Critiques on article

Impacts of Family and Community Violence Exposure on Child Coping and Mental Health

Esror Tamim Mohammad & Ester R. Shapiro & Laurel D. Wainwright & Alice S. Carter

Published online: 2 August 2014 # Springer Science+Business Media New York 2014

Abstract An ecological stress process model was employed to explore relations between children’s exposures to family and community violence and child mental health, and emotionally- regulated coping (ERC) as a protective factor among Latino, European–American, and African–American school-aged children (n=91; girls, n=50[54 %]) living in single-parent fami- lies who were either homeless and residing in emergency shelters or housed but living in poverty. Mothers reported domestic violence experiences and their child’s history of physical/sexual abuse, community violence exposures, and mental health. Children reported on exposure to community violence, internal- izing symptoms, and coping. The mental health impacts of multi-level violence exposures and ERC as a moderator of asso- ciations between violence exposures and child mental health was tested with structural equation modeling. Family abuse was uniquely associated with PTSD, and community violence with anxiety and aggression. Latent interaction tests revealed that ERC moderated relations between family abuse and anxiety, aggres- sion and PTSD. Emotionally-regulated coping appears to play a protective role for children’s mental health in contexts of violence exposure, offering opportunities for intervention and prevention.

Keywords Poverty . Violence exposure . Emotion regulation . Mental health . Ethnic minorities

Introduction

The prevalence and consequences of children’s exposure to violence (ETV), including experiencing or witnessing

physical violence and physical coercion across family and community contexts, has been an area of concern across research disciplines, services providers and policymakers (e.g., Finkelhor et al. 2009; Foster and Brooks-Gunn 2009). In a recent large-scale population study of a nationally representative sample of children and adolescents (n=4,549), 60.6 % reported at least one direct or witnessed victim- ization in the past year, 46.3 % experienced a physical assault, 25.3 % reported witnessing violence, and 9.8 % reported witnessing family violence (Finkelhor et al. 2009). Prevalence rates are even higher when examining the experiences of children and adolescents of African American, Latino and Native American descent, as so- cial-ecological factors that precipitate violence (e.g., so- cial inequality, racism and discrimination, poverty, and crime) are more commonplace in their environments (Brady et al. 2008; Hanson et al. 2006; Finkelhor et al. 2009; Foster and Brooks-Gunn 2009). Moreover, a large body of research shows that children’s ETV can be a significant risk factor for the development of emotional and behavioral problems, including internaliz- ing, externalizing, and post-traumatic stress symptoms (e.g., Chan and Yeung 2009; Cicchetti and Toth 2005; Kearney et al. 2010; Lynch 2003; Margolin and Gordis 2000; Yates et al. 2003). However, much of the litera- ture concentrates on the negative psychological and mental health impacts of ETV as an isolated or inde- pendent occurrence. Research on the role of child cop- ing as a protective factor in the presence of multiple violence exposures is scant; yet further research on child coping is needed to inform prevention and inter- vention strategies that minimize the negative mental health consequences of violence exposures. Thus, the present study applies an ecological stress process model of child ETV to examine (a) associations between fam- ily and community violence exposures and child

E. T. Mohammad (*): E. R. Shapiro: L. D. Wainwright: A. S. Carter Department of Psychology, University of Massachusetts Boston, 100 Morrissey Blvd., Boston, MA 02125-3393, USA e-mail: [email protected]

J Abnorm Child Psychol (2015) 43:203–215 DOI 10.1007/s10802-014-9889-2

mental health adjustment, and (b) the role of emotionally- regulated coping (ERC) as a protective factor in the associa- tion between ETV and child mental health symptoms.

Ecological Stress Process Model of Children’s Exposure to Violence

The stress process model of exposure to family and commu- nity violence (Foster and Brooks-Gunn 2009) offers an inte- grative framework for conceptualizing the effects of multiple types of violence exposure on children’s mental health adjust- ment and the role of child emotionally-regulated coping as a protective factor altering relationships between violence exposure and mental health outcomes. The model recognizes multiple sources of stress across ecological domains (e.g., racism, socioeconomic disadvantage, housing instability, family structure and adversity), mediators and moderators of stress (e.g., resources for social support and coping), as well as outcomes of stress (e.g., mental health symptoms), which are viewed as interconnected elements in understanding children’s ETV. Foster and Brooks-Gunn (2009) were particularly interested in addressing the co-occurrence of high stress exposure and low resource access for children and families of ethnic minority status living in poverty to guide targeted programs and policies addressing social structures and inequalities. In their model, coping resources applied across multiple levels of a child’s ecology are hypothesized to protect against negative child mental health outcomes in contexts where ETV was likely to occur in multiple settings. Children living in poverty have little control over contextual coping resources, exercising the most control over their own coping strategies. For example, children’s emotionally-regu- lated coping and children’s self-reported assessment of the effectiveness of their coping strategies in responding to both daily demands and more challenging stressful experiences or potentially traumatic events are within a child’s control, and may be associated with improved mental health outcomes (Eisenberg et al. 1997; Ryan-Wenger 1990; Sandler 2001). Consistent with Foster and Brooks-Gunn (2009), our model postulates that child coping may be an individual, cross-con- textual resource that could moderate, and thereby reduce, the adverse effects of multilevel violence exposure on child men- tal health symptoms.

Exposure to Violence (ETV) and Child Mental Health

Within an ecological stress process model, ETV can be con- ceptualized as a social/environmental stressor, which can be associated with mental health symptoms. Although multiple types of ETV are prevalent, most research investigating the consequences of ETV has concentrated on the detrimental

effects of specific types of violence (Foster and Brooks- Gunn 2009), which limits conclusions that can be drawn regarding the differential effects of interrelated and co-occur- ring violence exposures. For example, a preponderance of research has shown that exposure to family abuse (e.g., phys- ical and sexual violence) can have deleterious immediate and long-term influences on child mental health (for reviews, see Cicchetti and Toth 2005; Kearney et al. 2010). In addition, whether through direct victimization or witnessing, exposure to intimate partner violence and community violence are associated with significant risk for developing emotional and behavioral problems, including internalizing, externalizing, and post-traumatic stress symptoms (e.g., Chan and Yeung 2009; Yates et al. 2003; Lynch 2003; Margolin and Gordis 2000). The present study was designed to examine the differ- ential effects of multiple types of violence exposure that can emerge and co-occur across ecological settings.

The Role of Child Coping in the ETV-Mental Health Link

The ecological stress process model of ETV underscores the relationship between stressors and resources as contributors to mental health outcomes, enabling a consideration of coping resources, such as the child capacity to cope with the demands of violence-based stressors, which could moderate and alter ETV-mental health links. Consistent with this model, expo- sure to family abuse and intimate partner violence can alter developmental and mental health trajectories by influencing adaptational capacities, such as emotion self-regulation (e.g., Cicchetti and Toth 2005; Cummings et al. 2009). Children’s emotional reactivity and dysregulation have been identified as moderators of the relation between exposure to family violence and internalizing and exter- nalizing behavior problems (Egeland et al. 2002). In contrast, findings for the role of child coping in the community violence literature are mixed. While use of problem-focused coping in response to community vio- lence can increase risk for externalizing behavior prob- lems, the same coping strategies appear to be protective in the presence of community violence in relation to internalizing behavior problems and PTSD (Kliewer and Sullivan 2008; Kliewer et al. 2006). Coping conceptu- alizations applied in ETV research frequently lack theo- retical clarity regarding the availability, effectiveness and limitations of child coping efforts, including devel- opmental and contextual considerations when children confront stressful life events and circumstances (Eisenberg et al.1997; Foster and Brooks-Gunn 2009; Tolan et al. 2006). Greater understanding of the ways in which child resources for coping with violence-based stressors moderate the association between violence exposures and mental health is needed to identify protective factors,

204 J Abnorm Child Psychol (2015) 43:203–215

which can be strengthened and mobilized through prevention and intervention strategies.

