- Email: [email protected]
Racial variations in the link between domestic violence and children's behavioral outcomes
Children and Youth Services Review 44 (2014) 90–99
Contents lists available at ScienceDirect
Children and Youth Services Review journal homepage: www.elsevier.com/locate/childyouth
Racial variations in the link between domestic violence and children's behavioral outcomes Jeong Ah Yoo ⁎ School of Social Welfare, Ewha Womans University, Ewhayeodae-gil, Seodaemun-gu, Seoul 120-750, South Korea
a r t i c l e
i n f o
Article history: Received 25 October 2013 Received in revised form 9 May 2014 Accepted 1 June 2014 Available online 11 June 2014 Keywords: Domestic violence Children Racial variation Behavior problems
a b s t r a c t The link between domestic violence and children has been well documented, but little is known about racial variations in this relationship. Using multiple waves of the Fragile Families and Child Wellbeing Study, this study examined whether the effects of domestic violence on children's behavior problems differed by race. Results from structural equation modeling revealed that domestic violence had negative effects on children in all racial groups; but the mechanisms through which domestic violence exerted its effects varied across groups. Whereas the effects of domestic violence on children's behavior problems were partially mediated through maternal functioning for White and Black families, only direct effects were found for Hispanic families. This study highlights wide racial variations in the mechanisms through which domestic violence affects children and the importance of cultural context in understanding those variations. © 2014 Published by Elsevier Ltd.
1. Introduction Domestic violence has received considerable attention as a serious social problem that impacts not only women, who are the most frequent victims, but also their children. For years researchers have studied the mechanisms or processes through which domestic violence affects children's behavior problems (Huang, Wang, & Warrener, 2010; Levendosky, Huth-Bocks, Shapiro, & Semel, 2003; Levendosky, Leahy, Bogat, Davidson, & von Eye, 2006; Schoppe-Sullivan, 2007; Yoo & Huang, 2012, 2013). Many of these studies, largely based on ecological frameworks, suggested that mothers' responses to domestic violence may mediate the effects that the violence has on their children. In particular, disruptions in maternal mental health and parenting, which are mostly caused by domestic violence, have been considered key mechanisms that mediate the effects of domestic violence on children's behavioral outcomes. Although the impact of domestic violence on women and children has been well documented in the literature, much less attention has been given to racial variations in this relationship. Some researchers have suggested that women's responses to domestic violence may vary by their racial and cultural backgrounds (Kasturinrangan, Krishnan, & Riger, 2004; Sokoloff & Dupont, 2005). They have found that the risk of experiencing psychological and parenting difficulties as a result of domestic violence may be greater for White women than for minority women (Campbell et al., 2008; Edelson, Hokoda, & Romos-Lira, 2007; Taylor, Guterman, Lee, & Rathouz, 2009; Torres & Han, 2000). These studies have suggested that cultural contexts may influence abused women's psychological resources and ⁎ Tel.: +82 10 5465 7271. E-mail address: [email protected].
http://dx.doi.org/10.1016/j.childyouth.2014.06.001 0190-7409/© 2014 Published by Elsevier Ltd.
parenting that are closely associated with their children's outcomes. However, no study has focused on the potential racial variations in the mechanisms through which domestic violence affects children's behavioral outcomes. Ecological theory (Belsky, 1984; Bronfenbrenner, 1979) has been widely used to understand the effects of domestic violence on children's behavioral development. The theory suggests that children's behavioral outcomes can be influenced by children's characteristics (i.e., gender and temperament), parents' characteristics (i.e., maternal mental health and parenting), and contextual factors (e.g., domestic violence). The theory regards the mother's psychological resources and parenting functioning as the most influential determinants of children's outcomes because most children, especially young children, are completely dependent on their mother for all aspects of care. Ecological theory also posits that, as a contextual factor, domestic violence can have substantial effects on children's behavioral outcomes. Such effects can be direct, and they can also occur indirectly through the negative effects that domestic violence exerts on maternal psychological well-being and parenting functioning. That is, the ecological model assumes that domestic violence can impair maternal psychological well-being and quality of parenting and thus can have negative effects on children's behavioral outcomes. Previous studies have provided empirical support for the ecological framework, which explains the mechanisms through which domestic violence affects children (Huang et al., 2010; Levendosky et al., 2003, 2006; Schoppe-Sullivan, 2007; Yoo & Huang, 2012, 2013). For example, in a longitudinal study on high-risk young children, Huang et al. (2010) found that domestic violence at Wave 1 had a negative effect on maternal mental health and reduced parenting qualities at Wave 2, which in turn increased children's behavior problems at Wave 3.
J.A. Yoo / Children and Youth Services Review 44 (2014) 90–99
Ecological theory also suggests that the influence of family-level factors on children's outcomes may vary by their relationships to macro-environmental contexts (Bronfenbrenner, 1979), which includes cultural context and the family's socioeconomic status. However, no study has focused on examining whether the mechanisms based on ecological theory apply to culturally diverse racial groups. Most existing studies have included race in their control variables but have not directly compared the effects of domestic violence on children across racial groups. Thus, this study explored whether the mechanisms linking domestic violence to children's behavioral problems in the ecological framework vary by race or cultural background. Cross-cultural or multicultural domestic violence studies have suggested that women's responses to domestic violence are in part contingent on the cultural contexts in which they live (Campbell et al., 2008; Edelson et al., 2007; Kasturinrangan et al., 2004). In the United States, race represents a key factor that defines those contexts. The research has indicated that culture has an important influence on abused women's responses, such as mental health problems (Campbell et al., 2008; Taylor et al., 2009; Torres & Han, 2000) and parenting difficulties (Edelson et al., 2007; Taylor et al., 2009). For example, Campbell et al. (2008) found that Black women who are abused tend to experience lower levels of mental health problems than do White women. Likewise, Torres and Han (2000) found that White women are more likely to experience psychological distress than do Hispanic women. In addition, several studies have reported racial differences in abused women's parenting difficulties (Edelson et al., 2007; Taylor et al., 2009). Edelson et al. (2007) reported that Hispanic women were more likely than White women to experience parenting stress. But Taylor et al. (2009) found that Hispanic abused women were less aggressive toward their children than Black or White women. These studies suggest that certain cultural contexts, particularly within minority communities, may moderate the negative effects of domestic violence on their mental health and parenting. Given the influence of cultural contexts on abused women's mental health and parenting – both of which are strong factors of children's outcomes – it does not seem reasonable to assume that minority children are affected by domestic violence through the same mechanisms or processes as White children. Some researchers have suggested that racial variations in women's and children's outcomes may be attributable not only to racial or cultural backgrounds but also to differences in socioeconomic status across racial groups, such as income, education, or marital status (Campbell et al., 2008; Sokoloff & Dupont, 2005; Yoo & Huang, 2012). In their qualitative research, Campbell et al. (2008) suggested that the relatively weak effect of domestic violence on minority families can be in part attributable to racial differences in socioeconomic characteristics. They also found that domestic violence was often not the most salient psychological stressor for Black women and that lack of financial resources or social stressors (i.e., racism) was equally or more important. This finding suggests that comparisons among racial groups could be invalid if socioeconomic status is not taken into account. Few empirical studies have examined how race or culture plays a role in the links between domestic violence, women's psychological and parenting functioning, and their children's behavioral outcomes. The paucity of empirical evidence is unfortunate because such information could be critical to better understand the complex effects of domestic violence on children and to develop more culturally sensitive services for supporting children and their mothers. Therefore, the present study addressed this gap in the literature by analyzing the potential racial variations in the effects of domestic violence on women's functioning and thereby their children's behavioral outcomes while controlling for socioeconomic characteristics. This study paid special attention to whether the mechanisms linking domestic violence's effects on children in ecological frameworks apply to culturally diverse children and whether there are culture-specific mechanisms through
91
which domestic violence affects children. Three racial groups – Black, White, and Hispanic families that live in the United States – were compared in this study. Within-group differences were not considered in this study because the sample sizes were not sufficient for separate analyses. Although minority communities are heterogeneous in terms of cultural backgrounds (e.g., country of origin, immigration status, or acculturation level), it is expected that minority women do share common influences (e.g., racism or sexism) in their experiences of domestic violence within a sociopolitical context (Kasturinrangan et al., 2004, p 320).
2. Methods 2.1. Data Data for this study were drawn from the Fragile Families and Child Wellbeing Study, a longitudinal birth cohort survey of 4898 mothers of newborns in 20 cities in the United States. This study oversampled children born to unmarried parents by design with a 3:1 ratio (3711 unmarried and 1187 married), which resulted in a larger number of low-income children and Black or Hispanic children than would be the case in a nationally representative sample (Reichmann, Teitler, Garfinkel, & McLanahan, 2001). The baseline core interviews were conducted with parents in the hospital at the time of the focal child's birth between 1998 and 2000. Follow-up surveys were conducted when the child was 1, 3, 5, and 9 years old. Because the present study focused on behavioral outcomes of preschool children, the first four waves of the surveys were used: baseline, Year 1, Year 3, and Year 5. Of the eligible mothers at baseline, 4365 were interviewed at Year 1, 4231 were interviewed at Year 3, and 4139 were interviewed at Year 5. After the age-3 and age-5 core interviews, mothers were asked to participate in an in-home assessment of parenting and child development. For the in-home assessments, 3254 mothers participated at Year 3, and 2977 participated at Year 5. This study used the two in-home assessments, which provided valuable information on the observer's evaluations of parenting but resulted in substantial loss of data. A total of 2404 mothers completed all six data collections across the 5 survey years. The 402 mothers who were not involved in a relationship with the child's biological father at baseline and Year 1 were not included in the current study. An additional 768 mothers were dropped from the analysis due to incomplete information on the explanatory and dependent variables. Finally, 37 mothers who identified their race as something other than White, Black, or Hispanic of Mexican descent were excluded from the analysis. The final sample included 1197 mothers, who had complete information on all variables across four waves. Approximately 55% of the final sample of mothers were Black (n = 659), 22% were White (n = 265), and 23% were Hispanic (n = 273). The mean age of the sample was 25. About 32% of the sample mothers did not have a high school diploma, 32% had graduated high school, and 36% had more than a high school education. At Year 1, almost 69% of the mothers were unmarried, and 43% had an income below the poverty line. About half of the children in the sample were boys (51%). Demographic differences were explored between mothers who did (n = 180) and did not report experiencing (n = 1017) any type of domestic violence at the Year 1 survey. Compared to mothers who did not report domestic violence, mothers who reported domestic violence were more likely to be unmarried and to be in poverty at Year 1. No significant differences were found for other background characteristics. To examine the influence of sample attrition on the findings, baseline characteristics of the final sample and the dropped cases were also examined. The results revealed no significant differences between the final sample and the dropped cases for most of the background characteristics, such as age, education, income, and marital status. However, the dropped cases was less likely to be Black and more likely to be Hispanic than the final cases.
