Victimization, Poverty, and Resilience Resources: Stress Process Considerations for Adolescent Mental Health

Abstract

Growing evidence suggests that exposure to early life adversity poses risk to youth development, with impaired mental health a central concern. This population-representative study of adolescent students (n=11,222) investigates the effects of two key forms of early adversity— victimization and poverty—on adolescent mental health in a step-wise fashion, also accounting for mutable and accessible resilience resources. Victimization and poverty prevalence reflected social patterning wherein being female, racial and ethnic minority youth, and those with lower resilience resources all reported significantly higher levels of victimization and family poverty. Greater levels of these adversities were significantly associated with lower levels of resilience resources. Poverty and particularly victimization demonstrated significant cumulative and distinct contributions across three indicators of compromised mental health—depression, suicidality, and broader psychological well-being. Resilience resources of family bondedness, school engagement, and sleep sufficiency all carried significant effects and accompanied lesser explanatory strength of victimization and family poverty. In separate analyses, each of four forms of victimization—adult maltreatment, bullying, dating violence, and feeling unsafe at school—were significant contributors to mental health, with cumulative exposure conveying the strongest unique effects. Implications and opportunities for prevention and remedial strategies are discussed, with particular attention to school-based responding.

Childhood adversity is a significant social and public health problem, with roles in the development of psychological disorders not only in childhood and adolescence but well into adulthood (Green et al., 2010; McLaughlin et al., 2012). Roughly 80% of children and adolescents in the US have experienced adversity in the form of victimization (Turner, Finklehor, & Ormrod, 2010) and approximately 43% of children live in low-income families (NCCP, 2018). Approximately 1 in 5 youth are estimated to have a diagnosable mental health disorder that carries life course implications. Yet the great majority do not receive mental health services, particularly minority and lower socioeconomic individuals (Merikangas, He, Burstein, et al., 2011). Left unattended, mental health disorders can contribute to what become secondary, cascading stressors throughout adolescence and adulthood—such as erosion of success in academic functioning, in the workforce, and in social and partnering relationships (Michael, Merlo, Basch, Wentzel, & Wechsler, 2015) as well as elevated suicide risk (World Health Organization, 2018).

The transition from adolescent to adulthood is one that comes with a variety of challenges. Youth are faced with developing independence skills that prepare them for autonomous adult roles within society. Transitional outcomes include areas such as academic achievement, employment, relationship attainment and management, and community engagement (Davis, 2003). For those experiencing mental health issues, this transitional period is a particularly vulnerable time (Davis, 2003; Schulenberg, Sameroff, & Cicchetti, 2004). With this vulnerability in mind, the present study draws from theorizing regarding cumulative adversity to assess stress contributions of victimization and poverty on youth as well as mutable and accessible health-promotive resources.

Research on long-term implications of early life stress within life-course approaches has directed attention to cumulative adversity beginning early in life and to mechanisms through which stress is processed. The stress process model has been applied to integrate psychosocial and neurobiological models of stress impact linking early adversity such as adverse childhood experiences to adult outcomes such as mental health (e.g., Edwards, Holden, Felitti, & Anda, 2003; Nurius, Green, Logan-Greene, & Borja, 2015) and adolescent mental health (Logan-Greene, Tennyson, Nurius, & Borja, 2017). This has reframed thinking about adult chronic conditions as developmental disorders in the life course sense (Shonkoff, Boyce, & McEwen, 2009). There are increasing calls for child and youth-serving practice, policy, and systems toward earlier life preventive and resilience-building supports to address youth well-being, but to also guard against erosion of adulthood health, productivity, and psychosocial functioning (Shonkoff et al., 2012).

Exposure to stressors such as interpersonal maltreatment, particularly those cumulative in nature, can lead to alteration of neurodevelopmental trajectories that influence healthy cognitive and socioemotional abilities. These, in turn, contribute to poor physical and mental health outcomes that can have transgenerational transmission implications (McLaughlin, Sheridan, & Lambert, 2014). In applying a stress process model to children’s exposure to violence, Foster & Brooks-Gunn (2009) reinforce the value of assessing multi-form exposures, of social determinant correlates such as poverty, and the importance of intermediary resources that can support coping and buffering and thereby serve as resilience resources in the context of adversities. Similarly, research on developmental victimology highlights the co-occurrence of multiple forms of victimization as well as re-victimization at later points in life (Finkelhor, Ormrod, & Turner, 2007).