Conceptualization of Child Emotionally Regulated Coping in the Present Study

The stress process model (Foster and Brooks-Gunn 2009) highlights the importance of conducting research examining the function and efficacy of child coping as a protective factor, as children exercise greater control over their own coping strategies than over their life contexts. The theoretical frame- works and conceptual models that guide adult coping research (see Aldwin et al. 2011; Folkman and Moskowitz 2004) have influenced the conceptualization and measurement of child coping (Compas et al. 2001; Eisenberg et al. 1997; Ryan-Wenger 1990). Child coping models, which follow the adult literature in distinguishing primary control (attempts to alter objective conditions) and secondary control (at- tempts to adjust oneself to objective conditions) forms of coping, vary with respect to how core dimensions of cop- ing responses, (e.g., problem and emotion-focused coping, avoidance, appraisal and approach coping, control coping, cognitive and behavioral coping, engagement and disen- gagement coping, and emotion, behavior and situation regulation coping),are differentiated (Compas et al. 2001). Compas et al. (2001) point out that primary control coping has been theorized to include both efforts aimed at affecting actual events or circumstances and regulating one’s emotions. By contrast, secondary coping was con- ceptualized as strategies enacted to enhance an individual’s fit with or adaptation to environmental demands (Compas et al. 2001), and can include the child’s emotion regulation as a coping strategy. Although empirical studies have ex- amined the role of different dimensions of child coping with a wide array of stressors (e.g., medical conditions, parental cancer, bereavement, and parental divorce) (Ayers et al. 1996; Grant and Compas 1995; Grant et al. 2000; Sandler et al. 2000), few studies have examined the role of child coping with violence-based stressors (Brady et al. 2008). In a recent review of the literature on children’s responses to terrorism, a form of violence exposure, Eisenberg and Silver (2011) noted that children exposed to other forms of violence prior to or subsequent to a terrorist attack, and ethnic minority children living in con- texts exposing them to other stressors evidenced greater vulnerability, with more compromised mental health out- comes. Eisenberg and Silver (2011) argue that coping and regulation are highly overlapping constructs, designed to calibrate how children respond both behaviorally and emo- tionally to challenges in general, and violence exposures in particular. Eisenberg et al. (2007) view children’s emotion- related regulation as a multi-faceted adaptive process

involving both how children experience emotions and how they express them behaviorally. Aldwin and col- leagues (2011) suggested that the coping literature has overemphasized individual primary control and problem- solving. They proposed that adaptive self-regulation may require both control and accommodation strategies, and that these may be particularly important for children and adults when facing conditions outside of their control, such as violence in their families or communities. The available research suggests that adaptive coping strategies, such as child emotional and behavioral self-regulation in the pres- ence of family and community violence exposures, may buffer the extent to which ETV is associated with mental health symptoms.

An additional dimension of child coping presented in the literature is the child’s self-reported assessment of the effec- tiveness of their coping strategies (Compas et al. 2001; Ryan- Wenger 1990). Ryan-Wenger (1990), whose work integrated Lazarus and Folkman’s (1984) theoretical framework for con- ceptualizing stress-coping processes and outcomes with a developmental perspective, proposed to assess children’s self-reported perceptions of the adequacy of their coping efforts alongside the strategies themselves. Similarly, Sandler’s (2001) conceptualization of ecologies of adversity as presenting children with unique, stressor-specific chal- lenges interfering with children’s abilities to meet primary needs and goals emphasizes the importance of children’s perceptions of the effectiveness of their coping as a critical child developmental process disrupted by adverse experi- ences. Accordingly, recent researchers have found that chil- dren who use coping strategies aimed at controlling or chang- ing environmental stressors that are realistically outside of their control, such as interpersonal violence, can be at in- creased risk for psychopathology (Clarke 2006; Fosco et al. 2007). Thus, the evidence highlights the importance of iden- tifying the function and child self-reported, subjective experi- ence of the efficacy of specific child coping behaviors in the context of environmental stressors that can be beyond the child’s control (Clarke 2006; Fosco et al. 2007; Sandler 2001). Consistent with the ecological stress process model (Foster and Brooks-Gunn 2009; Pearlin 1999), the child’s sense of control or efficacy in coping is critical under condi- tions where stressors are beyond their control and therefore may moderate the relation between violence exposures and mental health outcomes.

The present study adopts a child coping model that empha- sizes child self-regulation capacities and processes (Eisenberg et al. 1997) and the following three broad domains of child coping: a) emotion regulation, b) situation regulation, and c) the regulation of emotionally driven behaviors . Children’s emotion regulation behaviors are understood to include strat- egies that diminish arousal and frustration, such as attention shifting and physical inhibition of behaviors to reduce

J Abnorm Child Psychol (2015) 43:203–215 205

escalation of emotional reactions (Eisenberg et al. 1997). When children’s emotion regulation attempts are not sufficient at reducing arousal in a stressful context, the regulation of emotionally-driven behaviors becomes relevant (e.g. behav- ioral regulation; Eisenberg et al. 1997). The child’s attempts to regulate the stressful context (i.e. situation regulation) and ability to regulate emotionally driven behaviors are employed after the child has appraised the stressful context, and has had emotional reactions that drive their behaviors (Eisenberg et al.1997). Behavioral regulation includes inhibiting and ex- pressing overt behaviors, which are relevant when emotion regulation methods have not reduced the effects of the envi- ronmental stressor (Eisenberg et al. 1997). The present study focuses on child coping in relation to the function and child self-reported sense of effectiveness of coping as a specific domain of self-regulation that may have protective effects in altering the process through which ETV is associated with mental health symptoms (Foster and Brooks-Gunn 2009).

Specifically, we proposed the following hypotheses:

1. Exposure to family and community violence will be as- sociated with higher levels of anxiety, aggression, and PTSD symptoms.

2. Emotionally regulated coping will moderate the relation between exposure to family and community violence and symptoms of anxiety, aggression and PTSD; whereby the relation between violence exposure and mental health symptoms will be attenuated among children endorsing more effective emotionally regulated coping.

Method

Participants

The data were provided by the Murray Center for the Study of Lives (now at the Institute for Quantitative Social Science, Harvard University) derived from the initial wave of the Worcester Family Research Project (WFRP, Bassuk et al. 1996), a longitudinal study of both currently homeless fami- lies residing in emergency shelters and low-income housed (never homeless) families conducted between 1992 and 1995 in Worcester Massachusetts (see Bassuk et al. 1996 for details of the original study; the Murray Center reviews the ethical conduct of all archived, de-identified data and approved the present study as consistent with the original study’s consent). From the total sample of 436 homeless and housed single- parenting families who participated in and completed multiple interview sessions during baseline data collection, there were 155 mothers with school-age children (n=236). Ninety-one mothers (age, M=34.47, SD=5.91) and their school-age chil- dren were retained for the present study because the children

met age criteria, (8 to 17-years-old; M=12.32, SD =2.45), that allowed for them to complete critical measures in this study. In cases in which mothers had multiple children (n=35), the oldest child was selected for inclusion to control for sibling effects.