92
J.A. Yoo / Children and Youth Services Review 44 (2014) 90–99
2.2. Measures 2.2.1. Children's behavior problems Children's behavior problems at age 5 were assessed with a partial scale of the Child Behavior Checklist for ages 4 to 18 (CBCL/4–8; Achenbach, 1991). The CBCL is a commonly used measure for assessing problematic behaviors. Two subscales, externalizing and internalizing behaviors, were used as the outcome variables for this study. Externalizing behavior problems were measured by aggressiveness (20 items) and delinquent behaviors (10 items). Internalizing behavior problems were measured by anxious and depressed behaviors (14 items) and withdrawal (9 items). The mothers were asked to rate several statements for how true they were (not true = 0, somewhat true = 1, very true or often true = 2). Each item was scored such that higher scores indicate more frequent or severe behavior problems. Total scores for the externalizing and internalizing subscales were calculated by summing the scores for all items. 2.2.2. Maternal mental health Maternal mental health at Year 3 was assessed by the mother's experience of a depressive episode (7 items) and generalized anxiety disorder (7 items). Both conditions are based on the Composite International Diagnostic Interview, Short Form (Walters, Kessler, Nelson, & Mroczek, 2002). For both depression and anxiety, the mothers who experienced three or more items were classified as having symptoms. The maternal mental health measure was coded yes if the mother experienced depressive symptoms or generalized anxiety disorder and no if she had never experienced any of the symptoms (1 = yes, 0 = no). 2.2.3. Parenting Parenting at Year 3 was assessed by two subdimensions: the interviewer assessment of parenting behavior and the mother's report of using spanking as a disciplinary tactic. The interviewer's assessment of parenting behavior was measured using subscales of the Home Observation for Measurement of the Environment (HOME; Bradley, 1993). The HOME instrument measures if a mother was unresponsive or harsh or exhibited a lack of verbal or social skills to the child during the home visit. The HOME instrument has been used widely to measure parenting and the home environment. The instrument has been shown to have few differences in psychometric properties across racial groups (Pachter, Auinger, Palmer, & Weitzman, 2006). One point was assigned for each negative response, which means that higher scores indicated worse parenting. Spanking was measured by the frequency with which the mother spanked the child when the child misbehaved or acted up during the month prior to the interview. Responses were coded into a 5-level variable: 0 (never), 1 (only once or twice), 2 (a few times in the past month), 3 (a few times a week), and 4 (nearly every day). 2.2.4. Domestic violence The primary independent variable of this study was domestic violence toward the mother at Year 1. Domestic violence was considered to be present if a mother reported that she had experienced any amount of physical, emotional, or sexual abuse inflicted by the child's biological father. The items “he slapped or kicked you” and “he hit you with his fist or a dangerous object” were used to measure physical violence. Items on emotional violence included (a) “he tried to isolate you from family and friends,” (b) “he tried to prevent you from going to work and/or school,” and (c) “he withheld money, made you ask for money, or took your money.” Sexual abuse was measured with the item, “he tried to make you have sex or do sexual things you didn't want to do.” There were three possible responses for each item: never, sometimes, and often. However, due to the smaller number of mothers who reported often, mothers who reported sometimes and often were combined into one category, yes. Domestic violence was coded as yes if a mother had
experienced any of the above items and as no if a mother reported never on all six items (1 = yes, 0 = no). 2.2.5. Race Racial identity was self-reported by the mother at the child's birth. Three racial groups were used in this study: non-Hispanic White (hereinafter White), non-Hispanic Black (hereinafter Black), and Hispanic mothers. The Fragile Families data allow Hispanic mothers to be disaggregated into subgroups of national origin (Mexican, Puerto Rican, South American, etc.), but the sample sizes were not sufficient for separate analyses. Although Hispanic women have diverse cultural backgrounds, they do share common influences (i.e., sexism or racism) when we consider domestic violence within a sociopolitical context (Kasturinrangan et al., 2004). 2.2.6. Controls The study controlled for socioeconomic status and child effects. Socioeconomic status included marital status, poverty, and educational attainment at Year 1. Marital status was treated as a dichotomous variable and was coded 1 if the mother was unmarried. Unmarried mothers in this study included women who were divorced, widowed, or never married. Poverty was treated as a dichotomous variable and was coded 1 if the family's income-to-needs ratio was less than 1. The family income-to-needs ratio was calculated by dividing the total family income by the official poverty threshold and adjusting the result for household size. Educational attainment was coded as less than high school, high school diploma, and more than high school and college degree. Child effects include child's gender and temperament at Year 1. Child temperament was treated as a continuous variable based on a subscale taken from the Emotionality, Activity, and Sociability Temperament Survey for Children (Buss & Plomin, 1984). Child temperament was measured by the average score of six items that asked mothers to rate the difficulty of the focal child's temperament. Higher scores indicated a more difficult temperament. Child's gender was coded 1 if the focal child was a boy. 2.3. Analysis The hypothesized path model that depicts the relationships between the main variables is presented in Fig. 1. This model hypothesizes both direct and indirect effects of domestic violence on children's behavior problems through maternal mental health and parenting. The model specified that domestic violence at Year 1 was related to maternal mental health, spanking, and parenting behavior at Year 3; that maternal mental health at Year 3, in turn, was associated with both spanking and parenting behavior at Year 3; and that both spanking and parenting behavior at Year 3 were related to children's behavior problems at Year 5. The direct effect of domestic violence at Year 1 on children's behavior problems at Year 5 was also specified. Control variables were included in the model but are not shown in the figure. Structural equation modeling with AMOS was used to examine whether the impact of domestic violence on women and children varied by race. Structural equation modeling (SEM) is a technique by which relationships between independent and dependent variables can be analyzed. SEM differs from regression techniques in that SEM allows one not only to assess direct effects but also to simultaneously analyze indirect effects through mediating variables (Pachter et al., 2006; Yoo & Huang, 2012). To determine the differences in the relationships between the variables across racial groups sampled, this study conducted separate SEM analyses for the White, Black, and Hispanic samples. And separate SEM analyses were conducted for externalizing and internalizing behavior problems by White, Black, and Hispanic groups. A total of six SEM models were analyzed. Model fit tests were conducted to determine whether the model depicted in Fig. 1 fit the observed data. A traditional chi-square test of model fit was used, which indicates adequate fit if chi-square is small
J.A. Yoo / Children and Youth Services Review 44 (2014) 90–99
Domestic violence at Yr 1
Spanking at Yr 3
Mental Health at Yr 3
93
Child behavior Problems at Yr 5
Parenting at Yr 3
Fig. 1. A hypothesized path model.
and insignificant. Several other model fit indices were also employed. The root mean square error of approximation (RMSEA) was used (values less than 0.08 indicate appropriate fit). The goodness of fit index (GFI) and the comparative fit index (CFI) were also used (good fit is indicated by values over 0.9).
3. Results 3.1. Descriptive analysis Data were analyzed for 265 White, 659 Black, and 273 Hispanic families. Table 1 displays the descriptive characteristics of these families. The results showed no large racial differences in the prevalence of domestic violence at Year 1, maternal mental health at Year 3, or children's behavior problems at Year 5. About 16% of Hispanic mothers reported that they experienced domestic violence during their child's first year of life. Almost 15% of Black and White mothers reported experiencing domestic violence, respectively. For maternal mental health, about 17% of Black and White mothers reported that they experienced mental health problems at Year 3, whereas approximately 12% of Hispanic mothers reported mental health problems at this time. Hispanic mothers rated their child's internalizing behavior problems at Year 5 slightly higher than White and Black mothers rated their children's, but there were no differences in the level of children's externalizing behavior problems across racial groups. However, maternal parenting at Year 3 differed considerably across the three racial groups. Regarding parenting behavior, Black mothers reported higher levels of negative parenting behaviors than White mothers, and Hispanic mothers fell in between the two groups. For spanking, White mothers reported more spanking than Hispanic mothers, and Black mothers were in the middle.
Most differences between the racial categories were found for socioeconomic characteristics, such as marital status, poverty, and education. The results in Table 1 show substantial variation in marital status across race. Approximately 40% of White mothers and 64% of Hispanic mothers were not married to their child's biological father at Year 1. In contrast, approximately 82% of Black mothers were unmarried at that time. Poverty also varied significantly by race. Whereas about 17% of White mothers had an income below the poverty line at the child's birth, 52% of Black mothers and 45% of Hispanic mothers reported a povertylevel income. Educational attainment also differed considerably by race. Almost 60% of White mothers had more than a high school education, compared to 30% of Black and 27% of Hispanic mothers. Approximately 33% of Black mothers and 46% of Hispanic mothers did not have a high school diploma. Finally, child temperament at Year 1 varied significantly by race. Black mothers reported child temperament problems more than White mothers did, and Hispanic mothers fell in between the two groups. No differences were found by child's gender.
3.2. Path models Results from the structural equation models of the effects of domestic violence on children's internalizing and externalizing behavior problems in White, Black, and Hispanic groups are displayed in Figs. 2 and 3. For all models, model-fit tests showed that a chi-square test for the model was nonsignificant, indicating that the model was compatible with the observed data. Likewise, the other model fit indices provided adequate fit to the data. The specific model fit for each model is shown in the figure. Paths with standardized coefficients that were significant at p ≤ .05 are highlighted by bold lines in the figure. This study controlled socioeconomic status and preexisting child effects in the
Table 1 Descriptive statistics for the variables by race.
Main variables Domestic violence at Year 1 Maternal mental health at Year 3 Parenting behavior at Year 3 (0–14) Spanking at Year 3 (0–4) Child externalizing problems at Year 5 (0–43) Child internalizing problems at Year 5 (0–26) Controls Unmarried at Year 1 Poverty at Year 1 Education at Year 1 (1–3) Child temperament at Year 1 (1–5) Boy
χ2 or F-test
White (N = 265)
Black (N = 659)
Hispanic (N = 273)
N (%)/M (SD)
N (%)/M (SD)
N (%)/M (SD)
39 (14.72) 44 (16.60) 0.61 (1.31) 0.98 (1.14) 12.29 (7.48) 17.40 (10.28)
97 (14.72) 114 (17.30) 2.02 (2.56) 1.07 (1.11) 13.10 (7.85) 18.27 (10.62)
44 (16.12) 32 (11.72) 1.13 (1.89) 0.77 (0.99) 12.87 (7.84) 19.36 (11.51)
0.78 4.63 44.42⁎⁎⁎ 7.38⁎⁎⁎
106 (40.0) 44 (16.6) 2.46 (0.72) 2.43 (0.7) 133 (50.19)
543 (82.4) 344 (52.20) 1.97 (0.79) 2.64 (0.77) 340 (51.59)
176 (64.47) 123 (45.05) 1.81 (0.83) 2.51 (0.7) 142 (52.01)
161.88⁎⁎⁎ 98.69⁎⁎⁎ 50.44⁎⁎⁎ 8.85⁎⁎⁎
Note. N (%) for domestic violence, maternal mental health, unmarried, poverty, and boy. Mean (SD) for parenting behavior, spanking, child externalizing and internalizing behavior problems, education, and child temperament. N = 1197. ⁎⁎⁎ p b .001.
1.03 2.26
0.21
94
J.A. Yoo / Children and Youth Services Review 44 (2014) 90–99
White 0.05 Domestic violence
0.16*
Child Externalizing Problems
0.21** Spanking 0.05
0.05
0.19**
0.15*
0.10 Mental Health
Parenting
−0.00
χ2 = 17.52 (df = 16, p > .05); CFI = 0.99; GFI = 0.98; RMSEA = 0.02
Black 0.06 Domestic violence
Child Externalizing Problems
0.18**
0.05 Spanking 0.09*
0.07
0.08*
0.11*
0.11** Mental Health
0.06
Parenting
χ2 = 19.06 (df = 16, p > .05); CFI = 0.99; GFI = 0.99; RMSEA = 0.02
Hispanic 0.14* Domestic violence
0.21**
Child Externalizing Problems
0.12 Spanking
0.02
0.07 0.01
0.13*
0.09 Mental Health
−0.01
Parenting
χ2 = 19.59 (df = 16, p > .05); CFI = 0.96; GFI = 0.97; RMSEA = 0.03
Fig. 2. Structural equation modeling of domestic violence on children's externalizing behavior problems in White, Black, and Hispanic groups. *p b .05. **p b .01. ***p b .001.
analyses; however, to maintain a clear focus on the effects of domestic violence, the paths between the control variables and main variables are not presented in the figures. 3.2.1. Children's externalizing behavior problems by race
control variables and children's externalizing problems. These results indicate that maternal mental health and parenting at Year 3 mediate the effects of domestic violence at Year 1 on children's outcomes at Year 5, when socioeconomic and preexisting child factors were taken into account.