As children are exposed to adversities, these disadvantages in development increase the likelihood of engagement in health risk behaviors and being exposed to situations that threaten personal safety, which then increases the odds of poor health outcomes later in life (Slack, Front & Jones, 2017). At the same time, evidence is indicating the value of resilience resources to attenuate the effects of victimization and other stressors and foster resilience (Hamby, Grych, & Banyard, 2018.) Social support, for example, is a broad spectrum resilience resource relative to home and community violence (Foster et al., 2009) as well as to perceived safety which itself reduces effects of violence (Overstreet & Baum, 2000; Shields, Nadasen & Pierce, 2008).

Shaping youth’s social environments includes school engagement, parental involvement, and creating culturally rich communities to support well-being and improve childhood resilience (Ungar, 2014). Positive school environments play a key role in promoting both educational attainment and emotional health resilience in response to early adversity (Khambati, Mahedy, Heron & Emond, 2018). Further, stable family environments and supportive relationships are associated with resilience following childhood maltreatment (Afifi & MacMillan, 2011).

Although a number of health behaviors have been examined in relation to adolescent mental health, sleep is becoming a growing target of attention (Meerlo, Sgoifo, & Suchecki, 2008). Life adversity has been associated, for example, with difficulty in calming to achieve sleep and with poor sleep (Chapman et al., 2011). Adolescents with sleep insufficiency have greater risk of academic failure (Titova et al, 2015), with poor sleep associated with disorders such as anxiety (Leahy & Gradisar, 2012) and depression (Lovato & Gradisar, 2014). These trends indicate a lesser availability of sleep as a stress buffer, with sleep insufficiency potentially exacerbating effects of current stress (Greenfield, Lee, Friedman, & Springer, 2011). Sleep is, thus, a mutable, broadly accessible resilience resource, supporting individual defenses against stressors and the reduction of dysregulations such as heightened reactivity to later stress (Chatburn, Coussens, & Kohler, 2015).

The Present Study

This study builds on growing recognition regarding the role of stressors in contributing to psychological distress. These stressors include chronic sources, such as socioeconomic disadvantage, and interpersonal, such as exposure to victimization. Within a school-based sample, we test the cumulative and distinctive contributions of these stressors on youth mental health, accounting for demographics, as well as the contributive effects of three mental health resilience resources (family bonding, school engagement and sleep sufficiency) which serve as partial mediators of stressors. We hypothesize that greater levels of these adversities will be significantly associated with lower levels of health-promotive resources, that poverty and victimization will sustain inter-related but distinct contributions to mental health, and that these resilience resources will be positively associated with positive mental health, reducing the effects of poverty and victimization.

Method

Survey Methods and Sample Characteristics

We analyzed data collected from 8th, 10th, and 12th grade students via the Washington State 2010 Healthy Youth Survey (HYS) in this cross-sectional study. By using a clustered sampling design, the HYS randomly selects Washington State schools and invites all students in participating schools to complete the survey. Because the number of items in this survey is too large to be asked of any one student, the HYS is segmented into four subsets, of which we accessed Forms B and NS (n = 11,222) that contained the variables of interest. In 2010 HYS, the school participation rates ranged from 81%-90% for 8th, 10th, and 12th grades, respectively. Students are notified that the survey will secure the anonymity of their answers. Parents of the children were also informed that they could refuse the participation of their children.

Average age of the sample is 14.93 years (SD=1.72) with approximately 27% of 13 years or younger and 28% of 17 year or older. The sample consists of 51.38% of female. The racial composition of the sample is as follows: 54.68% White or Caucasian, 13.38% Hispanic or Latino/Latina, 8.57% Asian or Asian American, 4.47% Black or African American, 2.25% American Indian or Alaskan Native, 2.09% Native Hawaiian or other Pacific Islander, 8.64% two or more races, and 5.93% other.

Measures

Mental health outcomes were assessed in three forms: Depression is a dichotomous variable capturing whether, during the past 12 months, a respondent ever felt “so sad or hopeless almost every day for two weeks or more in a row that you stopped doing some usual activities”. Suicidality is a cumulative index (range 0-3) based on students responding yes to one or more of the following: during the past 12 months, if a respondent experienced suicide ideation, identified a suicide plan, or made a suicide attempt. Psychological Well-Being is based on 5 items (looking forward to the future, feel good about myself, feel satisfied with the way my life is, feel alone— reverse coded, my life is much worse/better than that of others my age) from the Youth Quality of Life Instrument-Surveillance Version (YQOL-S). Whereas the original YQOL-S consists of 13 items conceptualizing broader quality of life indicators (Topolski, Patrick, Edwards, Huebner, Connell, & Mount, 2001), 2010 HYS uses only a partial set. We transformed the original items (measured on 0-10, not at all true to completely true) to 100-point scale items and obtained the mean of the transformed items, based on the recommended protocol by the Washington State Department of Health (Washington State Department of Health, n.d.)