Mothers in the study were predominantly young (age range=23-54, M=34.69, SD=6.24), single (94.5 %), and the primary caregiver of two or more children (95.6 % multiple child families). More than 62 % of the sample was housed (n=57), with 37.4 % of the sample consisting of homeless families residing in emergency shelters (n=34). Mothers tended to have less than 12 years of formal education (M=10.68, SD=2.63), and reported an annual family income less than$15,000 (M=$11,676, SD=$4,780), below the in- come threshold used to determine poverty status in the United States. Girls comprised 54.9 % of the sample (n=50). More than 50 % of the sample of children and adolescents was Puerto Rican or Other Latino (n=45), 25.8 % was European American (n=23), 10.1 % was African American (n=9), and 13.5 % other (n=12).

Procedures

Written informed consent was obtained from both mothers and children before interviews were conducted. Families par- ticipated in face-to-face interviews with psychiatric nurses, social workers, or psychology graduate students with clinical experience. All interviews were conducted over the course of three to four sessions totaling 10 h. Care was taken to provide private spaces for individual interviews for both mothers and children, in both homeless shelters and home settings. Homeless participants were interviewed in a private space within the homeless shelter in which they were residing. Housed participants were either interviewed in their homes or a community-based project office. As an incentive to par- ticipate at each interview, mothers received $10 vouchers and children received $5 vouchers redeemable for merchandise at local stores. Depending upon the language preference of the participant, interviews were conducted in English or Spanish.

Measures

Sociodemographic Information A modified version of the Personal History Form (Barrow et al. 1985; as cited in Bassuk et al. 1996) was administered to obtain information about housing, income, education, and family structure.

Exposure to Family Abuse Maternal responses to the fol- lowing two questions about the child’s history of abuse within the family were selected from the Background on Individual Child questionnaire (Bassuk et al. 1996): 1) “Do you think this child has ever been abused physical- ly?” and 2) “Do you think this child has ever been

206 J Abnorm Child Psychol (2015) 43:203–215

abused sexually?” Maternal responses to the follow-up question, “Who was the abuser?,” for each item were dummy coded according to whether the perpetrator of the abuse was a family member (e.g., mother, father, foster parent, grandparent) or not (e.g., stranger, neigh- bor, unknown). The presence of physical or sexual child abuse by a family member in either of these two items was summed to create a child Exposure to Family Abuse score (α=.63).

Exposure to Intimate Partner Violence The Conflict Tactic Scale (CTS; Straus 1979) Domestic Violence Subscale, com- prised of eleven items, was used to assess violence perpetrated by an intimate partner towards the mother herself. The major- ity of homeless mothers in this study did not complete the CTS measures about recent experiences of intimate partner violence due to concern that answers about recent or current experiences could result in a mandated report to child welfare services. For this reason, only CTS responses reporting on experiences of intimate partner violence since adulthood (e.g. since maternal age of 17 or older) were included in this study. The CTS intimate partner violence subscale had excellent internal consistency (α=.94).

Exposure to Community Violence To measure children’s exposure to community violence, scales were constructed using both child and maternal responses on the Life Events Questionnaire (LEQ; Masten et al.1994) and child re- sponses on the Strains and Worries Scale (Bassuk et al. 1996). The LEQ is a 45-item instrument that presents stressful life events that may have occurred in the past 12-months. Previous studies within this sample suggested that the children underreported violent community events when compared to maternal reports (Buckner et al. 2004; Buckner et al. 2003). Therefore, both maternal responses and child self-reports on six specific LEQ items reflecting child exposure to violence in the community were aggre- gated to create an ETV score (e.g., “During the past year, a member of the family was a victim of violence,” “witnessed something violent happen at school or in the neighborhood”).

The Strains and Worries Scale (SWS) is a 22-item mea- sure of life strains and worries (Bassuk et al. 1996; also, for a more detailed description of the scale, see Buckner et al. 2003). The SWS was administered to children to assess chronic adversities commonly experienced by poor and homeless children. When children endorsed a particular strain (e.g., Do you ever feel the place you are living is unsafe?), they were asked “how frequently?” (ranging from 0=never to 4=everyday), and “how stressful is this?” (rang- ing from 0=not stressful to 4=very stressful). Frequency scores for four items addressing how often children experi- enced strains and worries associated with violence in the

community were aggregated to create a community violence exposure score (α=.67).

Anxiety The Revised Children’s Manifest Anxiety Scale (RCMAS) is a widely used 37 item self-report instrument designed to assess the level and nature of anxiety in children and adolescents from 6 to 19-years-old. The Total Anxiety score is comprised of 28 items. The internal-consistency of the RCMAS in this sample (α=.88) is consistent with earlier reports (Pina et al. 2009; Reynolds and Richmond 1985).

Symptoms of Aggression and Post-Traumatic Stress The Child Behavior Checklist (CBCL, Achenbach 1991) is a widely used measure with strong psychometric quali- ties, including criterion-related validity and reliability. All mothers in the study were administered the CBCL. This study utilized the 20-item aggression subscale of the CBCL (α=.93) and 20- item Post-traumatic Stress Disorder (PTSD) subscale (α=.87; Achenbach and Rescorla 2007; Wolfe et al. 1989). Internal consistency in the current sample is consistent with earlier reports (Wolfe et al. 1989). Due to the high correlation between CBCL aggression and PTSD subscales, r=.72, p<.01, three overlapping items were excluded from the PTSD subscale. The revised 17-item PTSD maintained good internal consistency (α=.84) and its correlation with the 20-item Aggression subscale was in the moderate range, r=.57, p<.01.

Emotionally Regulated Coping The 26-item Schoolager’s Coping Strategies Inventory (SCSI; Ryan-Wenger 1990), a child-report instrument, was used to measure the frequency and effectiveness of coping strategies in relation to stress. For each of the twenty-six coping strategies listed in the SCSI, children were prompted indicate how often they use each strategy (frequency) and how helpful they perceived each strategy (effectiveness) when exposed to stressful experi- ences. Specifically, interviewers read the following script to each child:

“Think about when YOU feel stressed, nervous or worried. Circle HOW OFTEN you do each of these things before the stressful things happens, while you feel stressed, or after the stressful thing is over. Then, tell me HOW MUCH each thing helps you feel better when you feel stressed, nervous or worried.”

