I need someone to create a 2 page  mini-lecture targeted specifically for first year Marriage & Family Therapy  students. APA style

Include information or responses to the following items:

A list of the strengths or characteristics of resilience identified in your reading and reflection.

A comment on how cultural stress, financial stress, and social stress play roles in resilience (include your reflections on each of these aspects).

A declaration of the strengths you believe will be most helpful to families with whom you will be working in dealing with violence. Be sure to explain how you would apply these ideas to your work with diverse families.

A description of how you will assist families to develop/use these strengths to prevent or overcome the consequences of family violence.

Your goal would be to take these four points and apply them to helping you develop resilience for yourself, learn about family violence, and help your clients who are victims of family violence.

References:

Chambers, J. C. (2005). Strengths for.    coping with family conflict.

Stephens, K. (2010). Fostering resilience in children exposed to domestic violence.

Resilience Among Children Exposed to Domestic Violence: The Role of Risk

and Protective Factors

Cecilia Martinez-Torteya, G. Anne Bogat, Alexander von Eye, and Alytia A. Levendosky Michigan State University

Individual and family characteristics that predict resilience among children exposed to domestic violence (DV) were examined. Mother–child dyads (n = 190) were assessed when the children were 2, 3, and 4 years of age. DV-exposed children were 3.7 times more likely than nonexposed children to develop internalizing or externalizing problems. However, 54% of DV-exposed children maintained positive adaptation and were characterized by easy temperament (odds ratio [OR] = .39, d = .52) and nondepressed mothers (OR = 1.14, d = .07), as compared to their nonresilient counterparts. Chronic DV was associated with maternal depression, difficult child temperament, and internalizing or externalizing symptoms. Results underscore heterogeneous outcomes among DV-exposed children and the influence of individual and family characteristics on children’s adaptation.

Using a person-oriented approach, this longitudinal study examined the individual and family factors that predict resilience among young children (from ages 2 to 4) exposed to domestic violence (DV; defined as male aggression toward a female partner). Research has consistently documented the negative effects of DV exposure on children’s adaptation (e.g., Kitzmann, Gaylord, Holt, & Kenny, 2003), but very few studies have explored what characteristics typify DV-exposed children who maintain positive adaptation (Grych, Jouriles, Swank, McDonald, & Norwood, 2000; Hughes & Luke, 1998). Longitudinal studies with high-risk samples (not DV exposed) have previously identified a variety of individual and family factors that predict and hinder resilience (e.g., Tiet et al., 1998; Wyman et al., 1999), but the role of these factors has not been delineated in the context of DV exposure.

DV and Its Negative Effects on Children

DV-exposed children are approximately 2 times more likely than nonexposed children to exhibit internalizing and externalizing problems (Stern-

berg, Baradaran, Abbott, Lamb, & Guterman, 2006). Among young children, the trauma of exposure is likely to disrupt the development of basic compe- tencies, threatening the child’s ability to process and manage emotions effectively and increasing internalizing and externalizing behaviors (Cole, Zahn-Waxler, Fox, Usher, & Welsh, 1996). For example, young DV-exposed children experience more distress in response to interadult conflict than their nonexposed peers (DeJonghe, Bogat, Levendo- sky, von Eye, & Davidson, 2005), and trauma symptoms have been reported in children as young as age 1 (Bogat, DeJonghe, Levendosky, Davidson, & von Eye, 2006).

Overall, children exposed to DV are at risk to develop both internalizing and externalizing prob- lems (e.g., Sternberg et al., 2006). However, chil- dren’s outcomes vary widely, and many children have adequate behavioral and emotional function- ing despite witnessing DV. Research has reported resilience rates from 31% to 65% (Grych et al., 2000; Hughes & Luke, 1998) among school-age children. These studies suggest that positive adaptation is associated with less partner-to-mother physical aggression, shorter duration of DV exposure, per- ception of the conflict as less threatening, less self- blame, and absence of maternal depression (Grych et al., 2000; Hughes & Luke, 1998). However, the scope of these studies was limited, and many individual and family characteristics that have been

This research was supported by a grant from the National Institute of Justice (8-7958-MI-IJ) and Centers for Disease Control (R49 ⁄ CCR ⁄ 518519-03-1) to the second, third, and fourth authors. Portions of this paper were presented at the Society for Research in Child Development biennial meeting in Boston, Massachu- setts, in March 2006.