Childhood Adversity was assessed in two forms: Victimization Index is a cumulative sum of four subscales (range of each subscale: 0-2). These included adult to child physical violence exposure (direct physical violence or witnessed physical violence), dating aggression (physical aggression or psychological aggression), feeling unsafe (currently feeling unsafe at school or being absent due to an unsafe school environment)—all of which were assessed as 0=neither form, 1=1 form, 2=both forms of exposure within that subscale—and recent bullied experience (0=no bullying, l=one type of bullied experience, 2=2 or more types of bullied experience). The total range of this index is 0 to 8.

Poverty is a cumulative index (range: 0-2) of two dichotomized low income indicators: food insecurity (had to cut meal sizes or skip meals in past 12 months due to lack of money: 0=no, 1= yes) and current housing insecurity (0= living in parent’s or guardian’s home, 1 = any one of the following less stable sheltering: motel/hotel; shelter; car, park, campground or other public place; with others or on own because cannot afford housing, waiting to be placed in foster care). Household income or other income indicators were not available.

Resilience resources were assessed in three forms. Family bondedness is assessed by the mean of two items: typically having dinner with family (1-5, never to always) and getting along with one’s parents (0-10, not at all to completely true). Due to the differing metrics, we standardized the items and took their mean (range=−2.36 to 1.05). School engagement is the mean index of three variables related to the involvement in school life including academic grade average, school enjoyment (0-4, never to almost always enjoyed being in school), and afterschool activity participation (0, 1-2, 3 or more days a week). Because these items were also measured with different metrics, we standardized the items and took the mean (range=−1.97 to 1.19). Grades were measured 1-5, mostly F’s to mostly A’s. Sleep sufficiency is a mean of the average number of sleep hours on weekday and weekend nights (range=5 to 9 hours).

Statistical analysis

We analyzed the prevalence of victimization experience across sample characteristics in the first stage by comparing means between subgroups in each covariate used for the following regression models. After conducting bivariate correlation analyses study variables, we then undertook theory-guided hierarchical multiple regressions sequentially regressing the three measures of mental health, controlling for demographics, onto the two blocks of explanatory variables: adverse childhood experiences (victimization and poverty) followed by resilience resources (family bondedness, school engagement, and sleep sufficiency). In many situations, linear regression models provide results that are indistinguishable or even preferable to those of logistic regressions with dichotomous dependent variables, and are more straightforward to interpret (Hellevik, 2007). Long (1997) and Von Hippel (2015) argue that if the predicted probabilities one is modeling are extreme (close to either 0 or to 1) then use of logistic regression is favored. When probabilities are more moderate (between .20 and .80 or so), then linear models generally fit as well as logistic models and linear models are recommended for ease of interpretation. Because predicted probabilities in the regression models with depression were between .2-. 8, we report the linear regression outcomes.

Thus, a full model was tested for each dependent variable by suquentially adding three variable blocks. This procedure tests cumulative effects of the full set of predictors in each model as well as the explanatory utility of each variable set and the individual predictors within the sets, controlling for the shared variance among all model variables. In order to examine the explanatory utility of the victimization index relative to its four subscales (physical violence exposure from adults, bullied by peer, dating aggression, and unsafe at school), we ran another set of regression models substituting the victimization index with those four victimization subscales. All analyses were conducted with STATA version 14.1.

Results

xVictimization and Poverty Prevalence Profiles

Significant differences in prevalence of victimization were evident across study covariates as shown in the Table 1. Adolescents experienced higher mean levels of victimization and poverty if they were female and of certain racial/ethnic minorities (Black, Native American, Pacific Islander and mixed race). As hypothesized, greater adversity was significantly associated with all resilience resources-- getting less sleep, less engaged in school; and with weaker family bonding. Particularly, youth with poverty related problems of both food and housing insecurity were exposed to more than three times greater victimization than youth without those adversities. Comparisons across levels of sleep sufficiency, family bonding, and school engagement also revealed that youth with the lowest level of resilience resources experienced approximately double the level of victimization exposure. Youth across age groups did not evidence clear trends of increased/decreased overall victimization exposure as age increases.

Table 1.