The SCSI has adequate psychometric properties: Ryan- Wenger (1990) reported a Cronbach’s alpha coefficient of .79 and moderate test-retest reliability (.73 to .82). In this study, eight items were excluded due to overlap with symp- toms of psychopathology (e.g., bite my nails; run away; fight someone; pick on someone, and hit, throw or break some- thing; cry or feel sad), resulting in a total of eighteen items. The 18 items reflecting effective emotion regulated coping

J Abnorm Child Psychol (2015) 43:203–215 207

(e.g., talk to someone; try to relax or stay calm; watch TV, read, or listen to music; try to forget about it; be by myself; cuddle a pet or stuffed animal; pray; sleep) were subjected to principal components analysis (PCA), which indicated the presence of a single factor structure with an eigenvalue above 1.0, explaining 27 % of the variance, thereby verifying a unidimensional construct. The 18-item effectiveness of regulated coping strat- egies obtained a Cronbach’s alpha coefficient of .82. To create a measure of the child’s self-reported emotionally-regulated cop- ing that did not simply list frequencies of coping strategy use, (which could represent multiple unsuccessful attempts at cop- ing), a combination of frequency and helpfulness scores for each of the eighteen items were used to represent the child’s ratings of effectiveness of their coping strategies. More specif- ically, helpfulness ratings were aggregated and then divided by the number of coping strategies endorsed by each individual child as a way to derive a mean effectiveness score for each child (Table 1). Means and standard deviations for the 18-items theorized to represent the effectiveness of emotionally- regulated coping construct are presented in Table 1.

Data Analyses

Primary data analyses were conducted using Mplus 6.12. In addition, R 2.15.1 was used to conduct preliminary analyses

and to perform post-hoc simple-slope tests probing interaction effects. Structural equation modeling (SEM) in Mplus was used to specify the measurement structure of latent constructs. Next, structural modeling was implemented to test hypothesized as- sociations among latent constructs. The χ2 statistic was used as the primary indicator of comparative model fit. However, to complement χ2 statistic, the Comparative Fit Index (CFI; Hu and Bentler 1999), Root-Mean Squared Error of Approximation (RMSEA; Browne and Cudeck 1993), and the Weighted Root Mean Square Residual (WRMR; Muthen and Muthen 2007) were employed as incremental indices of fit. Specifically, CFI values equaling 0.90 or greater, values of RMSEA of less than 0.08, and WRMR values of less than 1.0 were required for model acceptance. Finally, to test the hypothesized role of emotion regulated coping as a latent moderator of the association between ETV and mental health adjustment, the LMS/QML approach provided by Mplus (Muthen and Muthen 2007) was applied to create separate latent variables representing the interactions between each type of ETV and ERC using finite mixtures of normal distributions of latent variables. Chi-square difference tests were computed using the loglikelihood value and scale correction score to compare nested and alternative models with and without the latent interaction effect specified (Little 2011; Muthen and Muthen 2007). In accordance with procedures prescribed for estimating, testing and probing interactions in regression models (Aiken and West 1991; Preacher et al. 2006), potential- ly significant interactions were examined for their simple slopes, region of significance, and confidence bands.

Results

Missing Data Imputation

The missing data rate was less than 5 % for all study variables except for two items: maternal reports on child’s exposure to family physical and sexual abuse had 8.8 % (n=8) and 5.5 % (n=5) missingness rates, respectively. Little’s (2011) Missing Completely At Random (MCAR) test indicated that there were no identifiable systematic patterns in missing values,χ2[491]=498.49, p=.39, suggesting that missing data were MCAR. Full Information Maximum Likelihood (FIML) was used to estimate missing data in Mplus.

Descriptive Data and Preliminary Analyses

Means, standard deviations, and intercorrelations for all mea- sures of the study’s primary constructs are presented in Table 2. Multiple regression analyses were conducted to ex- amine associations between sociodemographic variables and scores on dependent measures representing mental health

Table 1 Schoolager’s coping strategies inventory items retained to mea- sure child reported effectiveness of emotionally regulated coping strategies

“How helpful is it?” Mean SD

1. Watch T.V. or listen to music 2.05 1.03

2. Say I’m sorry or tell the truth 2.17 1.04

3. Think about it 1.80 .96

4. Try to relax, stay calm 1.73 1.08

5. Try to forget about it 1.60 1.07

6. Do something about it 1.96 1.16

7. Draw, write, or read something 1.73 1.18

8. Sleep or take a nap 1.73 1.15

9. Talk to someone .75 1.10

10. Eat or drink 1.41 1.17

11. Play a game or something 1.49 1.27

12. Be by myself; be alone 1.53 1.08

13. Pray 1.58 1.28

14. Work (around the house) 1.31 1.18

15. Walk, run or ride my bike 1.50 1.28

16. Daydream 1.04 1.09

17. Cuddle a pet .97 1.17

18. Talk to self .75 1.11

The response scale ranged from 0 to 3 for all helpfulness ratings. The scale used to assess helpfulness included the following anchors: 0=Never do it, 1=Does not help, 2=Helps a little, 3=Helps a lot

208 J Abnorm Child Psychol (2015) 43:203–215

symptoms to identify and account for potentially significant covariates. Each regression equation contained seven indepen- dent variables: maternal education and child age were entered as continuous variables, and child sex, ethnicity, housing status, birthplace, and language preference were entered as categori- cals. The full model predicting anxiety symptoms was statisti- cally significant, adjusted R2=0.104, F(9, 78)=2.122, p<.05; none of the demographic variables made a statistically signifi- cant independent contribution to the model predicting anxiety symptoms. The full model predicting aggression symptoms was statistically significant, adjusted R2=0.153, F(9, 78)=2.458, p <.05; housing status made a significant unique contribution to the model predicting aggression symptoms, ß=0.24, p<.05. The full model predicting PTSD symptoms approached statistical significance, adjusted R2=0.068, F(9, 78)=1.709, p=.06, with both age, ß=0.23, p<.05, and housing status, ß=0.32, p<.05, making significant contributions to the model in predicting PTSD symptoms. To increase model parsimony, sociodemographic variables not significantly associated with mental health symptoms were dropped from analysis; of the potential sociodemographic covar- iates, only housing status and age were included in subsequent hypothesis testing analyses.

Rates of Exposure to Violence Over 93 % of the sample was exposed to at least one type of violence, 53 % of children were exposed to at-least two different types of violence, and 14 %

were exposed to three types of violence. Based on maternal reports, 15 % of children were physically abused, 13 % were sexually abused, and 7.7 % of children were reportedly exposed to both types of family violence. Mothers reported high rates of intimate partner violence: 43 % reported being pushed or shoved, 21 % reported having been beaten up with multiple blows, and 17 % reported having been strangled or choked. The rates of child exposure to community vio- lence were also very high: over 30 % of children reported having witnessed violence at school or in the neighbor- hood, 38 % of children reported worrying about a robbery or burglary of their home or family, and 30 % reported worrying about their safety where they live.

Correlations among the six exogenous indicators of violence exposure are presented in Table 2. The measurement model consisted of the following seven latent constructs: (1) exposure to family abuse, (2) intimate partner violence, (3) exposure to community violence, (4) emotionally-regulated coping (ERC), and mental health symptoms of (5) anxiety, (6) aggression and (7) PTSD. The χ2 test for this model was not significant, χ2(149)=59.326, p=0.26, which along with indices of relative fit, CFI=0.958, RMSEA=0.028, 90 % CI=0.000-0.058, WRMR=0.607, supports adequate model fit. The factor loadings in the measurement model relating latent con- structs with observed indicators operationally defining each of the theorized constructs were strong, positive, and statistically significant (range =0.356 to 0.954).

