involvement in outpatient mental health services. Elsewhere, older youth have reported greater service utilisation, females greater service satisfaction, and visible majority youth have reported larger service use histories and greater service satisfaction than minority youth (Ungar et al., 2013). In a large meta-analysis of juvenile justice interventions, Lipsey (2006) reported that the demographic characteristics of juvenile offenders did not appear to have much impact on service effectiveness. That said, he also pointed to a marginally significant relationship in relation to gender such that programs with more males had poorer outcomes and counseling appeared less effective for male than for female offenders and more effective for minority than majority youth. The Current Study Using a methodology developed in Canada (Ungar et al., 2013) the current study examines whether or not services that adopt PYD approaches are related to improved outcomes for at-risk young people in New Zealand. The young people who are the focus of this study have been exposed to significant harm during childhood and adolescence and this includes exposure to chronic abuse and neglect. The analysis considers the possible role that PYD approaches play in bolstering youth resilience (defined ecologically). Risk is measured on a range of dimensions including individual (internalising and externalising) and contextual risks (family, education and neighborhood). The current paper is concerned with examining the following hypotheses: 1. Services that adopt PYD approaches are related to increased youth resilience and better outcomes. 2. Risk, resilience, service experiences and wellbeing outcomes are experienced differently by age, gender and ethnic group within this population of at-risk youth. Method Sample Following approval from the University Ethics Review Board, a purposive sample was drawn from across New Zealand of 605 at-risk youth who were clients of the child welfare, juvenile justice, education (support additional to mainstream classrooms) and mental health systems in mandated and non-mandated services including statutory and community providers (see Table 1). In addition to facing elevated individual and contextual risks, all participants were multiple service-using youth; that is, they had been clients of more than one of the four service systems listed in Table 1 in the six months prior to participating in the research. Because there was no national list of multiple service-using youth from which a random sample could be drawn, a community saturation approach was taken to recruitment whereby all organisations that worked with youth in a locality were identified and, with their permission, organisational records were examined to identify potential youth who were current clients. Once an organization’s list had been exhausted, researchers moved to the next organisation and reviewed their client lists to identify additional youth who could be recruited into the study. This generated a study population that was composed of all youth who could be located in each area who were multiple service users at the time of the survey and who were aged between 12 and 17 years. A range of locations were selected that included both rural and urban settings

44

J. Sanders et al. / Child Abuse & Neglect 42 (2015) 40–53

Table 2 Demographic characteristics.

Gender Female Male Age 15 years and under 16+ years Ethnicity ¯ a Maori Pacific Island ¯ Pakeh a¯ b Other ethnicity Total a b

n

%

223 382

36.86 63.14

336 269

55.54 44.46

294 106 188 17 605

48.60 17.52 31.07 2.80 100

The indigenous population of New Zealand. European/White New Zealanders.

Table 3 Living arrangements.

Family (incl. relatives) Family-like setting (e.g. foster care) Residential facility Living independently

n

%

422 58 90 35

69.75 9.58 14.87 5.78

and the demographic characteristics of youth in the final sample was consistent with the patterns we would have expected ¯ given the characteristics of youth in the major service systems. That is; it included an over representation of Maori and Pacific Island youth and of males. Practitioners from each agency first approached youth to seek their permission to meet with a researcher to learn more about the study, researchers then met with youth to seek their consent to participate. Data was collected via a self-report questionnaire administered by research staff with youth. To reduce the potential for administration and response bias, questions were read out to youth. All research staff received detailed training regarding the administration of the questionnaire. The refusal rate was 2.5%. Considerable effort was devoted to ensuring that youth answers to questions accurately reflected the names and service scope of the different services that they had used. This included creation of lists of all the services working with young people in each locality and also lists of names of practitioners in each of the services so that interviewers could accurately allocate youth answers to the correct service. Interviewers were also trained in the jargon used by youth in each locality to refer to services. However, for confidentiality reasons, it was not possible to cross check all 605 youth reports of services with all agency data bases, something that would be required to document actual rates of multiple service use across the sample. Caution is therefore needed in interpreting results. With this caveat in mind, at the time of the interview 307 (50.8%) young people were involved with three or four of the service systems (child welfare, mental health, juvenile justice, education additional to mainstream classroom participation; Table 2). In terms of experiences of risk factors; most youth reported experiences of abuse, maltreatment and neglect. For instance, 61.65% reported current mandated interventions for abuse and neglect and 83% had histories of involvement in child welfare services that featured investigations and interventions related to abuse and neglect. A detailed breakdown of the living arrangements is provided in Table 3. Just over half of the youth (n = 346; 57%) were living with one or both birth parents at the time of the study. Four measures (CES-D-12-NLSY; 4HSQ Delinquency, 4HSQ Risk, SDQ conduct problems–see below) assessed presence of internalising and externalising disorders, and the presence of each of these risk factors in the lives of youth are noted in Table 4. Measures The Pathways to Resilience Youth Measure (PRYM) is an instrument developed at the Resilience Research Centre at Dalhousie University in Halifax, Canada for investigating service use of at-risk youth. This instrument was adapted for use in the New Zealand study (Sanders et al., 2013). Table 5 provides details of means, standard deviations, alpha and CFA mean factor loadings for the continuous variables used in the current analysis. Demographic Variables Age, gender and ethnicity data was collected from all youth. For analysis purposes, the age distribution, of 12–17, years was dichotomised to younger (15 years and younger) and older (16 years and older). This demarcation represented the

J. Sanders et al. / Child Abuse & Neglect 42 (2015) 40–53

45

Table 4 Scores on selected risk measures. Mean CES-D-12-NLSCY At-risk group Comparison group 4HSQ Delinquency At-risk group Comparison group 4HSQ Risk At-risk group Comparison group SDQ Conduct Problems At-risk group Comparison group