Correspondence concerning this article should be addressed to Cecilia Martinez-Torteya, Clinical Psychology, Michigan State University, Psychology Building, East Lansing, MI. Electronic mail may be sent to [email protected].

Child Development, March/April 2009, Volume 80, Number 2, Pages 562–577

� 2009, Copyright the Author(s) Journal Compilation � 2009, Society for Research in Child Development, Inc. All rights reserved. 0009-3920/2009/8002-0019

identified as protective factors among other high-risk samples (e.g., positive parenting, easy temperament) were not assessed. In addition, these studies were cross-sectional, and early develop- mental trajectories associated with resilience or recovery were not investigated.

The present research explored a number of fac- tors relevant to young children’s adaptation identi- fied by previous research with high-risk children: child’s easy temperament, child’s cognitive ability, positive parenting, maternal depression, stressful life events, low income, and minority status. Indi- vidual- and family-level factors are consistent with the guidelines for research on resilience recom- mended by Luthar, Sawyer, and Brown (2006) including (a) saliency to life context (young chil- dren’s reliance on their caregivers, high prevalence of maternal depression and stressful life events in DV households), (b) enduring characteristics (rela- tive stability of temperament and cognitive ability), and (c) contribution to the development of other assets (impact on children’s self-regulation and cop- ing skills).

Defining Resilience

Resilience has been defined as the maintenance of healthy ⁄ successful functioning or adaptation within the context of a significant adversity or threat (Garmezy, 1993; Luthar, Cicchetti, & Becker, 2000; Masten & Obradovic, 2006). Thus, two elements must co-occur for resilience to be present: a circum- stance that has the potential to disrupt children’s development and reasonably successful adaptation (Luthar et al., 2000; Masten, 2001). Multiple defini- tions have been used to measure risk or adversity; an individual risk model explores the contribution of one risk factor to the development of negative outcomes (e.g., child maltreatment), whereas a cumulative risk model asserts that accumulation of adversity results in maladaptation (Rutter, 1979). Despite the widespread use of both models in the study of resilience, cumulative risk models provide limited insight into unique characteristics that provide protection in the context of a particular risk (i.e., DV); thus, individual risk models might be better suited to identify predictors of resilience in the children exposed to DV.

Likewise, positive adaptation has been defined in several ways, including absence of psychopathol- ogy (Tiet et al., 1998), behavioral and cognitive competence (Kim-Cohen, Moffitt, Caspi, & Taylor, 2004), and mastery of appropriate developmental tasks (Masten, 2001). Masten and Obradovic (2006)

emphasize the importance of both external adapta- tion to the environment and internal sense of well- being as part of a comprehensive assessment of resilience. Moreover, resilience is better character- ized as a dynamic process, because individuals can be resilient to specific environmental hazards or resilient at one time period but not another (Rutter, 2006).

Consistent with an individual risk model, the present study used DV exposure as the index of adversity; other risk factors (maternal depression, stressful life events, minority status, and low income) that might heighten the negative effects of DV or individually contribute to disruptions in chil- dren’s adaptation were also explored. To incorpo- rate indices of both external and internal functioning, positive adaptation is defined as the maintenance of nonclinical levels of both internaliz- ing and externalizing behaviors over time.