Table 2 Descriptive statistics and bivariate correlations of study variables

Variable 1 2 3 4 5 6 7 8 9

1. EFA-P - .036 .348** .097 .079 .239* .250* .359** .319**

2. CTS-P - .055 -.016 -.137 .097 -.099 .222* .007

3. LEQ–P - .381** .126 .165 .178 .397** .199

4. LEQ–C - .482** .004 .221* .365** -.009

5. SWS-C - -.064 .435** .229* -.004

6. ERC-C - -.022 -.069 -.031

7. RCMAS-C - .104 .125

8.CBCL– Agg./P - .571**

9.CBCL– PTSD/P -

Mean (SD)

.253 (.569)

2.87 (3.6)

.572 (1.01)

.674 (.914)

2.15 (2.99)

1.11 (.41)

10.84 (6.53)

9.71 (7.93)

5.05 (4.92)

N 91 91 91 89 89 91 89 89 89

p<0.10 *p<0.05 ** p<0.01. EFA, exposure to family abuse; CTS, conflict tactics scale; LEQ, life events questionnaire; ECV, exposure to community violence; SWS, strains and worries scale; ERC, emotionally regulated coping; RCMAS, revised children’s manifest anxiety scale; CBCL, child behavior checklist, Agg., aggression scale; C, child reported; P, parent reported

J Abnorm Child Psychol (2015) 43:203–215 209

Type of Violence Exposure and Child Mental Health Symptoms

In the structural model, the three endogenous latent constructs representing anxiety, aggression and PTSD were regressed onto the four exogenous latent constructs representing each type of ETV and ERC. Two sociodemographic covariates– housing status and child age–were specified in the structural model as exogenous observed variables on which the endoge- nous latent variables were regressed (see Fig. 1). The structural model represented adequate fit for the data, χ2(186)=216.218, p=.064, CFI=0.892, RMSEA=0.042, 90 % CI=0.000- 0.064,WRMR=0.737. On the basis of non-significant probabil- ity, which was conferred by the Wald test, parameters were trimmed one at a time in order to improve estimation of the hypothesized structural association among latent exogenous and endogenous constructs (Chou and Bentler 1990). This data- driven decision making process was situated within the broader context of theory, such that model modifications and selection of a “better” model were consistent with the theoretical framework informing the hypotheses and analytic procedure.

The following parameters were trimmed to improve identi- fication of the hypothesized structural associations among latent exogenous and endogenous constructs: the regression of

anxiety on exposure to family abuse and intimate partner vio- lence; the regression of aggression on exposure to family abuse and ERC; and the regression of PTSD on intimate partner violence and exposure to community violence. The main ef- fects trimmed structural model represented good fit for the data, χ2(192) =215.168, p=0.12, CFI=0.917, RMSEA=0.036, 90 % CI=0.00-0.060,WRMR=0.748. Improvement in model fit be- tween the trimmed and untrimmed structural models was sta- tistically significant, χ2d =57.03, dfd =6, p<.001.

As displayed in Fig. 2, there were statistically significant regression coefficients for the paths linking anxiety to exposure to community violence, ß=0.829, p<.001, and ERC, ß=−0.413, p<.05, aggression to intimate partner violence, ß=0.458, p=.052, and exposure to community violence, ß=1.303, p<.05, and PTSD to exposure to family abuse, ß=0.498, p<.001,and ERC, ß=−0.286, p<.05. The regression coefficient for the path linking aggression to child housing status was not significant, ß=0.726, p=.13. There were significant regression coefficients for the paths linking PTSD to child age, ß=0.115, p<0.05, and housing status, ß=0.708, p<.05. This main effects structural model was retained as the nested (null) model against which subsequent alternative models investigating the hypothe- sized moderating role of effective ERC in the association be- tween type of ETVand mental health symptoms were compared.

Sexual Abuse

Physical Abuse

Family Abuse

CTS3

CTS2

CTS1

Intimate Partner

Violence

Strains & Worries Scale (C)

Life Events Questionnaire (P)

Life Events Questionnaire (C)

Community Violence

RCMAS1

RCMAS2

RCMAS3

Anxiety

CBCL1

CBCL2

CBCL3

Aggression

CBCL4

CBCL5

CBCL6

PTSD

Age

Housing Status

SCSI1 SCSI2 SCSI3

Emotionally Regulated

Coping

Fig. 1 Trimmed model illustrating links between type of violence expo- sure and mental health symptoms. Nonsignificant path coefficients of endogenous variables regressed on demographic covariates were not trimmed in the final structural model. Solid lines represent statistically

significant regression paths. Dashed lines indicate non-significant paths. P, parent reported; C, child reported; CTS, conflict tactics scale; CBCL, child behavior checklist; RCMAS, revised children’s manifest anxiety scale; SCSI, school-ager’s coping strategies inventory

210 J Abnorm Child Psychol (2015) 43:203–215

Testing the Role of Emotionally-Regulated Coping (ERC) as a Moderator

There was a significant difference between the null and alter- native structural models, χ2d =12.86, dfd=6, p=.045, when the latent exogenous interaction term between exposure to family abuse and ERC was added to the main effects model (see Fig. 2). In contrast, the models with the other two interaction terms tested, (between intimate partner violence and ERC, χ2d =3.29, dfd=6, p=0.77, and between exposure to community violence and ERC, χ2d =5.72, dfd =6, p=0.46, did not significantly improve model fit when compared to the null, main effects model.

The conditional effect of exposure to family abuse in the model predicting anxiety was positive and significant, ß=0.754, p<.001, and the conditional effect of ERC was negative and nonsignificant, ß=−0.349, p=.11. Consistent with the hypothesized protective role of ERC, the regression coefficient for the path linking anxiety to the latent interaction term between exposure to family abuse and ERC was negative and significant, ß=−0.460, p<.01, indicating that the positive association between exposure to family abuse and anxiety is

reduced for children and youth with higher overall ERC. Post hoc simple slopes analyses were performed to determine whether the simple slopes for the regression of anxiety on exposure to family abuse at high and low levels of ERC were significantly different from zero. For children with higher levels of ERC (+1 standard deviations [SD]), the simple slope was 0.629, SE[ß]=0.161, t=3.92 , p<.001; children with lower levels of ERC (−1 SD), the simple slope was 0.878, SE[ß]=0.196, t=4.47, p<.001. The regions of significance for the ERC moderator ranged from 0.829 to 8.22, indicating that any difference in slopes outside this range is statistically significant. Since the observed values for the moderator ranged from -.51 to 0.8, the effect of exposure to family abuse on anxiety was statistically significant for all observed values of ERC.