Victimization and Poverty Scores Across Study Variables

	Variable categories	Victimization (max score = 8)		Poverty (max score = 2)
		Mean (SD)	Ftest	Mean(SD)	Ftest
Age	≤13	1.45(1.55)	F(3,7887)=1.69	.20 (.45)	F(3, 8007)=17.11***
	14-15	1.48(1.58)		.24 (.49)
	16-17	1.39(1.57)		.24 (.49)
	≥18	1.37(1.61)		.35 (.60)
Gender	Female	1.49(1.57)	F(l, 7889)=12.56***	.22 (.47)	F(l, 8011)=6.38*
	Male	1.37(1.57)		.25 (.51)
Race/Ethnicity	White	1.29(1.48)	F(7, 7834)=17.28***	.19 (.44)	F(7, 7948)=20.68***
	Asian	1.39(1.60)		.16 (.41)
	Pacific Islander	1.57(1.70)		.32 (.55)
	Hispanic	1.61(1.65)		.32 (.56)
	Mixed race	1.66(1.66)		.28 (.51)
	Nat, American	1.66(1.57)		.33 (.56)
	Other	1.77(1.63)		.31 (.55)
	Black	1.98(1.88)		.41 (.65)
Poverty[food, housinginsecurity]	0	1.16(1.34)	F(2, 7817)=707.86***	-	-
	1 insecurity	2.26(1.75)
	2 both forms	4.03 (1.97)
Sleep Sufficiency	< 6 hours	2.12(1.85)	F(3, 7778)=134.91***	.42 (.62)	F(3, 7894)=103.39***
	6-7 hours	1.59(1.60)		.27 (.52)
	7-8 hours	1.25(1.43)		.17 (.43)
	≥8 hours	1.07(1.34)		.14 (.39)
School Engagement	Low	1.91(1.76)	F(2, 7777)=214.96***	.37 (.60)	F(2, 7902)=191.61***
	Moderate	1.34(1.48)		.21 (.46)
	High	1.04(1.30)		.12 (.34)
Family Bondedness	Low	2.08(1.78)	F(2, 7838)=415.87***	.41 (.62)	F(2, 7956)=341.69***
	Moderate	1.25(1.39)		.17 (.41)
	High	0.94(1.24)		.11 (.32)

Note. Subgroups for family bondedness and school engagement were operationalized as trichotomies of the variable distributions.

^*p≤.05

^**p≤.01

^***p≤.001

Hierarchical Multiple Regressions

As Table 2 shows, all three full regression models for 1) depression, 2) suicidality, and 3) psychological well-being achieved significance. The magnitude of explained variance was significantly augmented by addition of the explanatory variable blocks, as reflected by increases in R². All the explanatory variables except poverty in the full model of depression remained significantly associated, demonstrating unique contribution after controlling for the effects of shared variance among the predictors.

Table 2.

Hierarchical Regressions Testing Incremental, Unique, and Cumulative Contributions of Study Variables to Mental Health

	Depression [β ], n=7525			Suicidality [β], n=7521			Psychological well-being [β], n=7506
	Stepl	Step2	Step3	Stepl	Step2	Step3	Stepl	Step2	Step3
F	20.97	127.12	134.72	5.56	116.96	111.14	9.72	215.37	429.18
R2	.02	.16	.20	.01	.15	.17	.01	.24	.45
Age	.05***	.06***	.01	.01	.01	−.02*	−.05***	−.05***	.04***
Female (Reference)
Male	−.13***	−.11 ***	−.12***	−.06***	−.05***	−.05***	.08***	.06***	.06***
White (Reference)
African	.04**	.00	−.01	.04**	.00	−.00	−.03**	.01	.03***
Hispanic	.06***	.03**	.01	.02	−.01	−.02	−.03**	.01	.04***
Asian	−.00	−.01	−.01	.02	.02	.02	−.03*	−.02*	−.02**
Native	.03**	.02	.01	.02*	.01	.00	−.03**	−.01	.00
Pacific Islander	−.00	−.01	−.01	.01	−.00	−.00	−.02	−.00	−.01
Other	.03*	.00	−.00	.03*	.00	.00	−.03**	−.00	.01
Multi-race	.05***	.03*	.02*	.03*	.01	.01	−.02	.01	.01
Victimization		.34***	.27***		.34***	.28***		−.36***	−.21***
Poverty		.06***	.01		.09***	.05***		−.21***	−.10***
Family Bonding			−.12***			−.12***			.38***
School Engagement			−.12***			−.07***			.17***
Sleep Sufficiency			−.09***			−.05***			.12***

Note. β=standardized coefficient. All model statistics and R²Δ values are significant at p ≤ .001.