SD

10.36 6.92

6.23 4.46

9.53 2.98

6.12 4.04

10.00 4.39

4.98 4.05

4.32 2.15

2.20 1.83

Table 5 Mean, standard deviations, ranges, alpha and mean CFA loadings for continuous variables. Variables Resilience Individual Relational Contextual Individual risk CESD-12 depression 4H risk 4H delinquency SDQ conduct problems Contextual risk Community risk Parental relationships Parental monitoring Education risk Service experience Empowering Respectful Service use Health Mental health Education Child welfare Youth justice Wellbeing SDQ Pro-social Social participation Community School Hobbies/work Future aspirations Positive peer group

Mean

SD

3.86 3.74 3.24

0.56 0.94 0.92

0.87 1.19 1.75 0.79

Total possible score

Alpha

CFA (mean)

5 5 5

0.79 0.81 0.74

0.28 0.61 0.57

0.52 0.62 1.12 0.42

3 3 4 2

0.78 0.81 0.88 0.66

0.5 0.53 0.7 0.5

13.83 7.77 4.7 8.59

3.52 3.59 5.82 3.15

23 20 15 15

0.58 0.78 0.9 0.6

0.58 0.86 0.87 0.58

20 11.77

6.06 2.63

30 15

0.86 0.76

0.72 0.72

5 1.85 2.5 2.6 3.18

1.82 1.8 1.56 1.89 2.41

10 10 10 10 10

0.59 0.68 0.57 0.69 0.84

0.4 0.48 0.22 0.46 0.31

1.54

0.35

2

0.6

0.49

0.92 4.5 9.19 21.05 3.65

0.97 5.07 5.33 6.03 2.54

3 18 27 35 10

0.66 0.73 0.61 0.79 0.86

0.66 0.71 0.48 0.6 0.75

point at which young people are able to leave school in New Zealand. Gender was measured as female and male. Ethnicity ¯ was grouped into four categories, the first three of which accounted for 97% of the study participants; Maori (indigenous ¯ a¯ (white/European, 31.07%), Pacific Island (the next largest ethnic group in the study, 17.52%) and other youth, 48.6%), Pakeh ethnicity (4%).

Resilience The PRYM included the Child and Youth Resilience Measure – 28 (CYRM-28; Liebenberg, Ungar, & Van de Vijver, 2012); an ecological measure which assessed resilience on three sub-scales. Items were rated on a 5-point scale from 1 = Does not describe me at all to 5 = Describes me a lot. The three CYRM sub-scales assess (1) individual resources including personal skills (such as ability to problem solve, cooperation, and awareness of personal strengths), peer support and social skills (2) relationships with parents or primary caregivers including physical and psychological caregiving, and (3) contextual resources that facilitate connection to culture, the role of religious and spiritual beliefs, and engagement with and relevance of education.

46

J. Sanders et al. / Child Abuse & Neglect 42 (2015) 40–53

Individual Risk Both internalising and externalising aspects of individual risk were assessed by questions in the PRYM. This included the 12-item version of the Centre for Epidemiological Studies Depression Scale (CES-D-12-NLSCY; Poulin, Hand, & Boudreau, 2005) which measured risk of depression by asking youth questions about experiences in the past week. Participants rated each item on a 4-point scale from 0 = rarely or none of the time to 3 = all of the time with some items being reverse scored. This measure compared favorably to other depression measures such as the Beck Depression Inventory (Wilcox, Field, Prodromidis, & Scafidi, 1998). Externalising risk was assessed using two sub-scales of the 4-H study of Positive Youth Development (Theokas & Lerner, 2006); Delinquency (frequency of behaviors such as theft, vandalism and aggression in the past year); and Risk (frequency of use of substances including alcohol, tobacco, marijuana and other drugs such as ecstasy, speed, heroin and crack in the past year). Individual items were rated on a 5-point scale from 1 = never to 5 = 5 or more times. Externalising risk was also assessed using the Conduct Problems sub-scale of the SDQ questionnaire (Goodman, 1997, 2001), which included shortness of temper and inclination for aggressive and violent responses, lying, theft and bullying. Items were measured on a 3-point scale from 0 = not true to 2 = certainly true with some items being reverse scored. Contextual Risk Contextual risk includes exposure to acute or chronic adversity within the family, school and community. The PRYM included a range of questions that measured these three aspects of contextual risk. Family Risk was assessed in two ways. First, the nature of relationships with parents was captured by asking about the presence or absence of a mother figure and a father figure as well as the nature of the relationship with these individuals and the amount of affection received from each. Second, it was assessed by asking questions concerned with parental monitoring, which asked youth to identify the level of parent/legal guardian presence when they woke up, returned from school or work, and went to sleep at night by ranking this on a 5 point scale from 0 = never, to 5 = five or more days a week. Education Risk considered youth’s sense of safety at school as well as sense of engagement with education up until the time of the interview, using reverse scored items. Questions explored teacher intervention in violent situations, the extent to which youth considered their school a good place to be, and the educational level they hoped to attain. A composite score measuring Sense of Community Danger was established using items from the Boston Youth Survey (BYS), with some items being reverse scored. Items assessed community cohesion as well as levels of community trust and interaction. Service Use History and Service Use Experience Service use history was measured by 35 questions in the PRYM that counted the total number of services that youth had used over their lifetimes up to the point of the interview. Service systems surveyed included child welfare, juvenile justice (including contact with the police), educational supports beyond regular classroom programming, (such as special education services, alternative education), mental health and general health services. Youth ranked their use of these 35 services on a four point scale from 0 = never needed it, to 3 = 3 or more times. A service quality score was derived from youth answers to 13 questions that assessed the extent to which service delivery adopted PYD components. This included the following components: personal agency (having a say in how the service is provided, relevance and accessibility of the service as well as overall satisfaction with the service) and staff respect (respect and sensitivity for youth and their family including their beliefs, and staff engaging in clear communication with youth). Adapted from the Youth Services Survey, this descriptive measure asked youth to assess the extent to which one service about which they had current experience met PYD criteria. In particular, questions asked youth to rank the extent to which service delivery was responsive to youth situations and whether services engaged appropriately with youth, their families or caregivers and their culture. Items were rated on a 5-point scale from 1 = Strongly disagree to 5 = Strongly agree. There is some evidence (see for example, Stiffman et al., 2000) that youth self-reports of service involvement may be unreliable, therefore in the current study considerable effort was expended to ensure accuracy of this data. Researchers were trained specifically in the names of organisations and the service system to which they belonged (for example, child welfare, juvenile justice, education services and mental health). When youth were unsure of the system to which their nominated service belonged, they were asked to name the service and its location and this was checked against locally available service information and information available on the internet and the appropriate allocation was then made. In addition, where youth were uncertain what the service may have been called, they were asked to name the workers they could remember who had worked with them and the researchers then searched for workers to locate the particular service in which they were employed. Wellbeing Outcomes (Wellbeing) A combined outcome measure was created from questions in the PRYM that captured the wellbeing of youth. While pathways through adolescence can be expected to be diverse and dynamic, subject to a range of internal and external influences, this outcome measure was designed to assess five aspects of youth lives that were considered to be markers of