In the model predicting aggression, the conditional effect of exposure to family abuse was positive and significant, ß=0.971, p<.001, and the conditional effect of ERC was negative and nonsignificant, ß=−0.096, p=.12. Also consis- tent with the hypothesized protective role of ERC, the regres- sion coefficient for the path linking aggression to the interac- tion term between exposure to family abuse and ERC was

Sexual Abuse

Physical Abuse

Family Abuse

CTS3

CTS2

CTS1

Intimate Partner

Violence

Strains & Worries Scale (C)

Life Events Questionnaire (P)

Life Events Questionnaire (C)

Community Violence

RCMAS1

RCMAS2

RCMAS3

Anxiety

CBCL1

CBCL2

CBCL3

Aggression

CBCL4

CBCL5

CBCL6

PTSD

Age

Housing Status

SCSI1 SCSI2 SCSI3

Emotionally Regulated

Coping

Fig. 2 Alternative model with the interaction term s exposure to family abuse and emotionally-regulated coping specified. All exogenous vari- ables that were trimmed in the main effects model, but involved in the interaction effect were retained in the alternatively model. Solid lines represent statistically significant regression paths. Dashed lines indicate

non-significant paths. Dotted lines reference statistically significant paths for mental health variables regressed on the latent interaction term. P, parent reported; C, child reported; CTS, conflict tactics scale; CBCL, child behavior checklist; RCMAS, revised children’s manifest anxiety scale; SCSI, school-ager’s coping strategies inventory

J Abnorm Child Psychol (2015) 43:203–215 211

negative and significant, ß=−0.376, p<.001, indicating that the positive association between exposure to family abuse and aggression is reduced for children and youth with higher overall ERC. Simple slope tests were performed to decompose the interaction between exposure to family abuse and ERC in the model predicting aggression: for children with higher levels of ERC (+1 SD), the simple slope was 0.869, SE[ß]=0.186, t=4.66, p<.001; for children with lower levels of ERC (−1 SD), the simple slope was 1.073, SE[ß]=0.164, t=6.56 , p<.001. The region of significance on the moderator ERC ranged from 1.23 ot 6.68, indicating that any given slope outside this range is statistically significant. Since observed values for the moderator ranged from .51 to 0.8, this indicates that the effect of exposure to family abuse on aggression is statistically significant for all observed values of ERC.

In the model predicting PTSD, the conditional effect of exposure to family abuse was positive and significant, ß=0.787, p<.001, and the conditional effect of ERC was negative and significant, ß=−0.169, p<.05. The regression coefficient for the path linking PTSD to the interaction term between exposure to family abuse and ERC was negative and significant, ß=−0.151, p<.01, indicating that the positive association between exposure to family abuse and PTSD is reduced for children and youth with higher overall ERC. Simple slope tests were performed to decompose the interac- tion between exposure to family abuse and ERC in the model predicting PTSD: for children with higher levels of ERC (+1 SD), the simple slope was 0.746, SE[ß]=0.554, t=1.35 , p=n.s.; for children with lower levels of ERC (−1 SDs), the simple slope was 0.828, SE[ß]=0.374, t=2.21 , p<.05. The regions of significance on the ERC moderator ranged from −0.696 to −0.082, indicating that any differences in slopes inside this range are statistically significant. However, since observed values for the moderator ranges from -.51 to 0.8, this indicates that the effect of exposure to family abuse on PTSD is statistically significant for lower overall values of ERC.

Post-hoc analyses of interactions indicated that the simple slope of exposure to family abuse on anxiety and aggression was statistically significant for higher and lower levels of

ERC. By contrast, probing tests indicated that the simple slope for the regression of PTSD on exposure to family abuse was significant only at the lower conditional values of the moder- ator ERC. Plots illustrating interactions between exposure to family abuse and ERC are displayed in Fig. 3.

Discussion

Consistent with the existing literature on child violence exposure, results of this study confirm that children living in conditions of poverty with an accumulation of demographic risk factors (e.g., single-parent head of household, racial-eth- nic minority status), regardless of homelessness or housing status, are at high risk for multiple violence exposures (Finkelhor et al. 2009; Hanson et al. 2006). Specifically, 93 % of the sample was exposed to at least one type of violence, 53 % of children were exposed to at least two different types of violence, and 14 % were exposed to three different types of violence.

As hypothesized, multi-level violence exposures were sig- nificantly associated with mental health maladjustment. Of note, exposure to family abuse was a significant unique pre- dictor of PTSD symptomatology and intimate partner violence was significantly associated with symptoms of aggression. In contrast, more distal violence exposures in the community were more generally associated with internalizing and exter- nalizing symptomatology. Moreover, the association between exposure to family abuse and internalizing and externalizing problems and PTSD symptoms was moderated by child ERC, suggesting that ERC does play a protective role following exposure to family abuse. These findings highlight the spec- ificity of proximal, family-based and more distal community- based violence exposures as ecologically-based risks for psy- chopathology. At the same time, these results suggest the critical role of child adaptive coping strategies, specifically emotionally-regulated coping, as a protective resource in re- lation to mental health symptoms in the presence of multiple forms of violence. Nonetheless, more rigorous prospective,

Fig. 3 Plots illustrating the interaction of exposure to family abuse and emotionally-regulated coping strategies. More positive slopes relate to lower levels of emotionally-regulated coping scores. ERC, emotionally regulated coping; RCMAS, revised childrenvs manifest anxiety scale;

CBCL, child behavior checklist. High and low levels of ERC correspond to +1 and −1 standard deviations away from the centered mean of ERC respectively

212 J Abnorm Child Psychol (2015) 43:203–215

longitudinal studies involving comprehensive assessment of violence exposures and child coping resources are needed.

Consistent with the ecological stress process model, these findings suggest that the ecological setting in which children are exposed to violence may present mental health challenges that are further compounded by limited resources across mul- tiple ecological levels that characterize communities impacted by poverty, discrimination, and other forms of social inequality (Foster and Brooks-Gunn 2009; Sandler 2001). Proximal vio- lence exposures, such as those encompassing direct victimiza- tion in the context of family relationships and/or the home environment may have different mental health impacts than more distal violence exposures, including indirect experiences as witnesses of violence in the neighborhood or school. Although previous developmental research has consistently shown that violence exposures are often associated with ele- vated internalizing and externalizing behavior problems (e.g., Margolin and Gordis 2000), there is also some support in the literature for specific links between different types of child violence exposure and particular domains of mental health maladjustment. In a review of recent child maltreatment re- search, Kearney and colleagues (2010) suggested that there may be a distinct relationship between children who are the targets of family abuse and mental health symptoms of PTSD. In addition, children who witness parental violence can be at particular risk for developing externalizing problems, such as aggressive behaviors (Chan and Yeung 2009; Yates et al. 2003). Similarly, this study’s findings of significant associa- tions between community violence exposures and mental health symptoms of aggression and anxiety corroborate find- ings reflected in much of the empirical literature on children’s experiences with community violence as a risk factor for behavioral and mental health maladjustment (Dempsey et al. 2000; Kliewer et al. 2006; Hanson et al. 2006).

The results of this study lend partial support for the second hypothesis, which postulated that children who perceived their emotional coping as more effective would report fewer mental health symptoms. In addition, the association between exposure to family abuse and mental health maladjustment was moderated by effectiveness of ERC; post-hoc decomposing of this interac- tion further reinforced the theorized role of ERC as a moderator. As child ratings of the effectiveness of ERC decreased, the slope relating exposure to family abuse to mental health symptoms became more strongly positive, suggesting that the impacts of family-based violence exposures may vary as a function of how children experience their ERC strategies as being effective and helpful. This result has not been reported in previous research.