Risk and Protective Factors

Broadly defined, protective or promotive factors refer to the characteristics that enhance adaptation, whereas the terms vulnerability and risk are used for the factors that increase the likelihood of maladap- tation. However, there are some inconsistencies associated with the use of these terms (Luthar et al., 2000). Some authors have used an internal versus external classification, in which vulnerability refers to the individual’s biologically based traits, whereas risk concerns environmental threats to adaptation (e.g., Shannon, Beauchaine, Brenner, Neuhaus, & Gatzke-Kopp, 2007). Other authors propose a classification based on an interaction with adversity, such that risk and compensatory fac- tors have a similar effect on the whole population (individuals who are and are not exposed to adver- sity), whereas vulnerability and protective factors influence outcomes only under high-risk conditions (e.g., Steinhausen & Metzke, 2001).

Despite conceptual inconsistencies, research has reliably reported a number of characteristics associ- ated with resilience, including a positive and sup- portive caregiver–child relationship, competent (structured and warm) parenting, positive caregiver mental health, child easy ⁄ engaging temperament, and child’s higher cognitive ability (Masten et al., 1999; Tiet et al., 1998; Wyman et al., 1999). Research has also found that positive adaptation is associated with lower levels of risk, including less parental psychopathology, life stress, and poverty, as well as being a member of a majority ethnic group (Bradley & Corwyn, 2002; Leech, Larkby, Day, & Day, 2006).

Child Resilience to Domestic Violence 563

Interestingly, the profiles of resilient children resemble those of competent children who are not exposed to adversity or high risk (Masten et al., 1999).

The present study focuses on characteristics at the family level (positive parenting, maternal depression, stressful life events, and low income) and individual level (easy temperament, high cog- nitive ability, race) that might influence children’s adaptation. The term risk is used to describe envi- ronmental characteristics that have been commonly associated with increases in child behavioral ⁄ emo- tional problems (maternal depression, stressful life events, minority status, and low income), whereas protective factor will be used for the environmental and individual characteristics that are associated with an increase in positive adaptation (including promoting effects, child easy temperament, positive parenting, child high cognitive ability).

Protective Factors: Positive Parenting, Child Easy Temperament, and Cognitive Ability

Positive Parenting

Parental warmth, positive expectations, support, and low derogation predict children’s behavioral and emotional adaptation under a wide variety of adverse circumstances (Katz & Gottman, 1997; Kim-Cohen et al., 2004). Among DV-exposed chil- dren, high maternal authority or control is associ- ated with more positive and less antisocial behaviors (Levendosky & Graham-Bermann, 2000), and effective parenting is associated with decreased externalizing behaviors (Levendosky, Huth-Bocks, Shapiro, & Semel, 2003). Thus, children whose mothers are available and supportive will be better able to develop self-regulation abilities within the context of effective mother–child interactions (Wyman et al., 1999).

Child Temperament

At-risk children with easy temperaments (regu- larity, approachability, high adaptability, positive mood, low reactivity; Thomas & Chess, 1985) show fewer behavior problems than children with diffi- cult temperaments (Kim-Cohen et al., 2004; Smith & Prior, 1995; Tschann, Kaiser, Chesney, Alkon, & Boyce, 1996). Children with easy temperaments are less reactive to stressors and more likely to utilize active and flexible coping strategies to deal with stress (Compas, Connor-Smith, & Jaser, 2004); they are also better able to regulate their feelings of sad-

ness and anger (Olson, Bates, Sandy, & Schilling, 2002). Research examining the moderating effect of child’s temperament on DV-exposed children’s adaptation is sparse. For example, DeJonghe et al. (2005) found that temperament predicted infants’ observed distress to verbal conflict among children not exposed to DV but not among exposed chil- dren. However, no studies to date have examined the link between temperament and internalizing and externalizing problems in the context of DV.

Cognitive Ability

High intelligence has also been associated with positive adaptation in the face of adversity (Jaffee, Caspi, Moffitt, Polo-Tomas, & Taylor, 2007; Tiet et al., 1998) and is predictive of lower levels of psy- chiatric disorders, lower rates of conduct problems, and higher levels of overall functioning (Malcarne, Hamilton, Ingram, & Taylor, 2000). Similar studies with DV-exposed samples are sparse but suggest that IQ is negatively associated with behavioral problems (Kolbo, 1996). Well-developed verbal cog- nitive abilities may facilitate verbal mediation of conflict and therefore lead to more appropriate behavioral choices and a wider range of coping strategies (Buckner, Mezzacappa, & Beardslee, 2003).