3.2.1.1. White families. As shown in Fig. 2, domestic violence at Year 1 had direct associations with poor maternal mental health (β = 0.15, p b .05) and more frequent use of spanking (β = 0.16, p b .05) at Year 3. In turn, poor maternal mental health (β = 0.19 p b .01) and more frequent use of spanking (β = 0.21, p b .01) were both directly related to greater externalizing behavior problems of children at Year 5. No direct associations were found between domestic violence at Year 1 and children's externalizing problems at Year 5. Regarding the effects of control variables, unmarried-mother status (β = −0.18, p b .01) and higher levels of child temperament (β = 0.14, p b .01) at Year 1 are strongly related to greater externalizing problems of White children at Year 5. Table 2 presents the indirect and total effects of domestic violence on children's externalizing problems. The indirect effect of domestic violence at Year 1 on children's externalizing problems at Year 5 was significant (β = 0.07, p b .01). No indirect effects were found between the
3.2.1.2. Black families. Domestic violence at Year 1 had a significant, direct relationship with poor maternal mental health at Year 3 (β = 0.11, p b .05). In turn, poor maternal mental health at Year 3 was associated with increased children's externalizing problems at Year 5 (β = 0.08, p b .05), as well as with greater use of spanking at Year 3 (β = 0.09, p b .05). Poor parenting at Year 3 – both spanking (β = 0.18, p b .01) and parenting behavior (β = 0.11, p b .01) – was related to increased children's externalizing problems at Year 5. No direct link was found between domestic violence and children's externalizing problems. With respect to the control variables, unmarried-mother status (β = − 0.09, p b .05), male gender (β = 0.09, p b .05), and higher level of child temperament (β = 0.14, p b .01) at Year 1 had direct relationships with greater externalizing problems of Black children at Year 5. As presented in Table 2, the indirect effect of domestic violence at Year 1 on children's externalizing problems at Year 5 was significant
J.A. Yoo / Children and Youth Services Review 44 (2014) 90–99
95
White 0.03 Domestic violence
0.16*
Child Internalizing Problems
0.13* Spanking 0.05
0.05
0.15*
0.15*
0.10 Mental Health
Parenting
−0.00
χ2 = 17.52 (df = 16, p > .05); CFI=0.99; GFI=0.99; RMSEA=0.02
Black 0.11** Domestic violence
Child Internalizing Problems
0.10*
0.05 Spanking 0.09*
0.07
0.05
0.11*
0.12** Mental Health
Parenting
χ2 = 19.06 (df = 16, p > .05); CFI=0.99; GFI=0.99; RMSEA=0.02
0.06
Hispanic 0.21** Domestic violence
0.21**
Child Internalizing Problems
0.02 Spanking
0.02
0.07 0.05
0.13*
0.15* Mental Health
−0.01
Parenting
χ2 = 19.59 (df = 16, p > .05); CFI=0.97; GFI=0.99; RMSEA=0.03
Fig. 3. Structural equation modeling of domestic violence on children's internalizing behavior problems in White, Black, and Hispanic groups. *p b .05. **p b .01. ***p b .001.
(β = 0.03, p b .01). Interestingly, the indirect effect of maternal mental health at Year 3 on children's externalizing problems at Year 5 was also significant in this group (β = 0.02, p b .01). No indirect effect was found between control variables and children's externalizing problems. These results suggest that domestic violence at Year 1 had a negative effect on maternal mental health at Year 3 and that mental health in turn tended to increase children's externalizing problems at Year 5 directly as well as indirectly through its influence on spanking at Year 3, even when socioeconomic and preexisting child factors were controlled in the model. 3.2.1.3. Hispanic families. A quite different picture emerged when predicting externalizing problems of the Hispanic children in the sample. For this group, domestic violence at Year 1 had direct associations with poor maternal mental health (β = 0.13, p b .05) and more frequent use of spanking at Year 3 (β = 0.21, p b .01). However, maternal mental health and parenting at Year 3 were not subsequently related to children's externalizing problems at Year 5. Notably, domestic violence at Year 1 had a strong direct effect on children's externalizing problems at Year 5, even after controlling mediators (β = 0.14, p b .05). No significant effects were found between control variables and children's externalizing problems. Whereas the indirect effect of domestic violence on
children's externalizing problems was not significant, its total effect on children was (β = 0.16, p b .01). These results indicate that domestic violence at Year 1 had a strong, direct effect on children's externalizing problems at Year 5, even after controlling mediators and other variables, and that this effect was not mediated through maternal mental health or parenting at Year 3. 3.2.2. Children's internalizing behavior problems by race 3.2.2.1. White families. The direct links among variables found in the model predicting internalizing problems for White children were similar to those found for this group's externalizing problems. As shown in Fig. 3, domestic violence at Year 1 had direct associations with poor maternal mental health (β = 0.15, p b .05) and more frequent use of spanking (β = 0.16, p b .05) at Year 3, both of which, in turn, were related to greater internalizing problems of children at Year 5 (β = 0.15, p b .05; β = 0.13, p b .05). Although the indirect effect of domestic violence at Year 1 for internalizing problems at Year 5 was also significant (β = 0.05, p b .01; Table 3), its total effect on children's internalizing problems was not. These results suggest that the power that domestic violence has on women and children was weaker for
−0.01 0.07 0.02 0.09 0.12 0.02⁎⁎
0.06 0.09⁎ 0.10⁎ 0.11⁎⁎ 0.18⁎⁎
−0.01 0.07 0.01 0.09 0.12
0.01
−0.01 0.01 0.03 0.01 0.01⁎ 0.03⁎⁎
0.13⁎ 0.02 0.21⁎⁎ 0.14⁎
3.2.2.3. Hispanic families. The results for Hispanic children's internalizing problems were very similar to the results for this group's externalizing problems. That is, domestic violence at Year 1 had a direct relationship with poor maternal mental health (β = 0.13, p b .05) and more frequent use of spanking at Year 3 (β = 0.21, p b .01), but neither outcome was subsequently related to children's internalizing problems at Year 5. Domestic violence had a strong effect on children's internalizing problems (β = 0.21, p b .01), even after controlling mediators and other variables in the analyses. The indirect effect of domestic violence at Year 1 for children's internalizing problems at Year 5 was not significant; however, its total effect on children's internalizing problems (β = 0.22, p b .01) was far larger for Hispanic children than for the White and Black groups. No direct or indirect effects were found between the control variables and children's internalizing problems. The results for Hispanic families indicate that domestic violence at Year 1 had a strong, direct effect on children's internalizing problems at Year 5, but this effect was not mediated through maternal mental health or parenting at Year 3.
0.15⁎ 0.05 0.17⁎ 0.12⁎
0.01
0.05 −0.00 0.05 0.19⁎⁎ 0.10 0.21⁎⁎
0.15⁎ 0.05 0.16⁎
0.00 0.01 0.07⁎⁎
Indirect effect
−0.00 0.05 0.20⁎⁎ 0.10 0.21⁎⁎
Direct effect
3.2.2.2. Black families. The specific pathways through which domestic violence affected Black children's internalizing problems differed from the pathways found in this group's externalizing problems. Domestic violence at Year 1 was directly associated with poor maternal mental health at Year 3 (β = 0.11, p b .05), and poor maternal mental health at Year 3 was related to more frequent use of spanking at Year 3 (β = 0.09, p b .05), but not to children's problems at Year 5. Poor parenting at Year 3 – both frequent use of spanking (β = 0.10, p b .05) and negative parenting behavior (β = 0.12, p b .01) – was associated with increased externalizing problems of children at Year 5. The direct effect of domestic violence at Year 1 for children's internalizing problems at Year 5 remained significant (β = 0.11, p b .05) even after controlling mediators. Regarding the effects of the control variables, only educational attainment had a direct relationship with internalizing problems of Black children (β = −0.11, p b .05). The indirect effect of domestic violence at Year 1 on children's internalizing problems at Year 5 was significant (β = 0.02, p b .01). The indirect effect of maternal mental health at Year 3 on children's internalizing problems at Year 5 via spanking at Year 3 was also significant (β = 0.02, p b .01). These results show that, even when socioeconomic and child factors were controlled in the model, domestic violence had a negative effect on maternal mental health and that mental health tended to increase children's internalizing problems primarily through its indirect effect on spanking.
0.11⁎ 0.07 0.05 0.06 0.06 0.09⁎ 0.08⁎ 0.11⁎⁎ 0.18⁎⁎
0.11⁎ 0.08 0.06 0.09⁎
internalizing than for externalizing problems. As for the effects of the control variables, only child temperament had a direct relationship with children's internalizing problems for this racial group (β = 0.18, p b .01).
Note. ⁎ p b .05. ⁎⁎ p b .01.
Parenting Spanking
Maternal mental health Parenting behavior Spanking Externalizing problems Parenting behavior Spanking Externalizing problems Externalizing problems Externalizing problems Domestic violence
Maternal mental health
Dependent variables
4. Discussion
Predictor
Table 2 Indirect and total effects of domestic violence on children's externalizing behavior problems.
Total effects
Total effects Direct effect Direct effect
Indirect effect
Hispanic Black White
Indirect effect
0.13⁎ 0.02 0.22⁎⁎ 0.17⁎⁎
J.A. Yoo / Children and Youth Services Review 44 (2014) 90–99 Total effects
96
The present study examined whether there were racial differences in the effects of domestic violence on children's externalizing and internalizing behavior problems when socioeconomic factors and preexisting child characteristics were controlled for. This longitudinal study provides the first empirical evidence of racial variations in the mechanisms through which domestic violence affects children's behavior problems. The results from this study indicate that domestic violence had negative effects on children in all racial groups, but the mechanisms through which domestic violence exerted its effects on children differed among White, Black, and Hispanic families. In particular, whereas the effects of domestic violence on children were mediated through maternal mental health and parenting for White and Black families, only direct effects were found for Hispanic families. These findings suggest that the conceptual framework that links domestic violence and children may not apply to all racial groups and that cultural contexts may contribute to children's and mothers' experiences of abuse in violent families.
0.02
0.00 0.07 0.12⁎⁎ 0.10⁎
0.02⁎⁎ 0.05 0.12⁎⁎ 0.10⁎ Note. ⁎ p b .05. ⁎⁎ p b .01.
0.10 0.13⁎ Parenting Spanking
Maternal mental health
Maternal mental health Parenting behavior Spanking Internalizing problems Parenting behavior Spanking Internalizing problems Internalizing problems Internalizing problems Domestic violence
0.15⁎ 0.05 0.16⁎ 0.03 −0.00 0.05 0.15⁎
0.01
0.00 0.01 0.05⁎⁎
0.10 0.13⁎
0.01 0.01⁎ 0.02⁎⁎
0.11⁎ 0.07 0.05 0.11⁎⁎ 0.06 0.09⁎
0.15⁎ 0.05 0.17⁎ 0.08 −0.00 0.05 0.15⁎
Indirect effect
0.02
0.13⁎ 0.02 0.22⁎⁎ 0.22⁎⁎ −0.01 0.07 0.05 0.15⁎ −0.00 0.01 0.01
0.13⁎ 0.02 0.21⁎⁎ 0.21⁎⁎ −0.01 0.07 0.05 0.15⁎ 0.11⁎ 0.08 0.06 0.13⁎⁎ 0.06 0.09⁎
Hispanic Black
Direct effect Direct effect
Dependent variables Predictor
Table 3 Indirect and total effects of domestic violence on children's internalizing behavior problems.