^*p<0.05

^**p<0.01

^***p<0.001

Consistent trends were evident relative to sex wherein female status accounted for greater mental health problems. Racial minority status was less clearly patterned although unfavorable for students of color relative to White youth indicating association with poorer mental health in varying degrees and with some change in the valence of the betas in the full models. In the full models of suicidality and psychological well-being (controlling for all predictors), age positively influenced better mental health status. Although older age significantly contributed to depression initially, significance was lost in the final model. All of which suggests that resilience resources were associated with less negative trends of age to mental health.

Greater victimization sustained the strongest unique association with each mental health indictor, net of the effects of other predictors. As expected, the victimization coefficient was reduced in each model with the addition of resilience resources. A similar trend was observed for the association of poverty to mental health across steps in the models, particularly for depression. Finally, each of three resilience resources functioned as significant predictors of lower depression and suicidality and better psychological well-being when all the other variables were accounted for, with family bondedness demonstrating the strongest coefficients.

Because the measures do not all provide clarity as to timing sequence of events, we are constrained in the extent to which resilience factors can be interpreted to be mediating (e.g., lessening) the relationship of victimization and poverty to mental health outcomes. However, because there is likely sufficient history in the lives of these factors for these youth to assume some degree of mediation, we undertook Sobel tests to determine if the observed reductions in the beta strengths of victimization and poverty when each resilience resource was added achieved significance, suggesting a statistically significant partial mediation. All Sobel tests achieved significance at the p>.001 level or better. In corpus, these results confirm study hypotheses that poverty and victimization would sustain inter-related but distinct contributions to mental health and that resilience resources are associated with more positive mental health as well as with weaker effects of poverty and victimization on mental health.

Component Analyses

Addressing our final research question, we tested component models that substituted the four subscales (adult physical violence, bullied by peer, dating aggression, and unsafe at school) for the victimization index (Table 3). Coefficients of all the other variables in these component models were very close to those of the original models (Table 2), so are not presented here. Each victimization subscale sustained independent significance indicating that each contributes value added explanation of the outcomes. As hypothesized, the cumulative victimization index demonstrated more robust effect sizes than those of each subscale for each mental health outcome. In addition, similar to the cumulative index, the effect of each victimization subscale diminished after resilience resources were explained.

Table 3.

Multivariate Regressions Substituting Victimization Components for the Aggregate Victimization Index

	Depression [β], n=7347			Suicidality [β], n=7344			Psychological well-being [β], n=7328
	Stepl	Step2	Step3	Stepl	Step2	Step3	Stepl	Step2	Step3
F	19.39	99.14	109.75	5.15	91.51	90.21	9.47	168.77	352.22
R2	.02	.16	.20	.01	.15	.17	.01	.24	.45
Adult Violence		.17***	.12***		.15***	.11***		−18***	−.07***
Bullying		.18***	.17***		.14***	.13***		−.18***	−.14***
Felt unsafe		.11***	.07***		.13***	.10***		−.15***	−.07***
Dating Violence		.03**	.03*		.10***	.10***		−.04***	−.02*
Poverty		.07***	.02		.08***	.04***		−.21***	−.10***

Note, /βtandardized coefficient. All model statistics and R²Δ values are significant p ≤ .001. Variables tested in these regression models are identical to those in table 2 except for interpersonal adversity subscales for the cumulative measure; coefficients for other variables are nearly equivalent to the table 2 so are not shown.

^*p<0.05

^**p<0.01

^***p<0.00l.

Discussion

This study considered youth victimization and low income as stress-evoking adversities that may combine to increase the risk of poor psychological health. The findings add to the evolving understanding of early life adversity effects on adolescent mental health by examining the cumulative and distinct contributions of youth victimization and family poverty, accounting for demographics as well as three constructs that tap theorized mental health resilience resources. Measured as a cumulative index, victimization was nearly exclusively the single strongest predictor—a particularly compelling finding within a general population as opposed to clinical or system-involved youth samples. Findings also support the value of mutable resilience resources in fostering positive mental health both at the bivariate level and within multifactor analysis such as these.