Risk Factors: Maternal Depression, Stressful Negative Life Events, Income, and Race

Maternal Depression

Maternal depression has been associated with negative child adjustment among high-risk chil- dren (Tiet et al., 1998) as well as DV-exposed chil- dren (e.g., Levendosky, Leahy, Bogat, Davidson, & von Eye, 2006). Levendosky et al. (2006) found that maternal functioning (posttraumatic stress disorder, depression, anxiety, and self-esteem) mediated the relation between DV exposure and preschoolers’ externalizing problems. A depressed mother’s unavailability may lead her child to expect rejection and feel helpless. Depressed mothers might also model maladaptive emotion regulation strategies; for example, Kliewer et al. (2004) found that mothers who are unable to man- age effectively their own feelings of sadness have children with more internalizing problems. Alter- natively, the depressed mother’s predisposition to psychopathology may be inherited by her child and activated by the stress of exposure to marital violence (Jaffee, 2005).

564 Martinez-Torteya, Bogat, von Eye, and Levendosky

Stressful Life Events

The accumulation of stressful life events is also associated with children’s emotional and behavioral problems (Smith & Prior, 1995). Among children exposed to DV, mothers’ reports of stressful life events are a significant predictor of children’s increased behavior problems and decreased social competence (Levendosky et al., 2003). The experi- ence of stressful life events increases the environ- mental demands on the mother–child dyad, such that the mother’s parenting can be negatively affected and the child might become sensitized to stressful situations (Davies, Winter, & Cicchetti, 2006).

Race and Low Income

Research on the effects of racial background on the development of internalizing and externalizing problems has shown mixed results. For example, Leech et al. (2006) found that being African Ameri- can was a predictor of higher levels of depression and anxiety. On the other hand, a higher incidence of depression and behavioral problems among White children, as compared to minority children, has also been reported (Gerard & Buehler, 2004). Additionally, children in poor families are more likely to develop behavioral and emotional prob- lems than children in middle and upper income families (Bradley & Corwyn, 2002), and low income is associated with a relatively high prevalence of other risk factors, such as maternal psychiatric symptoms, life stress, and ineffective parenting (e.g., McLoyd, 1998).

Person Orientation

Current research primarily uses a variable- oriented approach to study the effects of DV on chil- dren’s development (Levendosky, Bogat, & von Eye, 2007). However, aggregating all participants in a single group (e.g., children exposed to DV) may mis- represent the individuals within the group in many important aspects. Alternatively, the person-ori- ented approach assumes that individuals are unique and behavior can be understood through the pre- dictable patterns that occur across the dependent and independent variables (Bogat, Levendosky, & von Eye, 2005). Person-oriented research ascertains the complex and multifaceted nature of child devel- opment by examining the patterns that result from the interassociations among variables at various eco- logical levels (Bergman & Magnusson, 1997).

Person-focused approaches to resilience allow the researcher to explore specific patterns and local associations that exist within groups, identifying individuals with positive versus negative function- ing (Masten, 2001). Complex designs include four groups of children that differ in their levels of adversity and adaptation. However, the profiles of risk and protection have not been investigated in the context of DV exposure. For the present research, the four-group model proposed by Masten was used, and groups were defined as follows: Resilient children are exposed to DV and are positively adapted, nonresilient children are exposed to DV and are negatively adapted, compe- tent children are not exposed to DV and are posi- tively adapted, and vulnerable children are not exposed to DV and are negatively adapted.