Total effects White
Indirect effect
Total effects
Direct effect
Indirect effect
Total effects
J.A. Yoo / Children and Youth Services Review 44 (2014) 90–99
97
The results for White families in this study showed that the effects of domestic violence on children, especially for externalizing behavior problems, were fully mediated through maternal mental health and parenting, which coincides well with an ecological framework. The results indicate that domestic violence at Year 1 had negative effects on mothers' mental health and parenting (particularly spanking) at Year 3, both of which in turn tended to increase problematic behaviors of children at Year 5 for White families. The results could be taken as evidence that abuse-induced disruptions in maternal mental health and parenting could be key mechanisms linking domestic violence and children for White families. A different picture emerged when analyzing the effects of domestic violence on Black children. For Black families, the specific pathways through which domestic violence affected children differed by the type of behavior problem. Whereas the effects of domestic violence on externalizing problems were fully mediated through maternal mental health, the effects on internalizing problems were direct and were partially mediated through maternal mental health. Although the specific link between domestic violence and children somewhat differed by the type of behavior problem, the overall results for Black families suggested that the mediation effects occurred primarily through maternal mental health. That is, abused women in these Black families frequently suffered from mental health problems, such as depression and anxiety, and their mental health problems were found to be key mechanisms through which domestic violence led to poor parenting (particularly spanking) and thereby poor behavioral outcomes of children. The results clearly showed that Black mothers whose children's behaviors were compromised were struggling with mental health problems. However, the emotional pain caused by domestic violence for Black women may be underestimated in the literature (Campbell et al., 2008) because such difficulties are frequently confounded with several other stressors common for this group, such as lack of resources or social stress. The findings in this study have important implications for social service efforts that support women and children in violent families; particularly, such efforts should address Black victims' need for emotional support to improve their parenting functioning and to reduce their children's behavior problems. Unlike the results for White families, however, mediation effects on parenting were not supported in the results for Black families. In particular, domestic violence at Year 1 did not have direct effects on parenting at Year 3, although parenting was a strong predictor of children's behavior problems at Year 5. It is not clear why Black mothers' parenting was not directly affected by domestic violence. A possible explanation is that the weak association between domestic violence and parenting may have something to do with the cultural validity of parenting measures. Parenting is influenced largely by culture. For instance, spanking is commonly considered to be an acceptable form of disciplinary behavior in Black cultures (Whaley, 2000) and thus a Black mother's use of spanking might actually not be strongly affected by domestic violence. Alternatively, given that Black women experience multiple institutional prejudices, abused Black mothers may be less likely to self-report spanking or negative parenting behaviors (Cardoso, Padilla, & Sampson, 2010). Future studies should include multiple measures of parenting that are culturally sensitive for all races and more thoroughly investigate the mediation effects of parenting in Black families. Perhaps the most interesting and unexpected findings of the present study was that there were no mediation effects of maternal mental health and parenting in the link between domestic violence and children in the Hispanic group. Among the Hispanic families, domestic violence had strong effects on both the externalizing and internalizing behaviors of children; however, contrary to the findings for White and Black families, these effects were not mediated through maternal mental health or parenting. Abused Hispanic mothers in this study exhibited mental health problems and poor parenting (especially spanking), but their mental health and parenting difficulties were not subsequently
98
J.A. Yoo / Children and Youth Services Review 44 (2014) 90–99
related to increased child behavior problems. It is unclear why Hispanic children were not influenced by maternal mental health and parenting difficulties caused by abuse. Prior research has not provided specific information about how Hispanic children respond to or cope with domestic violence and its related maternal dysfunction. As a potential explanation, the unexpected results for Hispanic children might be associated with length of exposure to domestic violence. One of the limitations in this study was that this study did not measure whether the domestic violence was a onetime event that happened at Year 1 or whether it continued in subsequent waves. Prior studies suggested that Hispanic women may stay in violent relationships longer to keep their family intact (Edelson et al., 2007). Married Hispanic mothers in this study who experienced domestic violence at Year 1 were more likely to still be living with the child's father in subsequent waves, and in these cases it appeared that domestic violence was ongoing. The Hispanic children who were exposed to ongoing violence may have been more strongly affected by their father's violent behaviors than by their mother's responses to the violence. Another possible explanation is that Hispanic children might be better helped through a continuum of naturally occurring informal support within the family or their community. Literature suggests that many Hispanic women and children tend to live in extended families that may provide social support, such as emotional support, child care, and protection for abuse victims and children (Kasturinrangan et al., 2004; Sokoloff & Dupont, 2005). Such support might serve as protective factors for Hispanic children and make it easier for them to cope with maternal stress related to domestic violence. Still, more research is needed to investigate what cultural aspects within the Hispanic community contribute to reducing the effects of abuseinduced disruptions in maternal functioning on children's behavior problems. The results from this study must be interpreted in the context of the following limitations. First, the substantial loss of data due to sample attrition across six surveys, as well as incomplete data for the variables used in this study, might have influenced the results. Second, this study did not examine how domestic violence at Year 1 interrelates over time with other risk factors (e.g., racism) and protective factors (e.g., religion or family) that might intervene in any direct relationship between domestic violence at Year 1 and child outcomes at Year 5. For example, domestic violence and its interaction with religion or social support in minority communities might be associated with differences in groups in this study. Future research is necessary to include multiple risk and protective factors in the analyses and to assess how those factors play a role in the effects of domestic violence on children. Third, the measures of domestic violence, maternal mental health, and children's behavior problems relied solely on maternal reports, which might have led to common single-reporter bias. Fourth, this study only included Black, White, and Hispanic families because of insufficient sample sizes for other racial groups. Some research has suggested that Native American and Asian women are at a greater risk of domestic violence and have specific cultural barriers and strengths related to abuse (Grossman & Lundy, 2007). Future research is needed to include more diverse racial categories and examine any cultural-specific mechanisms that might be present within these groups. Finally, this study did not consider within-group differences. Minority communities tend to be heterogeneous in terms of cultural backgrounds. For example, Hispanic refers to individuals of several national origins such as Mexican, Puerto Rican, Cuban, and Central American. Even though these groups share some cultural similarities and the same language, these subgroups also differ in their country of origin, history of migration, acculturation level, and other factors that may influence the experiences of abused women and their children. In addition, there may be differences between long-standing U.S. families and newly arrived immigrant families within minority communities. Newly arrived immigrant women tend to experience limited English skills, social isolation, or lack of information of social services, which makes it difficult for them to cope with
domestic violence and the related stress (Kasturinrangan et al., 2004). Minority groups also differ in their religious beliefs which might influence the experience of abused women and their children (e.g., Catholic and Protestant groups within the Black families). Given the results that the effects of domestic violence on children varied by race, future research is required to investigate the within-group variability in the experiences of battered women and children. Despite these limitations, this study revealed that children's and mothers' experiences of domestic violence varied by race and that cultural contexts may contribute to the dynamics of children's and mother's experiences. Although more research is necessary to fully understand such racial variations, it is evident that social service professionals need to be culturally sensitive when working with abused women and their children from diverse racial groups. Social service providers must understand the racial and cultural differences in mothers' and children's experiences of domestic violence, as well as the particular needs that different groups have (Sokoloff & Dupont, 2005), without perpetuating negative stereotypes for racial groups. This study found that White children were more affected by maternal dysfunction caused by domestic violence than were their minority peers. Emotional support and parenting intervention programs for abused mothers can be effective in preventing or minimizing the negative effects of domestic violence on White children. For Black families, this study found that domestic violence is a psychologically stressful event that leads to maternal depression and anxiety. Thus, social service providers should first conduct open-minded assessments of women's need for emotional support when working with Black families. To address the needs of Black women's mental health problems, social service providers should understand their community or cultural values and be aware of cultural strengths (e.g., religion) or barriers to help-seeking (e.g., community pressures or racism) in the Black community. They should also attempt to involve community members who share the backgrounds of the victims, and develop culturally congruent forms of emotional support and resources. These efforts may have a positive effect on reducing mental health problems of abused Black women and thereby minimizing their children's behavior problems. For Hispanic families, this study revealed that children can be strongly affected by the presence of domestic violence within the family without being additionally affected by maternal functioning. These findings suggest that emotional support services or parenting intervention programs developed for White families might be not effective in reducing behavior problems of Hispanic children. Rather, when working with Hispanic families, social service providers should pay more attention to direct links between the father's violent behaviors and children's outcomes. This study also suggests that Hispanic children might be culturally protected from the effects of maternal dysfunction caused by domestic violence. Thus, social service providers should assess support systems of victims and children within the family or community and attempt to collaborate with the support groups to develop culturally appropriate services. These efforts should help abused victims and their children cope with domestic violence and the related challenges more effectively. Finally, it is important that social service providers should understand that many minority families experience other social stressors (e.g., lack of resources or racism) which may affect their lives as well as their ability to access resources (Campbell et al., 2008; Kasturinrangan et al., 2004). Thus, social service providers should note that services or interventions for supporting minority families can be more effective when their efforts are combined with other physical and environmental services. Acknowledgments This work was supported by the Brain Korea 21 Plus Project (Ewha Womans University, School of Social Welfare) of the National Research Foundation of Korea Grant (21B20130012299).
J.A. Yoo / Children and Youth Services Review 44 (2014) 90–99
References Achenbach, T. M. (1991). Manual for the child behavior checklist/4–18 and 1991 profile. Burlington, VT: University of Vermont Department of Psychiatry. Belsky, J. (1984). The determinants of parenting: A process model. Child Development, 55, 83–96. Bradley, R. H. (1993). Children's home environments, health, behavior, and intervention efforts: A review using the HOME inventory as a marker measure. Genetic, Social, and General Psychology Monographs, 119, 437–490. Bronfenbrenner, U. (1979). The ecology of human development: Experiments by nature and design. Cambridge, MA: Harvard University Press. Buss, A. H., & Plomin, R. (1984). Temperament: Early developing personality traits. Hillsdale, NJ: Erlbaum. Campbell, J., Campbell, D. W., Gary, F., Nedd, D., Price-Lea, P., Sharps, P. W., et al. (2008). African American women's responses to intimate partner violence: An examination of cultural context. Journal of Aggression, Maltreatment & Trauma, 16(3), 277–295. Cardoso, J. B., Padilla, Y. C., & Sampson, M. (2010). Racial and ethnic variation in the predictors of maternal parenting stress. Journal of Social Service Research, 36, 429–444 (f). Edelson, M. G., Hokoda, A., & Romos-Lira, L. (2007). Differences in effects of domestic violence between Latina and non-Latina women. Journal of Family Violence, 22, 1010. Grossman, S. F., & Lundy, M. (2007). Domestic violence across race and ethnicity: Implications for social work practice and policy. Violence Against Women, 13(10), 1029–1052. Huang, C., Wang, L. R., & Warrener, C. (2010). Effects of domestic violence on behavior problems of preschool-aged children: Do maternal mental health and parenting mediate the effects? Children and Youth Services Review, 32, 1317–1323. Kasturinrangan, A., Krishnan, S., & Riger, S. (2004). The impact of culture and minority status on women's experience of domestic violence. Trauma, Violence & Abuse, 5(4), 318–332. Levendosky, A. A., Huth-Bocks, A.C., Shapiro, D. L., & Semel, M.A. (2003). The impact of domestic violence on the maternal–child relationship and preschool-age children's functioning. Journal of Family Psychology, 17(3), 275–287.
99
Levendosky, A. A., Leahy, K. L., Bogat, G. A., Davidson, W. S., & von Eye, A. (2006). Domestic violence, maternal parenting, maternal mental health & infant externalizing behavior. Journal of Family Psychology, 20(4), 544–552. Pachter, L. M., Auinger, P., Palmer, R., & Weitzman, M. (2006). Do parenting and the home environment, maternal depression, neighborhood, and chronic poverty affect child behavioral problems differently in different racial–ethnic groups? Pediatrics, 117(4), 1329–1338. Reichmann, N. E., Teitler, J. O., Garfinkel, I., & McLanahan, S. S. (2001). Fragile families: Sample and design. Children and Youth Services Review, 23(4–5), 303–326. Schoppe-Sullivan, S. J. (2007). Marital conflict and children's adjustment: Evaluation of the parenting process model. Journal of Marriage and Family, 69, 1118–1134. Sokoloff, N. J., & Dupont, I. (2005). Domestic violence at the intersections of race, class, and gender. Violence Against Women, 11(1), 38–64. Taylor, C. A., Guterman, N.B., Lee, S. J., & Rathouz, P. (2009). Intimate partner violence, maternal stress, nativity, and risk for maternal maltreatment of young children. American Journal of Public Health, 99(1), 175–183. Torres, S., & Han, H. -R. (2000). Psychological distress in non-Hispanic White and Hispanic abused women. Archives of Psychiatric Nursing, XIV(1), 19–29. Walters, E. E., Kessler, R. C., Nelson, C. B., & Mroczek, D. (2002). Scoring the World Health Organization's Composite International Diagnostic Interview Short Form (CIDI-SF). http://www.who.int/msa/cidi/CIDISFScoringMemo12-03-02.pdf Whaley, A. L. (2000). Sociocultural differences in the developmental consequences of the use of physical discipline during childhood for African Americans. Cultural Diversity and Ethnic Minority Psychology, 6(1), 5–12. Yoo, J. A., & Huang, C. -C. (2012). The effects of domestic violence on children's behavior problems: Assessing the moderating roles of poverty and marital status. Children and Youth Services Review, 34(12), 2464–2473. Yoo, J. A., & Huang, C. -C. (2013). Long term relationships among domestic violence, maternal mental health and parenting, and preschool children's behavior problems. Families in Society, 94(4), 268–276.