By contrast, analyses did not render significant statistical support for the theorized role of ERC as a modifier of the relationship between either intimate partner violence or expo- sure to community violence and mental health symptoms. It is plausible to argue that due to limited sample size and power, the presence of the abovementioned interactions was not

detected. This finding is not consistent with some of the scant research that has examined child coping in the presence of specific types of violence exposure. In one study, adaptive coping strategies were found to operate as a protective factor in the presence of community violence, whereby community violence was associated with higher levels of externalizing behavior problems among youth who reported lower coping effectiveness (Brady et al. 2008). Another study reported that child emotional and physiological self-regulatory capacities mediated the association between marital emotional and phys- ical abuse and child externalizing problems (Cummings et al. 2009). Comparison across studies of child violence exposure and coping is challenging because conceptualization and mea- surement of both child coping and violence exposure vary across studies as do the nature of the samples studied.

It is possible that the way in which ERC was measured influenced the current findings. For example, ERC was not comprehensively measured through direct assessment of chil- dren’s coping strategies in the presence of violence in different settings, such as home, neighborhood, and classroom. Thus, our theoretical definition and operational measurement of adaptive coping strategies as ERC may have emphasized strategies that were particularly relevant for the home envi- ronment (e.g., go to sleep, watch TV) and not those that may be contextually appropriate for coping with exposure to com- munity violence (e.g., taking the bus to school instead of walking, participating in youth programs).

The strengths of the present study include a) studying a highly underrepresented, ethnically diverse sample of mothers who are single parents and children living in conditions of poverty, b) assessment of ETV across multiple ecological settings, and c) theory-driven examination of the role of child coping as an intervening process that could buffer the influ- ences of multiple violence exposures on mental health. As a necessary departure from the theoretical and methodological approaches undertaken in past research, the present study distinguished adaptive coping strategies from symptoms of psychopathology in our conceptualization of ERC as specific behaviors that could regulate the impacts of ETV without placing the child at further risk of additional violence expo- sures. Nonetheless, findings of the present study must be interpreted in light of several limitations, including small sample size, cross-sectional analyses, single parental infor- mant for family violence exposures and mental health symp- toms of aggression and PTSD, and limited assessment of violence exposures (e.g. failure to assess the intensity and chronicity of exposure or the child’s proximity to violence). The Conflict Tactics Scale (Straus 1979) used in the present study asks mothers about experiences as victims of intimate partner violence during adulthood, and not whether the child in the sample was aware of, or exposed to, this experience. Because the study did not have a standardized measure of child exposure to community violence, a measure was

J Abnorm Child Psychol (2015) 43:203–215 213

constructed from three different non-standardized life events subscales. Finally, as previously noted, the ERC construct suffers from limitations which may have influenced analyses of the role of child coping as a protective factor across eco- logical settings wherein violence can be perpetrated.

The body of empirical research upon which the conceptual framework and findings of this study are established suggest that supporting children’s coping strategies, whether as pre- vention or a treatment strategy, could buffer the detrimental effects of some violence exposures (Foster and Brooks-Gunn 2009). In treatment, children presenting with symptoms of anxiety, aggression and PTSD could benefit from incorporat- ing protocols enhancing their use of self-regulation strategies that emphasize emotion regulation (e.g. talk to someone, play, write, etc.). However, clinicians working with children may need to explore the different types of violence as well as the contexts of violence exposures, and the degree to which ERC strategies may be relevant and effective for coping with the stressors impacting their lives. Intervention efforts targeting children’s ERC may be particularly relevant as a preventative measure, because they can be applied in the context of family and community violence without situating the child for further risk for direct victimization and psychopathology (Sandler 2001). Future research should incorporate more comprehen- sive measures of children’s ETV and adaptive coping strate- gies over time through longitudinal research. Although this study distinguished between constructs representing family and community violence exposures, it is not clear whether the ETV-mental health links observed in the present study are due to cumulative impacts or differential influences of types of violence exposures. Subsequent work investigating the im- pacts of ETV should try to evaluate the ways in which vio- lence exposures combine and accumulate across ecological and developmental contexts. Towards that end, a more com- prehensive assessment approach should involve the participa- tion of children, siblings, and caregivers, as well as teachers because each offers an important perspective on the child’s ETV at home, in the neighborhood, on the bus, and in the school playground.

Author Note This study was based on secondary analysis of data archived by the Murray Center for the Study of Lives, Harvard Univer- sity, from the Worcester Family Research Project conducted by the National Center on Family Homelessness and the University of Massa- chusetts Medical Center at Worcester with funding from the National Institute of Mental Health (NIMH) and the U.S. Maternal and Child Health Bureau. We are indebted to the families and children who partic- ipated in the Worcester Family Research Project. We also thank Debra Gustafson, Kristopher Preacher, Ista Zahn and Howard Crumpton for their support and assistance with data preparation and analyses, including measurement modeling and latent interaction tests focusing on children’s own experiences of coping effectiveness.

Conflict of Interest The authors declare that they have no conflict of interest.

References

Achenbach, T. M. (1991). Integrative guide for the 1991 CBCL/4-18, YSR and TRF profiles. Burlington: Department of Psychiatry, University of Vermont.

Achenbach, T. M., & Rescorla, L. A. (2007). Multicultural Supplement to the Manual for the ASEBA School-Age Forms & Profiles. Burlington: University of Vermont, Research Center for Children, Youth, & Families.

Aiken, L. S., & West, S. G. (1991). Multiple regression: Testing and interpreting interactions. Newbury Park: Sage.

Aldwin, C., Skinner, E., Zimmer-Gembeck, M., & Taylor, A. (2011). Coping and self-regulation across the life-span. In Fingerman (Ed.). 563–589.

Ayers, T. S., Sandler, I. N., West, S. G., & Roosa, M. W. (1996). A dispositional and situational assessment of children’s coping: testing alternative models of coping. J Pers, 64, 923–958.

Barrow, S., Hellman, F., Lovell, A., Plapinger J., Robinson D., & Struening E. (1985). Personal history form. New York: New York State Psychiatric Institute.

Bassuk, E. L., Weinreb, L. F., Buckner, J. C., Browne, A., Salomon, A., & Bassuk, S. S. (1996). The characteristics and needs of sheltered homeless and low-income housed mothers. J Am Med Assoc, 276, 640–646.

Brady, S. S., Gorman-Smith, D., Henry, D. B., & Tolan, P. H. (2008). Adaptive coping reduces the impact of community violence expo- sure on violent behavior among African American and Latino male adolescents. J Abnorm Child Psychol, 36, 105–115.

Browne, M. W., & Cudeck, R. (1993). Alternative ways of assessing model fit. In K. A. Bollen & J. S. Long (Eds.), Testing structural equation models. Newbury Park: Sage.

Buckner, J., Mezzacappa, E., & Beardslee, W. (2003). Characteristics of resilient youths living in poverty: the role of self-regulatory process- es. Dev Psychopathol, 15, 139–162.

Buckner, J. C., Beardslee, W. R., & Bassuk, E. L. (2004). Exposure to violence and low-income children’s mental health: direct, moderated, and mediated relations. Am J Orthopsychiatry, 74, 413–423.

Chan, Y., & Yeung, J. (2009). Children living with violence within the family and its sequel: a meta-analysis from 1995–2006. Aggress Violent Behav, 14, 313–322.

Chou, C. P., & Bentler, P. M. (1990). Model modification in covariance structure modeling: a comparison among the likelihood ratio, Lagrange Multiplier, and Wald tests. Multivar Behav Res, 25, 115– 136.