J. Sanders et al. / Child Abuse & Neglect 42 (2015) 40–53

47

Table 6 MANOVA testing age, gender and ethnicity differences (cells contain effect sizes and p values). Dependent variables 2 , (p)

Source Individual risk Age (A) Gender (G) Ethnicity (E) A×G A×E G×E A×G×E

.010 (.014) .051 (.000)

Contextual risk

Service use history

Service use experience

.014 (.004)

.008 (.032)

.040 (.000)

Resilience

Wellbeing

.040 (.000)

.028 (.001)

.041 (.041) .014 (.039)

a baseline level of wellbeing for youth in New Zealand. Pro-social behavior was measured using the SDQ pro-social behavior subscale (Goodman, 1997, 2001) which assessed youth capacity for kindness, sharing and concern for others on a 3-point scale from 0 = Not true to 2 = Certainly true. Social participation was assessed on three dimensions using 11 questions that measured the extent to which youth were involved in extra-curricular activities, in activities with community organisations and through participation in hobbies and part-time employment. Questions asked youth to rank themselves on a 5 point scale where 1 = Does not describe me at all to 5 = Describes me a lot and to identify the frequency of their involvement in nominated activities. An adapted and reverse-scored list of questions from the fourth and fifth cycles of Statistics Canada’s National Longitudinal Survey of Children and Youth was used to obtain information surrounding positive peer group activity. Future aspirations were measured using two different sets of questions. First, by completion of the Satisfaction with Life measure (Diener, Emmons, Larsen, & Griffin, 1985) in which youth ranked five questions assessing their overall satisfaction with life on a five point scale from 1 = Strongly disagree to 5 = Strongly agree. Second, young people were asked to respond to the following two questions: “Most of the time I think things will work out OK for me”, and “I have some clear ideas about what the next 5 years might be like for me”. Youth were asked to rank their answers to these questions on a five point scale where 1 = Does not describe me at all to 5 = Describes me a lot. Educational involvement was assessed with an answer to a single yes/no question that asked if youth were enrolled in any school subjects at the time of the survey. Youth did not have to be attending a mainstream school to answer yes to this question. For example, they could be enrolled in correspondence school or attending an alternative education program in their local community. Data Analysis Missing data analysis showed data to be missing completely at random (MCAR). Missing values were replaced using maximum likelihood estimation (EM; Enders, 2010). Continuous variables were examined for normality, linearity, and homoscedasticity; all were within acceptable limits. The analysis is concerned with understanding the relationship between patterns of risk, service utilisation, resilience and outcomes. The analysis adopted a multi-step procedure. In order to confirm that the variables used in the analysis were able to be reduced into the selected measures (individual risk, contextual risk, service use, service experience, resilience and wellbeing), the measurement model of each sub-scale was validated using Confirmatory Factor Analysis (CFA) of the raw items for each scale. All fit statistics were satisfactory providing confidence that the variables could be reduced in this way (see Table 5 for CFA mean factor loadings for each subscale). Exploratory factor analysis (EFA) using oblique rotation (Direct Oblimin) supported mean clustering of theoretical components of the model (see previous discussion of instruments grouped by theoretical component measured). Scale scores were therefore included in the analysis as they reflected theoretical components of the proposed model (see Table 5). Path analysis allows for the theoretical and empirical investigation of the structure of complex relationships between exogenous and endogenous variables (Byrne, 2009). A path analysis was therefore conducted to allow an exploration of the possible mediating role of resilience between the risks in youth environments and services received with the wellbeing outcome measure. Finally, a MANOVA was conducted to assess the differences in experience of risk, resilience, service use and outcomes for youth by demographic variables. The data were analysed using PASW Statistics 22 and AMOS 22 for Windows. Results We tested a model in which the four exogenous variables: contextual risk, individual risk, service experience and service use history were allowed to correlate. Resilience was the mediating variable and wellbeing was the outcome variable. Results of this full mediation model showed a poor fit. Modification indices suggested that risk factors (individual and contextual) and service quality, variables already in the model, were directly related to the wellbeing outcome measure. Once these direct relationships were added to the model, an excellent fit was obtained (2 (1, N = 605) = .426, p = .514, AGFI = .995, CFI = 1.00, RMSEA = .000) (see Fig. 1). MANOVA was then used to test for differences in age (2 levels), gender (2 groups) and ethnicity (4 groups) on all 6 variables in the path analysis (Table 6) in order to understand how demographic variables impacted youths’ experiences of the various components in the model and if observed differences were significant. The six variables were treated as dependent variables,

48

J. Sanders et al. / Child Abuse & Neglect 42 (2015) 40–53

Fig. 1. Path analysis model of risk, service use, resilience and wellbeing with standardised maximum likelihood effects.