Contents lists available at ScienceDirect
Children and Youth Services Review journal homepage: www.elsevier.com/locate/childyouth
Racial variations in the link between domestic violence and children's behavioral outcomes Jeong Ah Yoo ⁎ School of Social Welfare, Ewha Womans University, Ewhayeodae-gil, Seodaemun-gu, Seoul 120-750, South Korea
a r t i c l e
i n f o
Article history: Received 25 October 2013 Received in revised form 9 May 2014 Accepted 1 June 2014 Available online 11 June 2014 Keywords: Domestic violence Children Racial variation Behavior problems
a b s t r a c t The link between domestic violence and children has been well documented, but little is known about racial variations in this relationship. Using multiple waves of the Fragile Families and Child Wellbeing Study, this study examined whether the effects of domestic violence on children's behavior problems differed by race. Results from structural equation modeling revealed that domestic violence had negative effects on children in all racial groups; but the mechanisms through which domestic violence exerted its effects varied across groups. Whereas the effects of domestic violence on children's behavior problems were partially mediated through maternal functioning for White and Black families, only direct effects were found for Hispanic families. This study highlights wide racial variations in the mechanisms through which domestic violence affects children and the importance of cultural context in understanding those variations. © 2014 Published by Elsevier Ltd.
1. Introduction Domestic violence has received considerable attention as a serious social problem that impacts not only women, who are the most frequent victims, but also their children. For years researchers have studied the mechanisms or processes through which domestic violence affects children's behavior problems (Huang, Wang, & Warrener, 2010; Levendosky, Huth-Bocks, Shapiro, & Semel, 2003; Levendosky, Leahy, Bogat, Davidson, & von Eye, 2006; Schoppe-Sullivan, 2007; Yoo & Huang, 2012, 2013). Many of these studies, largely based on ecological frameworks, suggested that mothers' responses to domestic violence may mediate the effects that the violence has on their children. In particular, disruptions in maternal mental health and parenting, which are mostly caused by domestic violence, have been considered key mechanisms that mediate the effects of domestic violence on children's behavioral outcomes. Although the impact of domestic violence on women and children has been well documented in the literature, much less attention has been given to racial variations in this relationship. Some researchers have suggested that women's responses to domestic violence may vary by their racial and cultural backgrounds (Kasturinrangan, Krishnan, & Riger, 2004; Sokoloff & Dupont, 2005). They have found that the risk of experiencing psychological and parenting difficulties as a result of domestic violence may be greater for White women than for minority women (Campbell et al., 2008; Edelson, Hokoda, & Romos-Lira, 2007; Taylor, Guterman, Lee, & Rathouz, 2009; Torres & Han, 2000). These studies have suggested that cultural contexts may influence abused women's psychological resources and ⁎ Tel.: +82 10 5465 7271. E-mail address: [email protected].
http://dx.doi.org/10.1016/j.childyouth.2014.06.001 0190-7409/© 2014 Published by Elsevier Ltd.
parenting that are closely associated with their children's outcomes. However, no study has focused on the potential racial variations in the mechanisms through which domestic violence affects children's behavioral outcomes. Ecological theory (Belsky, 1984; Bronfenbrenner, 1979) has been widely used to understand the effects of domestic violence on children's behavioral development. The theory suggests that children's behavioral outcomes can be influenced by children's characteristics (i.e., gender and temperament), parents' characteristics (i.e., maternal mental health and parenting), and contextual factors (e.g., domestic violence). The theory regards the mother's psychological resources and parenting functioning as the most influential determinants of children's outcomes because most children, especially young children, are completely dependent on their mother for all aspects of care. Ecological theory also posits that, as a contextual factor, domestic violence can have substantial effects on children's behavioral outcomes. Such effects can be direct, and they can also occur indirectly through the negative effects that domestic violence exerts on maternal psychological well-being and parenting functioning. That is, the ecological model assumes that domestic violence can impair maternal psychological well-being and quality of parenting and thus can have negative effects on children's behavioral outcomes. Previous studies have provided empirical support for the ecological framework, which explains the mechanisms through which domestic violence affects children (Huang et al., 2010; Levendosky et al., 2003, 2006; Schoppe-Sullivan, 2007; Yoo & Huang, 2012, 2013). For example, in a longitudinal study on high-risk young children, Huang et al. (2010) found that domestic violence at Wave 1 had a negative effect on maternal mental health and reduced parenting qualities at Wave 2, which in turn increased children's behavior problems at Wave 3.
J.A. Yoo / Children and Youth Services Review 44 (2014) 90–99
Ecological theory also suggests that the influence of family-level factors on children's outcomes may vary by their relationships to macro-environmental contexts (Bronfenbrenner, 1979), which includes cultural context and the family's socioeconomic status. However, no study has focused on examining whether the mechanisms based on ecological theory apply to culturally diverse racial groups. Most existing studies have included race in their control variables but have not directly compared the effects of domestic violence on children across racial groups. Thus, this study explored whether the mechanisms linking domestic violence to children's behavioral problems in the ecological framework vary by race or cultural background. Cross-cultural or multicultural domestic violence studies have suggested that women's responses to domestic violence are in part contingent on the cultural contexts in which they live (Campbell et al., 2008; Edelson et al., 2007; Kasturinrangan et al., 2004). In the United States, race represents a key factor that defines those contexts. The research has indicated that culture has an important influence on abused women's responses, such as mental health problems (Campbell et al., 2008; Taylor et al., 2009; Torres & Han, 2000) and parenting difficulties (Edelson et al., 2007; Taylor et al., 2009). For example, Campbell et al. (2008) found that Black women who are abused tend to experience lower levels of mental health problems than do White women. Likewise, Torres and Han (2000) found that White women are more likely to experience psychological distress than do Hispanic women. In addition, several studies have reported racial differences in abused women's parenting difficulties (Edelson et al., 2007; Taylor et al., 2009). Edelson et al. (2007) reported that Hispanic women were more likely than White women to experience parenting stress. But Taylor et al. (2009) found that Hispanic abused women were less aggressive toward their children than Black or White women. These studies suggest that certain cultural contexts, particularly within minority communities, may moderate the negative effects of domestic violence on their mental health and parenting. Given the influence of cultural contexts on abused women's mental health and parenting – both of which are strong factors of children's outcomes – it does not seem reasonable to assume that minority children are affected by domestic violence through the same mechanisms or processes as White children. Some researchers have suggested that racial variations in women's and children's outcomes may be attributable not only to racial or cultural backgrounds but also to differences in socioeconomic status across racial groups, such as income, education, or marital status (Campbell et al., 2008; Sokoloff & Dupont, 2005; Yoo & Huang, 2012). In their qualitative research, Campbell et al. (2008) suggested that the relatively weak effect of domestic violence on minority families can be in part attributable to racial differences in socioeconomic characteristics. They also found that domestic violence was often not the most salient psychological stressor for Black women and that lack of financial resources or social stressors (i.e., racism) was equally or more important. This finding suggests that comparisons among racial groups could be invalid if socioeconomic status is not taken into account. Few empirical studies have examined how race or culture plays a role in the links between domestic violence, women's psychological and parenting functioning, and their children's behavioral outcomes. The paucity of empirical evidence is unfortunate because such information could be critical to better understand the complex effects of domestic violence on children and to develop more culturally sensitive services for supporting children and their mothers. Therefore, the present study addressed this gap in the literature by analyzing the potential racial variations in the effects of domestic violence on women's functioning and thereby their children's behavioral outcomes while controlling for socioeconomic characteristics. This study paid special attention to whether the mechanisms linking domestic violence's effects on children in ecological frameworks apply to culturally diverse children and whether there are culture-specific mechanisms through
91
which domestic violence affects children. Three racial groups – Black, White, and Hispanic families that live in the United States – were compared in this study. Within-group differences were not considered in this study because the sample sizes were not sufficient for separate analyses. Although minority communities are heterogeneous in terms of cultural backgrounds (e.g., country of origin, immigration status, or acculturation level), it is expected that minority women do share common influences (e.g., racism or sexism) in their experiences of domestic violence within a sociopolitical context (Kasturinrangan et al., 2004, p 320).
2. Methods 2.1. Data Data for this study were drawn from the Fragile Families and Child Wellbeing Study, a longitudinal birth cohort survey of 4898 mothers of newborns in 20 cities in the United States. This study oversampled children born to unmarried parents by design with a 3:1 ratio (3711 unmarried and 1187 married), which resulted in a larger number of low-income children and Black or Hispanic children than would be the case in a nationally representative sample (Reichmann, Teitler, Garfinkel, & McLanahan, 2001). The baseline core interviews were conducted with parents in the hospital at the time of the focal child's birth between 1998 and 2000. Follow-up surveys were conducted when the child was 1, 3, 5, and 9 years old. Because the present study focused on behavioral outcomes of preschool children, the first four waves of the surveys were used: baseline, Year 1, Year 3, and Year 5. Of the eligible mothers at baseline, 4365 were interviewed at Year 1, 4231 were interviewed at Year 3, and 4139 were interviewed at Year 5. After the age-3 and age-5 core interviews, mothers were asked to participate in an in-home assessment of parenting and child development. For the in-home assessments, 3254 mothers participated at Year 3, and 2977 participated at Year 5. This study used the two in-home assessments, which provided valuable information on the observer's evaluations of parenting but resulted in substantial loss of data. A total of 2404 mothers completed all six data collections across the 5 survey years. The 402 mothers who were not involved in a relationship with the child's biological father at baseline and Year 1 were not included in the current study. An additional 768 mothers were dropped from the analysis due to incomplete information on the explanatory and dependent variables. Finally, 37 mothers who identified their race as something other than White, Black, or Hispanic of Mexican descent were excluded from the analysis. The final sample included 1197 mothers, who had complete information on all variables across four waves. Approximately 55% of the final sample of mothers were Black (n = 659), 22% were White (n = 265), and 23% were Hispanic (n = 273). The mean age of the sample was 25. About 32% of the sample mothers did not have a high school diploma, 32% had graduated high school, and 36% had more than a high school education. At Year 1, almost 69% of the mothers were unmarried, and 43% had an income below the poverty line. About half of the children in the sample were boys (51%). Demographic differences were explored between mothers who did (n = 180) and did not report experiencing (n = 1017) any type of domestic violence at the Year 1 survey. Compared to mothers who did not report domestic violence, mothers who reported domestic violence were more likely to be unmarried and to be in poverty at Year 1. No significant differences were found for other background characteristics. To examine the influence of sample attrition on the findings, baseline characteristics of the final sample and the dropped cases were also examined. The results revealed no significant differences between the final sample and the dropped cases for most of the background characteristics, such as age, education, income, and marital status. However, the dropped cases was less likely to be Black and more likely to be Hispanic than the final cases.