Prevalence of Victimization across Youth Characteristics

A cumulative adversity perspective urges attention to ways that variation in strains and resources may be socially patterned. Our findings are consistent with this premise, indicating childhood adversity assessed through youth victimization as part of this patterning. Higher levels of youth victimization were significantly evident, for example, among females, racial and ethnic minority youth, those with current socioeconomic disadvantage, those with greater sleep impairment, and those with lesser family and school supports—predictors that themselves individually and cumulatively are associated with undermined mental health. These results are supported by Turner, Finkelhor & Ormrod’s (2006; see also Umberson, Williams, Thomas, & Thomeer, 2014) findings that early exposures to stress were aligned with social status. Turner et al (2006) found that racial and ethnic minorities, children from lower socioeconomic status, and those living with a single parent or step-parent experienced more types of victimization as well as non-victimization adversity than children from higher social statuses, particularly among children 10-17.

Childhood adversities—both victimization and nonvictimization—tend to co-occur, arising during developmentally sensitive periods in the life course with the risk of catalyzing chains or linked pathways of subsequent adversity exposure (Ben-Shlomo & Kuh, 2002). This patterning mirrors stress proliferation processes wherein primary stressors lead to conditions or events that add to and scaffold the continuing impact of earlier life adversities, fostering socioeconomic, behavioral, and psychosocial pathways alongside biological mechanisms of stress reactivity (Aneshensel, 2009). Victimization and other childhood stressors appear to lay a foundation of vulnerability, with demonstrated association with poorer mental health in early adulthood (Schilling, Aseltine, & Gore, 2007) as well as prediction of greater adversity exposures, poverty and limited social supports in adulthood, accumulating to produce eroded adult mental health (Jones, Nurius, Song, & Fleming, 2018).

Adversity Contributors to Youth Mental Health

The inclusion of three related yet distinct features of adolescent mental health permits examination of the stability of findings and their interpretation across models. Findings indicate that both adversities – victimization and poverty – as well as resilience resources sustain independent contribution to all indicators of mental health, controlling for associations with all other study variables. Victimization largely conveyed greatest relative effect. Sex also sustained as a significant contributor, indicating a heightened mental health risk for female adolescents. Race and ethnicity effects apparent in the first step were substantially, although not completely, lower after poverty and victimization were entered into the models. Although weak, effects for African-American and Hispanic statuses were positive for psychological well-being once resilience resources were accounted for. These findings may suggest that the residual variance conveys cultural characteristics that hold protective potential for these youth, consistent with Arlington & Wilson (2000) and Ungar (2015). Similarly, in the final model when resilience resources were accounted for, older age among these teens lost its unfavorable associations with mental health, underscoring the value of multiple supports in later adolescence.

Examination of early life adversity effects on mental health among adults has demonstrated patterns of greater adversity exposure in adulthood as well as lesser adult supports and maintenance of healthy habits – yielding chains of disadvantage that undermine physical and mental health (Nurius et al., 2015; Turner, Thomas, & Brown, 2016; Umberson et al, 2014). Path analysis of early life adversities have demonstrated both direct and mediated pathways to adult mental health through lower adult socioeconomic status as well as lesser social support and greater adult adversities including victimization (Jones et al., 2018). The current findings in adolescence appear to be reflecting a developmental foundation setting for these later observed trends.

Victimization here is more prevalent across social determinant factors such as minority race/ethnicity status and poverty and with lesser resilience resources, and proves to be, in general, the comparatively most robust predictor of adolescent mental health risks. Those with higher victimization scores here are generally experiencing multiple forms of victimization, suggesting a broader base of maltreatment exposure across the people and contexts of youths’ lives. Three of the four subscales reflect peer victimization forms, accentuating the importance of school-based assessment and interventions both to prevent exposures and to build resilience fostering supports.

Although different forms of victimization frequently co-occur (Finkelhor, Ormrod, Turner, & Hamby, 2005), the current findings indicate that each form assessed here carries unique contributions toward increased risk of mental health problems including co-presence of other victimization forms. Findings that each victimization form adds significantly to three indicators of mental health problems and that effects for the cumulative index is more robust than any one victimization form reinforces theorizing that these type of stressors have cumulative, compounding effects. Because the majority of items assessed here pertain to victimization by peers and peer contexts (dating violence, being bullied, feeling unsafe at school), schools have particularly vital opportunities to engage with youth in terms of both victimization and perpetration of peer victimization. Both youth victimization and undermined mental health carry broad spectrum risks for other aspects of function such as educational attainment (Fletcher, 2008; Pate, Maras, Whitney, & Bradshaw, 2017).