while age, gender, and ethnicity were fitted as independent variables. Starting with a full model including all possible interactions, backward selection and Type I sums of squares were used. Any variables that had significant interactions were kept in the final model even if the variable was not significant on its own. All multivariate main effects were significant [age: Wilks’ Lambda = .950, F(6, 584) = 5.115, p ≤ .001, 2 = .050; gender: Wilks’ Lambda = .965, F(5, 584) = 3.571, p = .002, 2 = .035; ethnicity: Wilks’ lambda = .816, F(18, 1652.287) = 6.821, p ≤ .001, 2 = .065]. Triple interaction of the three main effects was significant [Wilks’ Lambda = .949, F(18, 1652.287) = 1.716, p = .031, 2 = .017]. As can be seen, the partial Eta squared showing the effect size of this triple interaction however was small (Cohen, 1988). Double interactions were not significant [age and ethnicity: Wilks’ Lambda = .974, F(18, 1652.287) = .875, p = .610; age and gender: Wilks’ Lambda = .988, F(6,584) = 1.138, p = .338, ethnicity and gender: Wilks’ Lambda = .953, F(18, 1652.287) = −1.571, p = .059]. The univariate test showed that older and younger youth were similar in terms of their scores on individual and contextual risk, resilience and wellbeing. They differed significantly in regard to their service use histories and their experience of service quality. Older youth reported greater service use histories (M = 3.18, SD = 1.20, p = .004) than younger youth (M = 2.91, SD = 1.14,) and were more likely to report receiving services delivered using PYD approaches (M = 16.27, SD = 3.94, p = .032) than were younger youth (M = 15.56, SD = 4.10). In terms of gender differences, males reported higher individual risk (M = 1.19, SD = .50, p= .014) than females (M = 1.10, SD = .50). There were no other gendered differences. ¯ The univariate effect of ethnicity was also significant for all variables with the exception of service use experience. Maori ¯ youth reported higher individual risk (M = 1.26, SD = .49, p ≤ .001) than Pakeh a¯ youth (M = .99, SD = .49). Pacific Island youth ¯ ¯ (M = 3.20, SD = 1.21, p ≤ .001) and Pakeh a¯ reported shorter service use histories (M = 2.60, SD = 1.15) compared to both Maori ¯ ¯ youth (M = 3.04, SD = 1.06, p = .010). Pakeh a¯ youth reported lower wellbeing (M = 5.53, SD = 1.51) than both Maori (M = 6.06, ¯ SD = 1.80, p = .003) and Pacific Island youth (M = 6.44, SD = 1.67, p = ≤.001). Pakeh a¯ youth also reported lower resilience ¯ (M = 3.455, SD = 0.61) than both Pacific Island (M = 3.79, SD = 0.65, p ≤ .001) and Maori youth (M = 3.66, SD = 0.58, p = .005). The triple interaction between age, gender and ethnicity was significant for the service use experience variable only, but the effect size for this relationship (2 = 017) was weak (Cohen, 1988).

J. Sanders et al. / Child Abuse & Neglect 42 (2015) 40–53

49

Discussion The results from the path analysis point to a complex set of relationships between service delivery and risk factors, and further, that the ways that these factors combine to interact with resilience is also complicated. The analysis suggests that resilience does, as hypothesised, potentially mediate the impact of risks and service delivery factors on wellbeing outcomes. Specifically, the model shows that within this group of very vulnerable youth, who have been exposed to high levels of abuse and neglect by the adults in their families and communities, those who received services that were empowering and respectful, reported enhanced resilience and this higher resilience was associated with better wellbeing outcomes. PYD intervention qualities such as encouraging personal agency and staff being respectful do appear to be related to higher levels of resilience for vulnerable youth. The model suggests that PYD approaches are related to outcomes for vulnerable youth in two ways. First, these approaches appear to exert a small direct influence on outcomes. Second, and more importantly, PYD intervention approaches appear to have a strong positive influence on outcomes that is mediated through the role these interventions play in enhancing youth resilience processes. Services that utilise PYD approaches are therefore likely to produce better outcomes through their work with vulnerable youth than services that do not utilise such approaches because of these direct and mediated impacts. This approach appears to work regardless of the service system involved and irrespective of whether services are delivered in a statutory or non-statutory, mandated or non-mandated context because data used in this analysis came from organisations located in the mental health, child welfare, education and juvenile justice systems and included mandated, non-mandated, statutory services and community provision. The model does raise some concerns regarding the relationship between high risks in general and contextual risks in particular and service delivery. Young people facing the highest contextual and individual risks had a greater likelihood of receiving more services, as might be expected given the greater risks they were facing. However, they also seemed less likely to experience services as empowering and respectful (PYD approaches). Furthermore, contextual risks had a strong negative relationship with resilience processes. This implies that youth confronting the highest contextual risks likely came to services with depleted resilience resources and they were then less likely to receive PYD interventions that would boost their resilience leaving them less able to achieve positive outcomes. Because they were less likely to receive PYD interventions they were also less likely to experience the (smaller) direct increase in outcomes from the service experience identified in the model. MANOVA indicated that there were some minor differences between the different demographic groups on some of the variables used in the path analysis, providing limited support for the second hypothesis which proposed that risk, resilience, service experiences and wellbeing outcomes would be experienced differently by youth depending on their age, gender and ethnic group. The deviations that were observed in the MANOVA suggest some areas for particular focus in terms of service delivery with specific sub-groups of at-risk youth. Specifically, older youth reported better service experiences than younger youth. Given that the service quality measure emphasised practices that encouraged active youth involvement; this result may be attributable to service providers providing more scope for older youth to be actively involved in the intervention process. Alternatively, given that older youth were more likely to have had longer service histories, it might reflect that over time and with greater experience as a service recipient youth learn strategies to engage with providers in ways that increase the chances that interventions will meet their needs. However, it might also represent a tendency by service professionals to accord lesser priority to engaging younger youth and providing them with opportunities to jointly make decisions through treatment. Higher wellbeing outcome levels for older youth were also notable and may reflect maturation processes. Alternatively, given the lower levels of service satisfaction reported by younger youth noted above, it might point to interventions that are less likely to meet the particular needs of younger youth who confront high levels of risk. Across general populations, age-related increases in risk behaviors through adolescence have been observed (Schulenberg, Wadsworth, O’Malley, Bachman, & Johnston, 1996). In the current study risk levels did not increase significantly with age. It is possible that the risks confronted by the youth in the current study emerged early in life and persisted over time explaining the absence of significant differences in levels of risk across the age span (Schulenberg, Maggs, & O’Malley, 2003). Consistent with findings reported elsewhere (Berzin, 2010) males reported more individual risk than females. Given that males are overrepresented in the youth justice population in New Zealand and many of the risk questions assessed behaviors that would bring youth into contact with the justice system this pattern is not surprising. The absence of gender-based differences on the contextual risk measures possibly reflects the fact that this group of at-risk young people lived in relatively homogenous communities where neighborhood risks influenced the lives of all youth. In terms of ethnicity, some MANOVA results confirm what is known about youth more generally in New Zealand. In ¯ particular, Maori youth feature disproportionately in national statistics on key indicators of distress such as crime, abuse ¯ youth reported significantly higher and risk behaviors (Johnson, 2013) and consistent with this in the current study Maori levels of individual risk. As observed in relation to risk profiles based on gender, the similarity across all four ethnic groups on the contextual risk measures suggests a concentration of at-risk youth in particular communities that face high levels of disadvantage and endemic distress. Pacific Island youth reported involvement in fewer services over their lifetimes than ¯ ¯ either Maori or Pakeh a¯ youth. This is of concern given that they did not confront significantly fewer risks than the other youth surveyed. It is not clear why this would be the case and the pattern suggests a need to consider service delivery targeted at the specific needs of Pacific youth. The absence of differences in service satisfaction by gender or ethnicity is a positive result suggesting that service delivery was targeted to youth culture and context. This contrasts with findings in other jurisdictions