92
J.A. Yoo / Children and Youth Services Review 44 (2014) 90–99
2.2. Measures 2.2.1. Children's behavior problems Children's behavior problems at age 5 were assessed with a partial scale of the Child Behavior Checklist for ages 4 to 18 (CBCL/4–8; Achenbach, 1991). The CBCL is a commonly used measure for assessing problematic behaviors. Two subscales, externalizing and internalizing behaviors, were used as the outcome variables for this study. Externalizing behavior problems were measured by aggressiveness (20 items) and delinquent behaviors (10 items). Internalizing behavior problems were measured by anxious and depressed behaviors (14 items) and withdrawal (9 items). The mothers were asked to rate several statements for how true they were (not true = 0, somewhat true = 1, very true or often true = 2). Each item was scored such that higher scores indicate more frequent or severe behavior problems. Total scores for the externalizing and internalizing subscales were calculated by summing the scores for all items. 2.2.2. Maternal mental health Maternal mental health at Year 3 was assessed by the mother's experience of a depressive episode (7 items) and generalized anxiety disorder (7 items). Both conditions are based on the Composite International Diagnostic Interview, Short Form (Walters, Kessler, Nelson, & Mroczek, 2002). For both depression and anxiety, the mothers who experienced three or more items were classified as having symptoms. The maternal mental health measure was coded yes if the mother experienced depressive symptoms or generalized anxiety disorder and no if she had never experienced any of the symptoms (1 = yes, 0 = no). 2.2.3. Parenting Parenting at Year 3 was assessed by two subdimensions: the interviewer assessment of parenting behavior and the mother's report of using spanking as a disciplinary tactic. The interviewer's assessment of parenting behavior was measured using subscales of the Home Observation for Measurement of the Environment (HOME; Bradley, 1993). The HOME instrument measures if a mother was unresponsive or harsh or exhibited a lack of verbal or social skills to the child during the home visit. The HOME instrument has been used widely to measure parenting and the home environment. The instrument has been shown to have few differences in psychometric properties across racial groups (Pachter, Auinger, Palmer, & Weitzman, 2006). One point was assigned for each negative response, which means that higher scores indicated worse parenting. Spanking was measured by the frequency with which the mother spanked the child when the child misbehaved or acted up during the month prior to the interview. Responses were coded into a 5-level variable: 0 (never), 1 (only once or twice), 2 (a few times in the past month), 3 (a few times a week), and 4 (nearly every day). 2.2.4. Domestic violence The primary independent variable of this study was domestic violence toward the mother at Year 1. Domestic violence was considered to be present if a mother reported that she had experienced any amount of physical, emotional, or sexual abuse inflicted by the child's biological father. The items “he slapped or kicked you” and “he hit you with his fist or a dangerous object” were used to measure physical violence. Items on emotional violence included (a) “he tried to isolate you from family and friends,” (b) “he tried to prevent you from going to work and/or school,” and (c) “he withheld money, made you ask for money, or took your money.” Sexual abuse was measured with the item, “he tried to make you have sex or do sexual things you didn't want to do.” There were three possible responses for each item: never, sometimes, and often. However, due to the smaller number of mothers who reported often, mothers who reported sometimes and often were combined into one category, yes. Domestic violence was coded as yes if a mother had
experienced any of the above items and as no if a mother reported never on all six items (1 = yes, 0 = no). 2.2.5. Race Racial identity was self-reported by the mother at the child's birth. Three racial groups were used in this study: non-Hispanic White (hereinafter White), non-Hispanic Black (hereinafter Black), and Hispanic mothers. The Fragile Families data allow Hispanic mothers to be disaggregated into subgroups of national origin (Mexican, Puerto Rican, South American, etc.), but the sample sizes were not sufficient for separate analyses. Although Hispanic women have diverse cultural backgrounds, they do share common influences (i.e., sexism or racism) when we consider domestic violence within a sociopolitical context (Kasturinrangan et al., 2004). 2.2.6. Controls The study controlled for socioeconomic status and child effects. Socioeconomic status included marital status, poverty, and educational attainment at Year 1. Marital status was treated as a dichotomous variable and was coded 1 if the mother was unmarried. Unmarried mothers in this study included women who were divorced, widowed, or never married. Poverty was treated as a dichotomous variable and was coded 1 if the family's income-to-needs ratio was less than 1. The family income-to-needs ratio was calculated by dividing the total family income by the official poverty threshold and adjusting the result for household size. Educational attainment was coded as less than high school, high school diploma, and more than high school and college degree. Child effects include child's gender and temperament at Year 1. Child temperament was treated as a continuous variable based on a subscale taken from the Emotionality, Activity, and Sociability Temperament Survey for Children (Buss & Plomin, 1984). Child temperament was measured by the average score of six items that asked mothers to rate the difficulty of the focal child's temperament. Higher scores indicated a more difficult temperament. Child's gender was coded 1 if the focal child was a boy. 2.3. Analysis The hypothesized path model that depicts the relationships between the main variables is presented in Fig. 1. This model hypothesizes both direct and indirect effects of domestic violence on children's behavior problems through maternal mental health and parenting. The model specified that domestic violence at Year 1 was related to maternal mental health, spanking, and parenting behavior at Year 3; that maternal mental health at Year 3, in turn, was associated with both spanking and parenting behavior at Year 3; and that both spanking and parenting behavior at Year 3 were related to children's behavior problems at Year 5. The direct effect of domestic violence at Year 1 on children's behavior problems at Year 5 was also specified. Control variables were included in the model but are not shown in the figure. Structural equation modeling with AMOS was used to examine whether the impact of domestic violence on women and children varied by race. Structural equation modeling (SEM) is a technique by which relationships between independent and dependent variables can be analyzed. SEM differs from regression techniques in that SEM allows one not only to assess direct effects but also to simultaneously analyze indirect effects through mediating variables (Pachter et al., 2006; Yoo & Huang, 2012). To determine the differences in the relationships between the variables across racial groups sampled, this study conducted separate SEM analyses for the White, Black, and Hispanic samples. And separate SEM analyses were conducted for externalizing and internalizing behavior problems by White, Black, and Hispanic groups. A total of six SEM models were analyzed. Model fit tests were conducted to determine whether the model depicted in Fig. 1 fit the observed data. A traditional chi-square test of model fit was used, which indicates adequate fit if chi-square is small
J.A. Yoo / Children and Youth Services Review 44 (2014) 90–99
Domestic violence at Yr 1
Spanking at Yr 3
Mental Health at Yr 3
93
Child behavior Problems at Yr 5
Parenting at Yr 3
Fig. 1. A hypothesized path model.
and insignificant. Several other model fit indices were also employed. The root mean square error of approximation (RMSEA) was used (values less than 0.08 indicate appropriate fit). The goodness of fit index (GFI) and the comparative fit index (CFI) were also used (good fit is indicated by values over 0.9).
3. Results 3.1. Descriptive analysis Data were analyzed for 265 White, 659 Black, and 273 Hispanic families. Table 1 displays the descriptive characteristics of these families. The results showed no large racial differences in the prevalence of domestic violence at Year 1, maternal mental health at Year 3, or children's behavior problems at Year 5. About 16% of Hispanic mothers reported that they experienced domestic violence during their child's first year of life. Almost 15% of Black and White mothers reported experiencing domestic violence, respectively. For maternal mental health, about 17% of Black and White mothers reported that they experienced mental health problems at Year 3, whereas approximately 12% of Hispanic mothers reported mental health problems at this time. Hispanic mothers rated their child's internalizing behavior problems at Year 5 slightly higher than White and Black mothers rated their children's, but there were no differences in the level of children's externalizing behavior problems across racial groups. However, maternal parenting at Year 3 differed considerably across the three racial groups. Regarding parenting behavior, Black mothers reported higher levels of negative parenting behaviors than White mothers, and Hispanic mothers fell in between the two groups. For spanking, White mothers reported more spanking than Hispanic mothers, and Black mothers were in the middle.
Most differences between the racial categories were found for socioeconomic characteristics, such as marital status, poverty, and education. The results in Table 1 show substantial variation in marital status across race. Approximately 40% of White mothers and 64% of Hispanic mothers were not married to their child's biological father at Year 1. In contrast, approximately 82% of Black mothers were unmarried at that time. Poverty also varied significantly by race. Whereas about 17% of White mothers had an income below the poverty line at the child's birth, 52% of Black mothers and 45% of Hispanic mothers reported a povertylevel income. Educational attainment also differed considerably by race. Almost 60% of White mothers had more than a high school education, compared to 30% of Black and 27% of Hispanic mothers. Approximately 33% of Black mothers and 46% of Hispanic mothers did not have a high school diploma. Finally, child temperament at Year 1 varied significantly by race. Black mothers reported child temperament problems more than White mothers did, and Hispanic mothers fell in between the two groups. No differences were found by child's gender.
3.2. Path models Results from the structural equation models of the effects of domestic violence on children's internalizing and externalizing behavior problems in White, Black, and Hispanic groups are displayed in Figs. 2 and 3. For all models, model-fit tests showed that a chi-square test for the model was nonsignificant, indicating that the model was compatible with the observed data. Likewise, the other model fit indices provided adequate fit to the data. The specific model fit for each model is shown in the figure. Paths with standardized coefficients that were significant at p ≤ .05 are highlighted by bold lines in the figure. This study controlled socioeconomic status and preexisting child effects in the
Table 1 Descriptive statistics for the variables by race.
Main variables Domestic violence at Year 1 Maternal mental health at Year 3 Parenting behavior at Year 3 (0–14) Spanking at Year 3 (0–4) Child externalizing problems at Year 5 (0–43) Child internalizing problems at Year 5 (0–26) Controls Unmarried at Year 1 Poverty at Year 1 Education at Year 1 (1–3) Child temperament at Year 1 (1–5) Boy
χ2 or F-test
White (N = 265)
Black (N = 659)
Hispanic (N = 273)
N (%)/M (SD)
N (%)/M (SD)
N (%)/M (SD)
39 (14.72) 44 (16.60) 0.61 (1.31) 0.98 (1.14) 12.29 (7.48) 17.40 (10.28)
97 (14.72) 114 (17.30) 2.02 (2.56) 1.07 (1.11) 13.10 (7.85) 18.27 (10.62)
44 (16.12) 32 (11.72) 1.13 (1.89) 0.77 (0.99) 12.87 (7.84) 19.36 (11.51)
0.78 4.63 44.42⁎⁎⁎ 7.38⁎⁎⁎
106 (40.0) 44 (16.6) 2.46 (0.72) 2.43 (0.7) 133 (50.19)
543 (82.4) 344 (52.20) 1.97 (0.79) 2.64 (0.77) 340 (51.59)
176 (64.47) 123 (45.05) 1.81 (0.83) 2.51 (0.7) 142 (52.01)
161.88⁎⁎⁎ 98.69⁎⁎⁎ 50.44⁎⁎⁎ 8.85⁎⁎⁎
Note. N (%) for domestic violence, maternal mental health, unmarried, poverty, and boy. Mean (SD) for parenting behavior, spanking, child externalizing and internalizing behavior problems, education, and child temperament. N = 1197. ⁎⁎⁎ p b .001.
1.03 2.26
0.21
94
J.A. Yoo / Children and Youth Services Review 44 (2014) 90–99
White 0.05 Domestic violence
0.16*
Child Externalizing Problems
0.21** Spanking 0.05
0.05
0.19**
0.15*
0.10 Mental Health
Parenting
−0.00
χ2 = 17.52 (df = 16, p > .05); CFI = 0.99; GFI = 0.98; RMSEA = 0.02
Black 0.06 Domestic violence
Child Externalizing Problems
0.18**
0.05 Spanking 0.09*
0.07
0.08*
0.11*
0.11** Mental Health
0.06
Parenting
χ2 = 19.06 (df = 16, p > .05); CFI = 0.99; GFI = 0.99; RMSEA = 0.02
Hispanic 0.14* Domestic violence
0.21**
Child Externalizing Problems
0.12 Spanking
0.02
0.07 0.01
0.13*
0.09 Mental Health
−0.01
Parenting
χ2 = 19.59 (df = 16, p > .05); CFI = 0.96; GFI = 0.97; RMSEA = 0.03
Fig. 2. Structural equation modeling of domestic violence on children's externalizing behavior problems in White, Black, and Hispanic groups. *p b .05. **p b .01. ***p b .001.
analyses; however, to maintain a clear focus on the effects of domestic violence, the paths between the control variables and main variables are not presented in the figures. 3.2.1. Children's externalizing behavior problems by race
control variables and children's externalizing problems. These results indicate that maternal mental health and parenting at Year 3 mediate the effects of domestic violence at Year 1 on children's outcomes at Year 5, when socioeconomic and preexisting child factors were taken into account.