Promising Possibilities for Resilience Factors

Our findings support the value of “poly-strengths”—diversity of one’s overall portfolio of strengths (Hamby, Grych, & Banyard, 2018) --for reducing the mental health threats carried by victimization and poverty given that effects for both were reduced with addition of resilience resources in the forms here of school engagement, family supports, and sleep sufficiency. Yet, young people who have experienced greater childhood adversity report feeling more stressed and less supported (Karatekin & Ahluwalia, 2016), all of which is associated with worse mental health outcomes. The family and school predictors here reflect aspects of respondents’ perceived sense of support and positive engagement, which proved to be valuable for youth mental health, including within the context of adversity exposures. More broadly, social support has been found useful in the contexts of several forms of violence exposure as well as toward promoting safety perceptions (Benhorin & McMahon, 2008; Rosario, Salzinger, Feldman, & Ng-Mak, 2008).

Yet lesser family support and school engagement signal greater isolation and lack of coping supports. These may also signal stressors such as harsh parenting or stressed caregivers at home or a poor sense of belonging or safety at school, which compound adversity effects (Bright, Knapp, Hinojosa, Alford, & Bonner 2016). Research has indicated that parents of minor children who had both higher ACEs in their childhood as well as higher similar adversities in adulthood (e.g., physical or sexual victimization) experienced markedly poorer household socioeconomic resources, greater behavioral risk habits and poorer physical and mental health— all of which become part of their children’s adversity exposure (Borja, Nurius, Song, & Lengua, 2019), suggesting intergenerational links. Beyond individual and family level costs, adversities such as victimization carry large economic and societal costs (Fang, Brown, Florence, & Mercy, 2012).

Sleep disturbance also often signals stressful contexts and is associated with being socially withdrawn (Chauliac, 2017) and with increased rates of mental health disturbances such as depression (McGlinchey, Reyes-Portillo, Turner, & Mufson, 2017; Orchard, Pass, Marshall, & Reynolds, 2017) and suicidal ideation (Im, Oh & Suk; 2017). Sleep insufficiency has been shown to impair emotional regulation (Walker & Van Der Helm, 2009), impulse control (Abe, Hagihara & Nobutomo, 2010), and problem-solving ability (Sio, Monaghan & Ormerod, 2013), which have detrimental implications for stress coping and success outcomes like academic performance. Sleep disturbances have been shown to seriously affect learning capacity, school performance and neurobehavioral functioning (Curcio, Ferrara & De Gennaro, 2006; Fallone, Owens & Deane 2002; Wolfson & Carskadon, 2003). Conversely, adequate sleep is an essential resource for both healthy development and stress amelioration that helps to offset disturbance generated by neurobiological alternations and hyperarousal (Koskenvuo, Hublin, Partinen, Paunio, & Koskenvuo, 2010), and foster psychological well-being and resilience to life stressors. These findings suggest support for interventions such as later school start times toward increasing sleep duration, which is associated with improved attendance and decreased incidences of drowsiness during school, decreases depressive symptoms, and some evidence of stronger academic performance (Wheaton, Chapman & Croft, 2016).

Implications and Opportunities for School Based Responding

The greatest opportunity for mitigating the effects of significant adversities occurs during childhood and adolescence (Walker & Walsh, 2015). Schools are an important point of contact for prevention, identification and treatment of mental health concerns and disorders, and ensuing academic problems (Bruns, Walrath, Glass-Siegel, Weist, 2004). Schools present a unique opportunity to promote resilience resources; e.g., through building supportive relationships. These include bonding with positive role models, setting high expectations that have both clear and consistent boundaries, opportunities for student engagement and autonomy, teaching social and emotional skills and working collaboratively with families (Roffey, 2016). Furthermore, schools have increasingly become the focus for health interventions and services because of their availability and accessibility to vulnerable student populations (Denny et al., 2011). School-based health centers with a focus on mental health services as a structural intervention may have the potential to address inequalities in pediatric mental health and address disparities in school achievement that perpetuate income inequality disparities and increased exposure to trauma (Larson, Chapman, Spetz, & Brindis, 2017).

The need for a more comprehensive approach to addressing trauma within the school setting has been widely documented (Chafouleas, Johnson & Santos, 2016; Jaycox, Katonka, Stein, Langley, & Wong, 2012; Overstreet & Chafouleas, 2016). These school-based mental health services provide the opportune placement for trauma informed practices toward preventing, identifying and ameliorating the effects of childhood adversities. Trauma informed practices aim to not only address the adversity, but also seek to address the biological, cognitive and behavioral adaptations that result from cumulative exposures. Programs aimed at mitigating the effects of trauma on development focus on fostering quality relationships, age-appropriate regulation of emotions and progressive mastery of developmental skills, as well as on positive school climate and classroom management (Blodgett & Dorado, 2016). Further, Overstreet and Mathews (2011), argue that utilizing a public health framework to address trauma in schools will promote prevention and early identification through addressing the underlying causal processes that lead to social, emotional and cognitive maladjustment.