50

J. Sanders et al. / Child Abuse & Neglect 42 (2015) 40–53

(Blackstock & Trocmé, 2005) that point to a lack of sensitivity to culture and context by service providers in their engagement with at-risk youth. It is possible that several decades of concerted effort on building cultural competence in practitioners and culturally responsive service provision in New Zealand (De Souza, 2008) has produced some benefits for at-risk youth. While there were no differences across ethnic groups in terms of quality of service experiences, there were significant differences within this population of at-risk youth in terms of the resilience and wellbeing outcome measures. The compar¯ ative advantage on wellbeing and other outcomes apparent in the general population for Pakeh a¯ (white/European, majority) youth in New Zealand (Pearson, 2012) was not apparent when considering this sub-group of youth who had been exposed ¯ ¯ a¯ youth reported significantly lower wellbeing and resilience compared to both Maori to significant adversity. Rather, Pakeh ¯ and Pacific Island youth. It appears that there may be cultural resources here that are protective for both Pacific and Maori youth within the population of disadvantaged youth that give them a relative benefit compared to their similarly at-risk ¯ Pakeh a¯ peers and this is particularly important given the possible mediating role that resilience played between risks, services and outcomes. There are clearly implications here for how services are delivered, in terms of the focus that providers have in regard to different aspects of youth lives depending on their cultural connections. In particular, it suggests a need ¯ ¯ youth. Work with Pakeh a¯ youth might benefit from a focus on for attention to be paid to risk when working with Maori building resilience resources and on the components that contribute to wellbeing outcomes, while work with Pacific youth may need to address specific aspects of service engagement and quality. Limitations The study reported here is cross-sectional in design and therefore causality cannot be demonstrated. The study provides a novel opportunity to gain insight into the relationship between risks, resilience and service use among young people who were using multiple services at the time of sampling. It is not known, however, whether such youth are a special population whose characteristics and challenges influence their patterns of service engagement, their risks and resilience or if the focus by service providers on individual risk factors creates the conditions under which these youth become clients of multiple services. It was not possible to generate a national list of all eligible multiple service-using youth from which a random sample could be drawn for the study. Generalizability of findings to general populations is therefore limited and the sampling methods that were adopted pose some challenges to the external validity of findings. Finally, the sample was drawn from the client lists of service providers in order to ensure that the at-risk group was composed of the youth who were facing the most adversity in their neighborhoods. It is possible, although unlikely, that the most at-risk youth were in fact not receiving any services and therefore would not be represented in the study. Implications for Practice The systematic investigation of the ways in which multiple service systems can work together to enhance resilience of vulnerable youth and to minimise the risks they confront is in its infancy. Evidence is beginning to show that when multiple services are well co-ordinated improvements in the circumstances of such youth can be achieved (Garland et al., 2003; Stathis et al., 2006). The current study suggests that the quality of service delivery, rather than the numbers of services provided, may be more important in building resilience, leading to better outcomes for youth with complex needs. Having contact with more services over the lifetime did make a contribution to resilience but it did not on its own contribute to better wellbeing outcomes. On the other hand, experiencing services that used PYD approaches such as the encouragement of personal agency, and empowering and respectful practices, made a (small) direct contribution to wellbeing outcomes but more importantly, it enhanced resilience which in turn played a very powerful role in the achievement of better wellbeing outcomes. This important finding suggests that service quality rather than simply the numbers of services involved with youth is the key to achieving good outcomes for youth exposed to numerous risks and with complex needs. That is, when services worked to promote and build upon the positive resources already present in youth lives they made a large contribution to wellbeing outcomes. This requires that practitioners: • recognise and take advantage of the positive resources youth bring into interventions; • harness these strengths in the support process; and • identify when resources are missing and compensate for these. These types of practices are likely to achieve superior outcomes. Focusing on how services are delivered is therefore an important part of securing good outcomes for youth facing significant challenges. Furthermore, services that are able to address a range of risks (contextual and individual) and provide interventions that address a number of issues may be more effective than a numerous providers each addressing specific risks and challenges individually. Of particular note in the current study was the significant decrease in service quality for youth which the highest risks (both contextual and individual). The path analysis suggests that systems may respond to higher risks by allocating more services to youth. However, it was the quality not the quantity of services that had the largest impact on levels of youth resilience and on positive outcomes. The analysis presented here also points to a need for greater attention to be paid to the impact which contextual risks (the risks youth confront in their neighborhoods, at school and in their families) have upon youth capacity to cope with stress