Cicchetti, D., & Toth, S. L. (2005). Child maltreatment. Annu Rev Clin Psychol, 1, 409–438.

Clarke, A. (2006). Coping with interpersonal stress and psychosocial health among children and adolescents: A Meta-Analysis. J Youth Adolesc, 35, 11–24.

Compas, B., O’Connor-Smith, J. K., Saltzman, S., Thomsen, A. H., & Wadsworth, M. E. (2001). Coping with stress during childhood and adolescence: problems, progress, and potential in theory and re- search. Psychol Bull, 127, 87–127.

Cummings, E. M., El-Sheikh, M., Kouros, C. D., & Buckhalt, J. A. (2009). Children and violence: the role of children’s regulation in the marital aggression child adjustment link. Clin Child Fam Psychol Rev, 12, 13–15.

Dempsey, M., Overstreet, S., & Moely, B. (2000). “Approach” and “avoidance” coping and PTSD symptoms in inner-city youth. Curr Psychol: Dev, Learn, Personal, Soc, 19, 28–45.

Egeland, B., Yates, T., Appleyard, K., & van Dulmen, M. (2002). The long-term consequences of maltreatment in the early years: a devel- opmental pathway model to antisocial behavior. Child Serv: Soc Pol, Res, & Pract, 5, 249–260.

214 J Abnorm Child Psychol (2015) 43:203–215

Eisenberg, N., & Silver, R. (2011). Growing up in the shadow of terror- ism: youth in America after 9/11. Am Psychol, 66, 468–481.

Eisenberg, N., Fabes, R., & Guthrie, I. (1997). Coping with stress: the roles of regulation and development. In S. A. Wolchik & I. N. Sandler (Eds.), Handbook of children’s coping: linking theory, research and interventions (pp. 41–70). New York: Plenum Press.

Eisenberg, N., Hofer, C., & Vaughan, J. (2007). Effortful control and its socioemotional consequences. In J. J. Gross (Ed.), Handbook of emotion regulation (pp. 287–306). New York: Guilford Press.

Finkelhor, D., Turner, H., Ormrod, R., Hamby, S. & Kracke, K. (2009). Children’s exposure to violence: A comprehensive national survey. Office of Juvenile Justice, US Department of Justice October, 2009.

Folkman, S., & Moskowitz, J. T. (2004). Coping: pitfalls and promise. Annu Rev Psychol, 55, 745–774.

Fosco, G., DeBoard, R., & Grych, J. (2007). Making sense of family violence: implications of children’s appraisals of interparental aggres- sion for their short- and long-term functioning. Eur Psychol, 12, 6–16.

Foster, H., & Brooks-Gunn, J. (2009). Toward a stress process model of children’s exposure to physical family and community violence. Clin Child Fam Psychol Rev, 12, 71–94.

Grant, K. E., & Compas, B. E. (1995). Stress and symptoms of depres- sion/anxiety among adolescent: searching for mechanisms of risk. J Consult Clin Psychol, 63, 1015–1021.

Grant, K. E., O’Koon, J. H., Davis, T. H., Roache, N. A., Poindexter, L. M., Armstrong, M. L., et al. (2000). Protective factors affecting low- income urban African American youth exposed to stress. J Early Adolesc, 20, 388–417.

Hanson, R., Self-Brown, S., Fricker-Elhai, A., Kilpatrick, D., Saunders, B., & Resnick, H. (2006). Relations among parental substance use, violence exposure and mental health: the national survey of adoles- cents. Addict Behav, 31, 1988–2001.

Hu, L., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covari- ance structure analysis: conventional criteria versus new alterna- tives. Struct Equ Model, 6, 1–55.

Kearney, C., Wechsler, A., Kaur, H., & Lemos-Miller, A. (2010). Posttraumatic stress disorder in maltreated youth: a review of contem- porary research and thought. Clin Child Fam Psychol Rev, 13, 46–76.

Kliewer, W., & Sullivan, T. (2008). Community violence exposure, threat appraisal, and adjustment in adolescents. J Clin Child Adolesc Psychol, 37, 860–873.

Kliewer, W., Parrish, K., Taylor, K., Jackson, K., Walker, J., & Shivy, V. (2006). Socialization of coping with community violence: influ- ences of caregiver coaching, modeling, and family context. Child Dev, 77, 605–623.

Lazarus, R. S., & Folkman, S. (1984). Stress, appraisal, and coping. New York: Springer.

Little, T. D. (2011). Longitudinal structural equation modeling. New York: Guilford press.

Lynch, M. (2003). Consequences of children’s exposure to community violence. Clin Child Fam Psychol Rev, 6, 265–266.

Margolin, G., & Gordis, E. B. (2000). The effects of family and commu- nity violence on children. Annu Rev Psychol, 51, 445–479.

Masten, A. S., Neeman, J., & Andenas, S. (1994). Life events and adjustment in adolescents: the significance of event independence, desirability, and chronicity. J Res Adolesc, 4, 71–97.

Muthen, L. K., & Muthen, B. O. (2007). Mplus User’s Guide (5th ed.). Los Angeles: Muthen & Muthen.

Pearlin, L. I. (1999). The stress process revisited: Reflections on concepts and their interrelationships. In C. S. Anenshensel & J. C. Phelan (Eds.), Handbook of the sociology of mental health (pp. 395–415). New York: Kluwer.

Pina, A. A., Little, M., Knight, G. P., & Silverman, W. K. (2009). Cross- ethnic measurement equivalence of the RCMAS in Latino and White youth with anxiety disorders. J Pers Assess, 91, 58–61.

Preacher, K. J., Curran, P. J., & Bauer, D. J. (2006). Computational tools for probing interaction effects in multiple linear regression, multi- level modeling, and latent curve analysis. J Educ Behav Stat, 31, 437–448.

Reynolds, C. R., & Richmond, B. O. (1985). Revised Children’s Manifest Anxiety Scale. RCMAS Manual. Los Angeles: Western Psychological Services.

Ryan-Wenger, N. M. (1990). Development and psychometric properties of the schoolager’s coping strategies inventory. Nurs Res, 39, 346– 349.

Sandler, I. N. (2001). Qualities and ecologies of adversity as common mechanism of risk and resilience. Am J Community Psychol, 29, 19–61.

Sandler, I. N., Tein, J., Mehta, P., Wolchik, S., & Ayers, T. (2000). Coping efficacy and psychological problems of children of divorce. Child Dev, 71, 1099–1118.

Straus, M. A. (1979). Measuring intrafamily conflict and violence: the conflict tactics (ct) scales. J Marriage Fam, 41, 75–88.

Tolan, P., Gorman-Smith, D., & Henry, D. (2006). Family violence. Annu Rev Psychol, 57, 557–583.

Wolfe, V. V., Gentile, C., & Wolfe, D. A. (1989). The impact of sexual abuse on children: a PTSD formulation. Behav Ther, 20, 215–228.

Yates, T. M., Dodds, M. F., Sroufe, L. A., & Egeland, B. (2003). Exposure to partner violence and child behavior problems: a prospective study controlling for child physical abuse and neglect, child cognitive ability, socioeconomic status, and life stress. Dev Psychopathol, 15, 199–218.

J Abnorm Child Psychol (2015) 43:203–215 215

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.