Schools provide the opportunity to identify children and families at risk and provide them with referral services, including referral and supportive services to address trauma (Oral et al., 2016). Training school personnel includes tools for identifying behaviors and coping strategies resulting from toxic stress (Walkley & Cox, 2013). Critical also is use of individualized interaction or instruction when needed that manage the social and physical environment to support positive psychosocial and learning outcomes (Blodgett & Dorado, 2016).

In addition to mental health, maltreatment in childhood is associated with a number of adverse academic factors including decreased cognitive function (Veltman & Browne, 2001), lower grades, lower academic achievement and an increase in the use of special educational services (Ryan et al., 2018) as well as disciplinary problems (Ireland, Smith & Thornberry, 2002). Students with victimization experiences struggle with self-regulation, social skills and truancy in school, which places them at greater risk for punitive and exclusionary action (Fantuzzo, Perlman, & Dobbins, 2011; Veltman & Browne 2001). Youth who experience childhood adversity including victimization, like many of our respondents, may well come to schools’ attention through teachers who are encountering these struggles.

Educators today face incredibly challenging, under-resourced environments. Recent research has focused on framing school-based trauma informed services within a multi-tiered model of support that aligns with current educational practices to promote resilience (Reinbergs & Fefer, 2018). Two widely recognized and implemented trauma informed interventions in the school setting include Social Emotional Learning (SEL) and Positive Behavior Interventions and Supports (PBIS). Social emotional learning provides a coordinated and systematic approach to helping children understand and regulate their emotions as well as social problem solving to increase student success and well-being (Durlak et al., 2011; Payton et al., 2008).

Positive Behavior Interventions and Supports (PBIS) is a set of intervention practices and systems for facilitating positive and safe social cultures and behavioral supports necessary to achieve academic and social success at the whole school level (Horner, Sugai & Anderson, 2010). It promotes social behaviors through developing and teaching consistent expectations, while increasing positive teacher and student interactions (Sugai & Horner, 2009). These programs provide instructional opportunities aimed at addressing social, emotional and behavioral expectations that have been shown to promote resiliency, consistency, adaptive coping strategies, student connectedness, positive behaviors and well-being for students, including those who have been exposed to trauma (Reinbergs & Fefer, 2018).

Use of Surveillance Data and Limitations

The current study draws on population-based state surveillance data of school based youth. Data from surveys such as this offer a practical strategy toward developing population-level examination of early life adversity association with health attending to both adversity exposure and resilience resources. This approach yields insights about general population youth that are value-added etiology and potential intervention targets. Findings of youth in specialized system or clinical settings are valuable in their own right, but tend to capture information only about individuals with formally reported victimization and are often restrictive in respondent characteristics and etiological factors. Community-based surveillance findings have relevance for prevention and resilience-building supports within the school settings in which youth are assessed as well as other youth-serving providers such as primary care professionals, family and mental health services, and other child and adolescent support services.

These benefits reside alongside some study limitations. Current results are based on self-reported, cross-sectional data. Although recall or reporting bias is always a consideration, the relatively short recall period may reduce the risk of bias. Substantial testing of retrospective adversity accounts has demonstrated stable linear trends, minimal association with participants’ psychological state at the point of assessment, and recall bias largely in the direction of underreporting occurrences (Corso, Edwards, Fang, & Mercy, 2008; Yancura & Aldwin, 2009), including milder forms of childhood adversity associated with later mental health problems (Taylor et al., 2006) and comparison of retrospective and prospective results (Hardt, Vellaisamy, & Schoon, 2010). The ability to examine both victimization and poverty as cumulative risk factors alongside protective factors is a strength.

Another limitation relates to measurement availability within secondary analysis. Indicators available to represent the constructs of youth victimization, poverty, family and school engagement, and mental health reflect, for example, a circumscribed portion of those domains. Depression was assessed using a single item and although this is common in epidemiological research, merits caution. Use of multi-indicator indices should serve to increase the stability of the measures and trends in the results. However, caution is prudent regarding the extent of domain interpretation. Finally, population characteristics of Washington State will vary in some respects from those of other states. The representation of racial/ethnic minority youth is about 45% in this sample, though the specific racial/ethnic composition may vary from that of other states. More diffuse diversity is a growing reality for many U.S. cities and state, thus, racial/ethnic variations may or may not bear upon generalizability of the findings reported here.