J. Sanders et al. / Child Abuse & Neglect 42 (2015) 40–53

51

and to achieve positive outcomes. Contextual risks are more difficult to identify and require that practitioners work with families, friends, educators and others in the community as well as youth themselves. Individual risks, on the other hand, are those behaviors which bring youth quickly to the attention of a range of organisations and the community at large and do not generally directly implicate others. It would appear that service providers were able to respond to these behaviors but less able to work to address the risks in young people’s social environments. The path analysis indicated that individual risks had a much stronger relationship with history of service use than did contextual risks. Given that higher levels of individual risk were associated with higher levels of contextual risk and that, at least in part, these individual risks had their origins in contextual risks, focusing attention on contextual risks may well provide opportunities for a more systemic response to the challenges presented by youth with complex needs. That is, increased attention to contextual risks may result in reductions in individual risks. The tendency for services to respond more readily to individual than contextual risks may reduce service effectiveness and certainly in the current study, contextual risks had a very strong inverse relationship with resilience. This pattern may explain the relatively positive results achieved by more intensive home, school and community-based interventions (Berzin, 2010; Thornton et al., 2000) which are arguably in a strong position to address both contextual and individual risks in the same intervention. Resilience played an important role in increasing the capacity of young people to achieve positive outcomes when services were involved in their lives. This analysis suggests a powerful role for services that are able to adopt empowering and respectful intervention styles in terms of their potential to impact upon outcomes through the influence they have on youth resilience. On the other hand, the model suggests that those services that are not able to work in empowering and respectful ways may be exacerbating the risks youth confront, or at the very least not producing benefits in terms of enhanced outcomes. Finally, as youth faced greater risks (individual and contextual) their experience of service quality dropped. More detailed understanding of the ways in which providers respond to risk and harness resilience, and in particular, of the types of service responses that are most effective in achieving this, is required and this is a focus of further work in this research program. Conclusion This analysis of data from vulnerable young people who experience multiple concurrent service system involvement suggests that better quality services rather than more services make the most difference to risk reduction, resilience enhancement and improved wellbeing outcomes. While coordination across services and accessibility are clearly important components of effective service delivery to these youth (Cheon, 2008; Hussey, Drinkard, & Flannery, 2007), the findings from this study constitute a timely reminder that the quality of interactions professionals have with youth do make an important difference. In particular, professional practices that provide spaces for youth engagement and decision making and that work in respectful ways taking account of youth circumstances, bring benefits in terms of enhanced resilience (the availability of resources around youth that help them respond to crises and solve problems) and this improves youth wellbeing outcomes. Acknowledgements The authors thank Ministry of Business, Innovation and Employment for funding this research programme. They also gratefully thank all the young people and their supporters who participated in the research. Finally they acknowledge the contribution of The Donald Beasley Institute, The Research Trust of Victoria University and its staff, Youthline and all the other researchers who helped with the research. References ˜ Beale Spencer, M., Harpalani, V., Cassidy, E., Jacobs, C., Donde, S., Goss, T., Munoz-Miller, M., Charles, N., & Wilson, S. (2006). Understanding vulnerability and resilience from a normative developmental perspective: Implications for racially and ethnically diverse youth. In D. Cicchetti, & D. Cohen (Eds.), Developmental psychopathology: Theory and method (Vol. 1) (pp. 627–672). Hoboken, NJ: John Wiley & Sons, Inc. Berzin, S. C. (2008). Difficulties in the transition to adulthood: Using propensity scoring to understand what makes foster youth vulnerable. Social Service Review, 82, 171–196. Berzin, S. (2010). Vulnerability in the transition to adulthood: Defining risk based on youth profiles. Children and Youth Services Review, 32, 487–495. Blackstock, C., & Trocmé, N. (2005). Community-based child welfare for Aboriginal Children: Supporting resilience through structural change. In M. Ungar (Ed.), Handbook for working with children and youth: Pathways to resilience across cultures and contexts (pp. 105–120). Thousand Oaks, CA: Sage. Botticello, A. (2009). A Multilevel analysis of gender differences in psychological distress over time. Journal of Research on Adolescence, 19(2), 217–247. Bottrell, D. (2009). Understanding ‘marginal’ perspectives: Toward a social theory of resilience. Qualitative Social Work: Research & Practice, 8, 321–339. Bullis, M., & Yovanoff, P. (2002). Those who do not return: Correlates of the work and school engagement of formerly incarcerated youth who remain in the community. Journal of Emotional and Behavioral Disorders, 10(2), 66–79. Byrne, B. (2009). Structural equation modeling with AMOS: Basic concepts, applications, and programming (2nd ed.). New York: Routledge/Taylor & Francis. Capaldi, D. (1992). Co-occurrence of conduct problems and depressive symptoms in early adolescent boys: II. A 2-year follow-up at Grade 8. Development and Psychopathology, 4, 125–144. Case, S. (2006). Young people ‘at risk’ of what? Challenging risk-focused early intervention as crime prevention. Youth Justice, 6, 171–179. Cheon, J. (2008). Best practices in community-based prevention for youth substance reduction: Towards strengths-based positive development policy. Journal of Community Psychology, 36, 761–779. Cohen, J. (1988). Statistical power analysis for the behavioural sciences. Hillsdale, NJ: Erlbaum. Courtney, M., & Dworsky, A. (2006). Early outcomes for young adults transitioning from out-of-home care in the USA. Child and Family Social Work, 11(3), 209–219. Department of Labour. (2012). Youth labour market fact sheet. Wellington, New Zealand: Department of Labour.