3.2.1.1. White families. As shown in Fig. 2, domestic violence at Year 1 had direct associations with poor maternal mental health (β = 0.15, p b .05) and more frequent use of spanking (β = 0.16, p b .05) at Year 3. In turn, poor maternal mental health (β = 0.19 p b .01) and more frequent use of spanking (β = 0.21, p b .01) were both directly related to greater externalizing behavior problems of children at Year 5. No direct associations were found between domestic violence at Year 1 and children's externalizing problems at Year 5. Regarding the effects of control variables, unmarried-mother status (β = −0.18, p b .01) and higher levels of child temperament (β = 0.14, p b .01) at Year 1 are strongly related to greater externalizing problems of White children at Year 5. Table 2 presents the indirect and total effects of domestic violence on children's externalizing problems. The indirect effect of domestic violence at Year 1 on children's externalizing problems at Year 5 was significant (β = 0.07, p b .01). No indirect effects were found between the
3.2.1.2. Black families. Domestic violence at Year 1 had a significant, direct relationship with poor maternal mental health at Year 3 (β = 0.11, p b .05). In turn, poor maternal mental health at Year 3 was associated with increased children's externalizing problems at Year 5 (β = 0.08, p b .05), as well as with greater use of spanking at Year 3 (β = 0.09, p b .05). Poor parenting at Year 3 – both spanking (β = 0.18, p b .01) and parenting behavior (β = 0.11, p b .01) – was related to increased children's externalizing problems at Year 5. No direct link was found between domestic violence and children's externalizing problems. With respect to the control variables, unmarried-mother status (β = − 0.09, p b .05), male gender (β = 0.09, p b .05), and higher level of child temperament (β = 0.14, p b .01) at Year 1 had direct relationships with greater externalizing problems of Black children at Year 5. As presented in Table 2, the indirect effect of domestic violence at Year 1 on children's externalizing problems at Year 5 was significant
J.A. Yoo / Children and Youth Services Review 44 (2014) 90–99
95
White 0.03 Domestic violence
0.16*
Child Internalizing Problems
0.13* Spanking 0.05
0.05
0.15*
0.15*
0.10 Mental Health
Parenting
−0.00
χ2 = 17.52 (df = 16, p > .05); CFI=0.99; GFI=0.99; RMSEA=0.02
Black 0.11** Domestic violence
Child Internalizing Problems
0.10*
0.05 Spanking 0.09*
0.07
0.05
0.11*
0.12** Mental Health
Parenting
χ2 = 19.06 (df = 16, p > .05); CFI=0.99; GFI=0.99; RMSEA=0.02
0.06
Hispanic 0.21** Domestic violence
0.21**
Child Internalizing Problems
0.02 Spanking
0.02
0.07 0.05
0.13*
0.15* Mental Health
−0.01
Parenting
χ2 = 19.59 (df = 16, p > .05); CFI=0.97; GFI=0.99; RMSEA=0.03
Fig. 3. Structural equation modeling of domestic violence on children's internalizing behavior problems in White, Black, and Hispanic groups. *p b .05. **p b .01. ***p b .001.
(β = 0.03, p b .01). Interestingly, the indirect effect of maternal mental health at Year 3 on children's externalizing problems at Year 5 was also significant in this group (β = 0.02, p b .01). No indirect effect was found between control variables and children's externalizing problems. These results suggest that domestic violence at Year 1 had a negative effect on maternal mental health at Year 3 and that mental health in turn tended to increase children's externalizing problems at Year 5 directly as well as indirectly through its influence on spanking at Year 3, even when socioeconomic and preexisting child factors were controlled in the model. 3.2.1.3. Hispanic families. A quite different picture emerged when predicting externalizing problems of the Hispanic children in the sample. For this group, domestic violence at Year 1 had direct associations with poor maternal mental health (β = 0.13, p b .05) and more frequent use of spanking at Year 3 (β = 0.21, p b .01). However, maternal mental health and parenting at Year 3 were not subsequently related to children's externalizing problems at Year 5. Notably, domestic violence at Year 1 had a strong direct effect on children's externalizing problems at Year 5, even after controlling mediators (β = 0.14, p b .05). No significant effects were found between control variables and children's externalizing problems. Whereas the indirect effect of domestic violence on
children's externalizing problems was not significant, its total effect on children was (β = 0.16, p b .01). These results indicate that domestic violence at Year 1 had a strong, direct effect on children's externalizing problems at Year 5, even after controlling mediators and other variables, and that this effect was not mediated through maternal mental health or parenting at Year 3. 3.2.2. Children's internalizing behavior problems by race 3.2.2.1. White families. The direct links among variables found in the model predicting internalizing problems for White children were similar to those found for this group's externalizing problems. As shown in Fig. 3, domestic violence at Year 1 had direct associations with poor maternal mental health (β = 0.15, p b .05) and more frequent use of spanking (β = 0.16, p b .05) at Year 3, both of which, in turn, were related to greater internalizing problems of children at Year 5 (β = 0.15, p b .05; β = 0.13, p b .05). Although the indirect effect of domestic violence at Year 1 for internalizing problems at Year 5 was also significant (β = 0.05, p b .01; Table 3), its total effect on children's internalizing problems was not. These results suggest that the power that domestic violence has on women and children was weaker for
−0.01 0.07 0.02 0.09 0.12 0.02⁎⁎
0.06 0.09⁎ 0.10⁎ 0.11⁎⁎ 0.18⁎⁎
−0.01 0.07 0.01 0.09 0.12
0.01
−0.01 0.01 0.03 0.01 0.01⁎ 0.03⁎⁎
0.13⁎ 0.02 0.21⁎⁎ 0.14⁎
3.2.2.3. Hispanic families. The results for Hispanic children's internalizing problems were very similar to the results for this group's externalizing problems. That is, domestic violence at Year 1 had a direct relationship with poor maternal mental health (β = 0.13, p b .05) and more frequent use of spanking at Year 3 (β = 0.21, p b .01), but neither outcome was subsequently related to children's internalizing problems at Year 5. Domestic violence had a strong effect on children's internalizing problems (β = 0.21, p b .01), even after controlling mediators and other variables in the analyses. The indirect effect of domestic violence at Year 1 for children's internalizing problems at Year 5 was not significant; however, its total effect on children's internalizing problems (β = 0.22, p b .01) was far larger for Hispanic children than for the White and Black groups. No direct or indirect effects were found between the control variables and children's internalizing problems. The results for Hispanic families indicate that domestic violence at Year 1 had a strong, direct effect on children's internalizing problems at Year 5, but this effect was not mediated through maternal mental health or parenting at Year 3.
0.15⁎ 0.05 0.17⁎ 0.12⁎
0.01
0.05 −0.00 0.05 0.19⁎⁎ 0.10 0.21⁎⁎
0.15⁎ 0.05 0.16⁎
0.00 0.01 0.07⁎⁎
Indirect effect
−0.00 0.05 0.20⁎⁎ 0.10 0.21⁎⁎
Direct effect
3.2.2.2. Black families. The specific pathways through which domestic violence affected Black children's internalizing problems differed from the pathways found in this group's externalizing problems. Domestic violence at Year 1 was directly associated with poor maternal mental health at Year 3 (β = 0.11, p b .05), and poor maternal mental health at Year 3 was related to more frequent use of spanking at Year 3 (β = 0.09, p b .05), but not to children's problems at Year 5. Poor parenting at Year 3 – both frequent use of spanking (β = 0.10, p b .05) and negative parenting behavior (β = 0.12, p b .01) – was associated with increased externalizing problems of children at Year 5. The direct effect of domestic violence at Year 1 for children's internalizing problems at Year 5 remained significant (β = 0.11, p b .05) even after controlling mediators. Regarding the effects of the control variables, only educational attainment had a direct relationship with internalizing problems of Black children (β = −0.11, p b .05). The indirect effect of domestic violence at Year 1 on children's internalizing problems at Year 5 was significant (β = 0.02, p b .01). The indirect effect of maternal mental health at Year 3 on children's internalizing problems at Year 5 via spanking at Year 3 was also significant (β = 0.02, p b .01). These results show that, even when socioeconomic and child factors were controlled in the model, domestic violence had a negative effect on maternal mental health and that mental health tended to increase children's internalizing problems primarily through its indirect effect on spanking.
0.11⁎ 0.07 0.05 0.06 0.06 0.09⁎ 0.08⁎ 0.11⁎⁎ 0.18⁎⁎
0.11⁎ 0.08 0.06 0.09⁎
internalizing than for externalizing problems. As for the effects of the control variables, only child temperament had a direct relationship with children's internalizing problems for this racial group (β = 0.18, p b .01).
Note. ⁎ p b .05. ⁎⁎ p b .01.
Parenting Spanking
Maternal mental health Parenting behavior Spanking Externalizing problems Parenting behavior Spanking Externalizing problems Externalizing problems Externalizing problems Domestic violence
Maternal mental health
Dependent variables
4. Discussion
Predictor
Table 2 Indirect and total effects of domestic violence on children's externalizing behavior problems.
Total effects
Total effects Direct effect Direct effect
Indirect effect
Hispanic Black White
Indirect effect
0.13⁎ 0.02 0.22⁎⁎ 0.17⁎⁎
J.A. Yoo / Children and Youth Services Review 44 (2014) 90–99 Total effects
96
The present study examined whether there were racial differences in the effects of domestic violence on children's externalizing and internalizing behavior problems when socioeconomic factors and preexisting child characteristics were controlled for. This longitudinal study provides the first empirical evidence of racial variations in the mechanisms through which domestic violence affects children's behavior problems. The results from this study indicate that domestic violence had negative effects on children in all racial groups, but the mechanisms through which domestic violence exerted its effects on children differed among White, Black, and Hispanic families. In particular, whereas the effects of domestic violence on children were mediated through maternal mental health and parenting for White and Black families, only direct effects were found for Hispanic families. These findings suggest that the conceptual framework that links domestic violence and children may not apply to all racial groups and that cultural contexts may contribute to children's and mothers' experiences of abuse in violent families.
0.02
0.00 0.07 0.12⁎⁎ 0.10⁎
0.02⁎⁎ 0.05 0.12⁎⁎ 0.10⁎ Note. ⁎ p b .05. ⁎⁎ p b .01.
0.10 0.13⁎ Parenting Spanking
Maternal mental health
Maternal mental health Parenting behavior Spanking Internalizing problems Parenting behavior Spanking Internalizing problems Internalizing problems Internalizing problems Domestic violence
0.15⁎ 0.05 0.16⁎ 0.03 −0.00 0.05 0.15⁎
0.01
0.00 0.01 0.05⁎⁎
0.10 0.13⁎
0.01 0.01⁎ 0.02⁎⁎
0.11⁎ 0.07 0.05 0.11⁎⁎ 0.06 0.09⁎
0.15⁎ 0.05 0.17⁎ 0.08 −0.00 0.05 0.15⁎
Indirect effect
0.02
0.13⁎ 0.02 0.22⁎⁎ 0.22⁎⁎ −0.01 0.07 0.05 0.15⁎ −0.00 0.01 0.01
0.13⁎ 0.02 0.21⁎⁎ 0.21⁎⁎ −0.01 0.07 0.05 0.15⁎ 0.11⁎ 0.08 0.06 0.13⁎⁎ 0.06 0.09⁎
Hispanic Black
Direct effect Direct effect
Dependent variables Predictor
Table 3 Indirect and total effects of domestic violence on children's internalizing behavior problems.