52

J. Sanders et al. / Child Abuse & Neglect 42 (2015) 40–53

De Souza, R. (2008). Wellness for all: The possibilities of cultural safety and cultural competence in New Zealand. Journal of Research in Nursing, 13(2), 125–135. Diener, E., Emmons, R., Larsen, R., & Griffin, S. (1985). The satisfaction with life scale. Journal of Personality Assessment, 49, 71–75. Duncan, B., Miller, S., & Sparks, J. (2004). The heroic client. New York, NY: Jossey-Bass. Eccles, J., Barber, B., Stone, M., & Hunt, J. (2003). Extracurricular activities and adolescent development. Journal of Social Issues, 59, 865–889. Enders, C. (2010). Applied missing data analysis. New York, NY: The Guildford Press. Fagan, J., & Freeman, R. (1999). Crime and work. In M. Tonry (Ed.), Crime and justice: A review of research (Vol. 25) (pp. 225–290). Chicago, IL: University of Chicago Press. Fergusson, D., & Horwood, L. (2003). Resilience to childhood adversity: Results of a 21 year study. In S. S. Luthar (Ed.), Resilience and vulnerability: Adaptation in the context of childhood adversities (pp. 130–155). New York: Cambridge University Press. Garland, A., Hough, R., Landsverk, J., & Brown, A. (2001). Multi-sector complexity of systems of care for youth with mental health needs. Children’s Services: Social Policy, Research and Practice, 4, 123–140. Garland, A., Aarons, G., Brown, S., Wood, P., & Hough, R. (2003). Diagnostic profiles associated with use of mental health and substance abuse services among high-risk youths. Psychiatric Services, 54, 562–564. Gilligan, R. (2004). Promoting resilience in child and family social work: Issues for social work practice, education and policy. Social Work Education, 23(1), 93–104. Goerge, R., Bilaver, L., Lee, B., Needell, B., Brookhart, A., & Jackman, W. (2002). Employment outcomes for youth aging out of foster care. Washington, DC: US Department of Health and Human Services, Office of the Assistant Secretary for Planning and Evaluation. Goodman, R. (1997). The strengths and difficulties questionnaire: A research note. Journal of Child Psychology and Psychiatry, 38, 581–586. Goodman, R. (2001). Psychometric properties of the strengths and difficulties questionnaire. Journal of the American Academy of Child and Adolescent Psychiatry, 40, 1337–1345. Haapasalo, J. (2000). Young offenders’ experiences of child protection services. Journal of Youth and Adolescence, 29, 355–371. Hazen, A., Hough, R., Landsverk, J., & Wood, P. (2004). Use of mental health services by youths in public sectors of care. Mental Health Services Research, 6, 213–226. Heinze, H. (2013). Beyond a bed: Support for positive development for youth residing in emergency shelters. Children and Youth Services Review, 35, 278–286. Hussey, D., Drinkard, A., & Flannery, D. (2007). Comorbid substance use and mental disorders among offending youth. Journal of Social Work Practice in the Addictions, 7, 117–138. Johnson, A. (2013). She’ll be right: A state of the nation report from the Salvation Army. New Zealand: The Salvation Army, Social Policy and Parliamentary Unit. Kroll, L., Rothwell, J., Bradley, D., Shah, P., Bailey, S., & Harrington, R. (2002). Mental health needs of boys in secure care for serious or persistent offending: A prospective, longitudinal study. Lancet, 359, 1975–1979. Lerner, R., Almerigi, J., Theokas, C., & Lerner, J. (2005). Positive youth development: A view of the issues. Journal of Early Adolescence, 25(1), 10–16. Lerner, R. (2005). Promoting positive youth development: Theoretical and empirical bases. White paper prepared for: Workshop on the science of adolescent health and development. Washington DC: National Research Council. Lerner, R. (2006). Resilience as an attribute of the developmental system: Comments on the papers of Professors Masten & Wachs. In B. Lester, A. Masten, & B. McEwen (Eds.), Resilience in children (pp. 40–51). Boston, MA: Blackwell. Liebenberg, L., Ungar, M., & Van de Vijver, F. R. R. (2012). Validation of the Child and Youth Resilience Measure-28 (CYRM-28) among Canadian youth with complex needs. Research on Social Work Practice, 22(2), 219–226. http://dx.doi.org/10.1177/1049731511428619 Lipsey, M. (2006). The primary factors that characterise effective interventions with juvenile offenders: A meta-analytic overview. Victims & Offenders, 42(2), 124–147. http://dx.doi.org/10.1080/15564880802612573 Loeber, R., Farrington, D., Stouthamer-Loeber, P., & Van Kammen, M. (1998). Multiple risk factors for multi-problem boys: Co-occurrence of delinquency, substance use, attention deficit, conduct problems, physical aggression, covert behavior, depressed mood and shy/withdrawn behavior. In R. Jessor (Ed.), New perspectives in adolescent risk behavior (pp. 90–148). New York, NY: Cambridge University Press. Mechanic, D., & Tanner, J. (2007). Vulnerable people, groups, and populations: Societal view. Health Affairs, 5, 1220–1230. Mitchell, P. (2011). Evidence-based practice in real-world services for young people with complex needs: New opportunities suggested by recent implementation science. Children and Youth Services Review, 33, 207–216. Mossakowski, K. (2003). Coping with perceived discrimination: Does ethnic identity protect mental health? Journal of Health and Social Behavior, 44, 318–331. Pearson, D. (2012). Ethnic inequalities – Race, ethnicity and inequality. In Te Ara – the Encyclopedia of New Zealand. http://www.TeAra.govt.nz/en/ ethnic-inequalities/page-1 (updated 13.06.12) Pittman, K. (1999). Youth today: The power of engagement. Forum for Youth Investment. Retrieved from. http://forumfyi.org/content/ youth-today-power-enga Poulin, C., Hand, D., & Boudreau, B. (2005). Validity of a 12-item version of the CES-D used in the National Longitudinal Study of Children and Youth. Chronic Diseases in Canada, 26(2/3), 65–72. Rogers, R. (2011). ‘I remember thinking, why isn’t there someone to help me? Why isn’t there someone who can help me make sense of what I’m going through?: Instant adulthood and the transition of young people out of state care’. Journal of Sociology, 47(4), 411–426. Sanders, J., & Munford, R. (2015). The interaction between culture, resilience, risks and outcomes: A New Zealand study. In L. Theron, M. Ungar, & L. Liebenberg (Eds.), Youth, Resilience and Culture - Commonalitieis and Complexities. USA: Springer. Sanders, J., Munford, R., Thimasarn-Anwar, T., Liebenberg, L., Ungar, M., Osborne, A., Dewhurst, K., Youthline New Zealand Henaghan, M., Mirfin-Veitch, B., Tikao, K., Aberdein, J., Stevens, K., & Urry, Y. (2013). Methodological Overview: The Pathways to Resilience Study Technical Report 2. New Zealand: Massey University. Available online at www.youthsay.co.nz Scales, P., Benson, P., & Mannes, M. (2006). The contribution to adolescent wellbeing made by nonfamily adults: An examination of developmental assets as contexts and processes. Journal of Community Psychology, 34, 401–413. Schofield, G., & Beek, M. (2009). Growing up in foster care: providing a secure base through adolescence. Child and Family Social Work, 3(14), 255–266. Schulenberg, J., Maggs, J., & O’Malley, P. (2003). How and why the understanding of developmental continuity and discontinuity is important: The sample case of long-term consequences of adolescent substance use. In J. Mortimer, & M. Shanahan (Eds.), Handbook of the life course (pp. 413–436). New York, NY: Plenum. Schulenberg, J., Wadsworth, K., O’Malley, P., Bachman, J., & Johnston, L. (1996). Adolescent risk factors for binge drinking during the transition to young adulthood: Variable- and pattern-centered approaches to change. Developmental Psychology, 32, 659–674. Sellers, R., & Shelton, J. (2003). The role of racial identity in perceived racial discrimination. Journal of Personality and Social Psychology, 84, 1079–1092. Sellers, R., Caldwell, C., Schmeelk-Cone, K., & Zimmerman, M. (2003). Racial identity, racial discrimination, perceived stress, and psychological distress among African American young adults. Journal of Health and Social Behavior, 44, 302–317. Stathis, S., Letters, P., Doolan, I., & Whittingham, D. (2006). Developing an integrated substance use and mental health service in the specialised setting of a youth detention centre. Drug and Alcohol Review, 25, 149–155. Stein, M., Ward, H., & Courtney, M. (2011). ‘Editorial: International perspectives on young people’s transitions from care to adulthood’. Children and Youth Services Review, 33(12), 2409–2411. Stiffman, A., Hadley-Ives, E., Doré, P., Polgar, M., Horvath, V., Striley, C., & Elze, D. (2000). Youths’ access to mental health services: The role of providers’ training, resource connectivity, and assessment of need. Mental Health Services Research, 2, 141–154. Sveinsson, K. (Ed.). (2012). Criminal Justice vs Racial Justice: Minority ethnic overrepresentation in the criminal justice system. London: Runnymeade Trust.

J. Sanders et al. / Child Abuse & Neglect 42 (2015) 40–53

53

Theokas, C., & Lerner, R. (2006). Observed ecological assets in families, schools, and neighborhoods: Conceptualization, measurement and relations with positive and negative developmental outcomes. Applied Developmental Science, 10(2), 61–74. Thornton, T., Craft, C., Dahlberg, L., Lynch, B., & Baer, K. (2000). Best practices of youth violence prevention: A sourcebook for community action. Atlanta, GA: Centers for Disease Control and Prevention, National Center for Injury Prevention and Control. Twenge, J., & Nolen-Hoeksema, S. (2002). Age, gender, race, socioeconomic status, and birth cohort difference on the Children’s Depressive Inventory: A meta-analysis. Journal of Abnormal Psychology, 111(4), 578–588. Ungar, M. (2011). The social ecology of resilience: Addressing contextual and cultural ambiguity of a nascent construct. American Journal of Orthopsychiatry, 81, 1–17. Ungar, M., Liebenberg, L., Dudding, P., Armstrong, M., & Van de Vijver, F. (2013). Patterns of service use, individual and contextual risk factors, and resilience among adolescents using multiple psychosocial services. Child Abuse and Neglect, 37, 150–159. Wagner, M., Newman, L., Cameto, R., Garza, N., & Levine, P. (2005). After high school: A first look at the postschool experiences of youth with disabilities. A Report from the National Longitudinal Transition Study-2 (NLTS2). Menlo Park, CA: SRI International. Wong, C., Eccles, J., & Sameroff, A. (2003). The influence of ethnic discrimination and ethnic identification on African American adolescents’ school and socioemotional adjustment. Journal of Personality, 71(6), 1197–1232. Wiesner, M. (2003). A longitudinal latent variable analysis of reciprocal relations between depressive symptoms and delinquency during adolescence. Journal of Abnormal Psychology, 112, 633–645. Wilcox, H., Field, T., Prodromidis, M., & Scafidi, F. (1998). Correlations between the BDI and CES-D in a sample of adolescent mothers. Adolescence, 33, 565–575.