Total effects White
Indirect effect
Total effects
Direct effect
Indirect effect
Total effects
J.A. Yoo / Children and Youth Services Review 44 (2014) 90–99
97
The results for White families in this study showed that the effects of domestic violence on children, especially for externalizing behavior problems, were fully mediated through maternal mental health and parenting, which coincides well with an ecological framework. The results indicate that domestic violence at Year 1 had negative effects on mothers' mental health and parenting (particularly spanking) at Year 3, both of which in turn tended to increase problematic behaviors of children at Year 5 for White families. The results could be taken as evidence that abuse-induced disruptions in maternal mental health and parenting could be key mechanisms linking domestic violence and children for White families. A different picture emerged when analyzing the effects of domestic violence on Black children. For Black families, the specific pathways through which domestic violence affected children differed by the type of behavior problem. Whereas the effects of domestic violence on externalizing problems were fully mediated through maternal mental health, the effects on internalizing problems were direct and were partially mediated through maternal mental health. Although the specific link between domestic violence and children somewhat differed by the type of behavior problem, the overall results for Black families suggested that the mediation effects occurred primarily through maternal mental health. That is, abused women in these Black families frequently suffered from mental health problems, such as depression and anxiety, and their mental health problems were found to be key mechanisms through which domestic violence led to poor parenting (particularly spanking) and thereby poor behavioral outcomes of children. The results clearly showed that Black mothers whose children's behaviors were compromised were struggling with mental health problems. However, the emotional pain caused by domestic violence for Black women may be underestimated in the literature (Campbell et al., 2008) because such difficulties are frequently confounded with several other stressors common for this group, such as lack of resources or social stress. The findings in this study have important implications for social service efforts that support women and children in violent families; particularly, such efforts should address Black victims' need for emotional support to improve their parenting functioning and to reduce their children's behavior problems. Unlike the results for White families, however, mediation effects on parenting were not supported in the results for Black families. In particular, domestic violence at Year 1 did not have direct effects on parenting at Year 3, although parenting was a strong predictor of children's behavior problems at Year 5. It is not clear why Black mothers' parenting was not directly affected by domestic violence. A possible explanation is that the weak association between domestic violence and parenting may have something to do with the cultural validity of parenting measures. Parenting is influenced largely by culture. For instance, spanking is commonly considered to be an acceptable form of disciplinary behavior in Black cultures (Whaley, 2000) and thus a Black mother's use of spanking might actually not be strongly affected by domestic violence. Alternatively, given that Black women experience multiple institutional prejudices, abused Black mothers may be less likely to self-report spanking or negative parenting behaviors (Cardoso, Padilla, & Sampson, 2010). Future studies should include multiple measures of parenting that are culturally sensitive for all races and more thoroughly investigate the mediation effects of parenting in Black families. Perhaps the most interesting and unexpected findings of the present study was that there were no mediation effects of maternal mental health and parenting in the link between domestic violence and children in the Hispanic group. Among the Hispanic families, domestic violence had strong effects on both the externalizing and internalizing behaviors of children; however, contrary to the findings for White and Black families, these effects were not mediated through maternal mental health or parenting. Abused Hispanic mothers in this study exhibited mental health problems and poor parenting (especially spanking), but their mental health and parenting difficulties were not subsequently
98
J.A. Yoo / Children and Youth Services Review 44 (2014) 90–99
related to increased child behavior problems. It is unclear why Hispanic children were not influenced by maternal mental health and parenting difficulties caused by abuse. Prior research has not provided specific information about how Hispanic children respond to or cope with domestic violence and its related maternal dysfunction. As a potential explanation, the unexpected results for Hispanic children might be associated with length of exposure to domestic violence. One of the limitations in this study was that this study did not measure whether the domestic violence was a onetime event that happened at Year 1 or whether it continued in subsequent waves. Prior studies suggested that Hispanic women may stay in violent relationships longer to keep their family intact (Edelson et al., 2007). Married Hispanic mothers in this study who experienced domestic violence at Year 1 were more likely to still be living with the child's father in subsequent waves, and in these cases it appeared that domestic violence was ongoing. The Hispanic children who were exposed to ongoing violence may have been more strongly affected by their father's violent behaviors than by their mother's responses to the violence. Another possible explanation is that Hispanic children might be better helped through a continuum of naturally occurring informal support within the family or their community. Literature suggests that many Hispanic women and children tend to live in extended families that may provide social support, such as emotional support, child care, and protection for abuse victims and children (Kasturinrangan et al., 2004; Sokoloff & Dupont, 2005). Such support might serve as protective factors for Hispanic children and make it easier for them to cope with maternal stress related to domestic violence. Still, more research is needed to investigate what cultural aspects within the Hispanic community contribute to reducing the effects of abuseinduced disruptions in maternal functioning on children's behavior problems. The results from this study must be interpreted in the context of the following limitations. First, the substantial loss of data due to sample attrition across six surveys, as well as incomplete data for the variables used in this study, might have influenced the results. Second, this study did not examine how domestic violence at Year 1 interrelates over time with other risk factors (e.g., racism) and protective factors (e.g., religion or family) that might intervene in any direct relationship between domestic violence at Year 1 and child outcomes at Year 5. For example, domestic violence and its interaction with religion or social support in minority communities might be associated with differences in groups in this study. Future research is necessary to include multiple risk and protective factors in the analyses and to assess how those factors play a role in the effects of domestic violence on children. Third, the measures of domestic violence, maternal mental health, and children's behavior problems relied solely on maternal reports, which might have led to common single-reporter bias. Fourth, this study only included Black, White, and Hispanic families because of insufficient sample sizes for other racial groups. Some research has suggested that Native American and Asian women are at a greater risk of domestic violence and have specific cultural barriers and strengths related to abuse (Grossman & Lundy, 2007). Future research is needed to include more diverse racial categories and examine any cultural-specific mechanisms that might be present within these groups. Finally, this study did not consider within-group differences. Minority communities tend to be heterogeneous in terms of cultural backgrounds. For example, Hispanic refers to individuals of several national origins such as Mexican, Puerto Rican, Cuban, and Central American. Even though these groups share some cultural similarities and the same language, these subgroups also differ in their country of origin, history of migration, acculturation level, and other factors that may influence the experiences of abused women and their children. In addition, there may be differences between long-standing U.S. families and newly arrived immigrant families within minority communities. Newly arrived immigrant women tend to experience limited English skills, social isolation, or lack of information of social services, which makes it difficult for them to cope with
domestic violence and the related stress (Kasturinrangan et al., 2004). Minority groups also differ in their religious beliefs which might influence the experience of abused women and their children (e.g., Catholic and Protestant groups within the Black families). Given the results that the effects of domestic violence on children varied by race, future research is required to investigate the within-group variability in the experiences of battered women and children. Despite these limitations, this study revealed that children's and mothers' experiences of domestic violence varied by race and that cultural contexts may contribute to the dynamics of children's and mother's experiences. Although more research is necessary to fully understand such racial variations, it is evident that social service professionals need to be culturally sensitive when working with abused women and their children from diverse racial groups. Social service providers must understand the racial and cultural differences in mothers' and children's experiences of domestic violence, as well as the particular needs that different groups have (Sokoloff & Dupont, 2005), without perpetuating negative stereotypes for racial groups. This study found that White children were more affected by maternal dysfunction caused by domestic violence than were their minority peers. Emotional support and parenting intervention programs for abused mothers can be effective in preventing or minimizing the negative effects of domestic violence on White children. For Black families, this study found that domestic violence is a psychologically stressful event that leads to maternal depression and anxiety. Thus, social service providers should first conduct open-minded assessments of women's need for emotional support when working with Black families. To address the needs of Black women's mental health problems, social service providers should understand their community or cultural values and be aware of cultural strengths (e.g., religion) or barriers to help-seeking (e.g., community pressures or racism) in the Black community. They should also attempt to involve community members who share the backgrounds of the victims, and develop culturally congruent forms of emotional support and resources. These efforts may have a positive effect on reducing mental health problems of abused Black women and thereby minimizing their children's behavior problems. For Hispanic families, this study revealed that children can be strongly affected by the presence of domestic violence within the family without being additionally affected by maternal functioning. These findings suggest that emotional support services or parenting intervention programs developed for White families might be not effective in reducing behavior problems of Hispanic children. Rather, when working with Hispanic families, social service providers should pay more attention to direct links between the father's violent behaviors and children's outcomes. This study also suggests that Hispanic children might be culturally protected from the effects of maternal dysfunction caused by domestic violence. Thus, social service providers should assess support systems of victims and children within the family or community and attempt to collaborate with the support groups to develop culturally appropriate services. These efforts should help abused victims and their children cope with domestic violence and the related challenges more effectively. Finally, it is important that social service providers should understand that many minority families experience other social stressors (e.g., lack of resources or racism) which may affect their lives as well as their ability to access resources (Campbell et al., 2008; Kasturinrangan et al., 2004). Thus, social service providers should note that services or interventions for supporting minority families can be more effective when their efforts are combined with other physical and environmental services. Acknowledgments This work was supported by the Brain Korea 21 Plus Project (Ewha Womans University, School of Social Welfare) of the National Research Foundation of Korea Grant (21B20130012299).
J.A. Yoo / Children and Youth Services Review 44 (2014) 90–99
References Achenbach, T. M. (1991). Manual for the child behavior checklist/4–18 and 1991 profile. Burlington, VT: University of Vermont Department of Psychiatry. Belsky, J. (1984). The determinants of parenting: A process model. Child Development, 55, 83–96. Bradley, R. H. (1993). Children's home environments, health, behavior, and intervention efforts: A review using the HOME inventory as a marker measure. Genetic, Social, and General Psychology Monographs, 119, 437–490. Bronfenbrenner, U. (1979). The ecology of human development: Experiments by nature and design. Cambridge, MA: Harvard University Press. Buss, A. H., & Plomin, R. (1984). Temperament: Early developing personality traits. Hillsdale, NJ: Erlbaum. Campbell, J., Campbell, D. W., Gary, F., Nedd, D., Price-Lea, P., Sharps, P. W., et al. (2008). African American women's responses to intimate partner violence: An examination of cultural context. Journal of Aggression, Maltreatment & Trauma, 16(3), 277–295. Cardoso, J. B., Padilla, Y. C., & Sampson, M. (2010). Racial and ethnic variation in the predictors of maternal parenting stress. Journal of Social Service Research, 36, 429–444 (f). Edelson, M. G., Hokoda, A., & Romos-Lira, L. (2007). Differences in effects of domestic violence between Latina and non-Latina women. Journal of Family Violence, 22, 1010. Grossman, S. F., & Lundy, M. (2007). Domestic violence across race and ethnicity: Implications for social work practice and policy. Violence Against Women, 13(10), 1029–1052. Huang, C., Wang, L. R., & Warrener, C. (2010). Effects of domestic violence on behavior problems of preschool-aged children: Do maternal mental health and parenting mediate the effects? Children and Youth Services Review, 32, 1317–1323. Kasturinrangan, A., Krishnan, S., & Riger, S. (2004). The impact of culture and minority status on women's experience of domestic violence. Trauma, Violence & Abuse, 5(4), 318–332. Levendosky, A. A., Huth-Bocks, A.C., Shapiro, D. L., & Semel, M.A. (2003). The impact of domestic violence on the maternal–child relationship and preschool-age children's functioning. Journal of Family Psychology, 17(3), 275–287.
99
Levendosky, A. A., Leahy, K. L., Bogat, G. A., Davidson, W. S., & von Eye, A. (2006). Domestic violence, maternal parenting, maternal mental health & infant externalizing behavior. Journal of Family Psychology, 20(4), 544–552. Pachter, L. M., Auinger, P., Palmer, R., & Weitzman, M. (2006). Do parenting and the home environment, maternal depression, neighborhood, and chronic poverty affect child behavioral problems differently in different racial–ethnic groups? Pediatrics, 117(4), 1329–1338. Reichmann, N. E., Teitler, J. O., Garfinkel, I., & McLanahan, S. S. (2001). Fragile families: Sample and design. Children and Youth Services Review, 23(4–5), 303–326. Schoppe-Sullivan, S. J. (2007). Marital conflict and children's adjustment: Evaluation of the parenting process model. Journal of Marriage and Family, 69, 1118–1134. Sokoloff, N. J., & Dupont, I. (2005). Domestic violence at the intersections of race, class, and gender. Violence Against Women, 11(1), 38–64. Taylor, C. A., Guterman, N.B., Lee, S. J., & Rathouz, P. (2009). Intimate partner violence, maternal stress, nativity, and risk for maternal maltreatment of young children. American Journal of Public Health, 99(1), 175–183. Torres, S., & Han, H. -R. (2000). Psychological distress in non-Hispanic White and Hispanic abused women. Archives of Psychiatric Nursing, XIV(1), 19–29. Walters, E. E., Kessler, R. C., Nelson, C. B., & Mroczek, D. (2002). Scoring the World Health Organization's Composite International Diagnostic Interview Short Form (CIDI-SF). http://www.who.int/msa/cidi/CIDISFScoringMemo12-03-02.pdf Whaley, A. L. (2000). Sociocultural differences in the developmental consequences of the use of physical discipline during childhood for African Americans. Cultural Diversity and Ethnic Minority Psychology, 6(1), 5–12. Yoo, J. A., & Huang, C. -C. (2012). The effects of domestic violence on children's behavior problems: Assessing the moderating roles of poverty and marital status. Children and Youth Services Review, 34(12), 2464–2473. Yoo, J. A., & Huang, C. -C. (2013). Long term relationships among domestic violence, maternal mental health and parenting, and preschool children's behavior problems. Families in Society, 94(4), 268–276.