Hypotheses

Consistent with previous research, we expected to find a group of young children who showed resilience to DV. However, DV-exposed children were hypothesized to be more likely to develop emotional and behavioral problems than were non- exposed children, and longer duration and higher frequency of DV exposure were expected to pre- dict internalizing and externalizing symptoms. Family and individual characteristics that were hypothesized to predict resilience included posi- tive parenting, child’s easy temperament, and child’s high cognitive ability. Nonresilience was expected to be characterized by maternal depres- sion, more stressful negative life events, low income, and ethnic minority status. Resilient chil- dren were predicted to display characteristics simi- lar to competent nonexposed children. Additionally, specific configurations of adversity (e.g., early vs. late DV exposure), risk, and protec- tive factors were hypothesized to be associated with positive and negative adaptation.

Method

Participants

The present research is part of a larger, longitudi- nal study that explores the effects of DV on women and their children (Bogat, Levendosky, & Davidson, 1999; Levendosky, Bogat, Davidson, & von Eye, 2000). Participants of the original study were 206 pregnant women recruited from urban, suburban, and rural areas in a Midwestern state. The study oversampled for DV in order to obtain two

Child Resilience to Domestic Violence 565

relatively equal numbered groups: women who expe- rienced violence during pregnancy and nonexposed women with similar demographic characteristics. Sixteen mother–child dyads from the original sample were excluded because of mother’s death (n = 2), child’s death (n = 2), loss of custody (n = 6), mother’s imprisonment (n = 1), or withdrawal from the study before the child’s first birthday (n = 5). The excluded participants did not differ from the current study sample on demographic characteris- tics, such as maternal age, maternal race, child’s race or child’s gender. Excluded women reported lower family income as well as higher levels of DV and depression at the first assessment period (third trimester of pregnancy).

Participants for the current study were 190 chil- dren (95 boys) and their mothers who were assessed yearly when the children were ages 2, 3, and 4. Forty-seven percent of the children were Caucasian, 25% African American, 23% multiracial, 2% Hispanic, 2% Native American, and 1% Asian American. The median monthly family income when children were 2 years old was $2,542 (SD = $2,187), and the average age of mothers was 27.38 (SD = 4.99).

Measures

Screening Instrument: DV

A subset of items from the 14-item Verbal and Physical Aggression scales of the Conflict Tactics Scales (CTS; Straus, 1979) were used to classify women’s DV experiences prior to the interview (Items 6 to 14; e.g., ‘‘Threatened you with a gun or knife’’). Straus, Hamby, Boney-McCoy, and Sugar- man (1996) reported an internal consistency of a = .58 for the verbal and physical aggression scale. For the present study, internal consistency was good (a = .88).

Grouping Variables

DV. The 46-item Severity of Violence Against Women Scales (SVAWS; Marshall, 1992) assessed violent behaviors and threats that the woman had experienced from her partner during the last year on a 4-point frequency scale. A total score is obtained by adding all items (0 to 138); higher scores represent more frequent abuse. High internal consistency (a = .97) has been previously reported (Huth-Bocks, Levendosky, & Semel, 2001). For the present study scores ranged from 0 to 94 and

internal consistency was excellent (a = .95 at age 2, a = .94 at age 3, a = .94 at age 4). Children were assigned to the DV group if their mothers endorsed any DV item at any of the three time periods. A DV dichotomous score (present or absent) was also created for each time period. A duration of DV score (0–3) was computed by adding the number of time periods in which the woman reported experiencing DV.

Child’s behavioral adaptation. Developmentally appropriate versions of the Child Behavior Checklist (99-item CBCL 2–3, Achenbach, 1992; 113-item CBCL 4–18, Achenbach, 1991) assessed children’s behavioral and emotional functioning at each time period. Mothers rated each item on a 3-point scale. There are two broadband subscales: Internalizing and Externalizing. T scores (30–100) were used with higher scores reflecting more frequent ⁄ severe symptoms. Excellent internal consistency (as = .90–.96) has been reported for the broadband scales (Achenbach, 1991, 1992). In this study, scores ranged from 30 to 72 for Inter- nalizing (a = .81 at age 2, a = .77 at age 3, a = .69 at age 4) and 30 to 77 for Externalizing (a = .89 at age 2, a = .90 at age 3, a = .85 at age 4). Children were classified as positively adapted if their CBCL scores were lower than 60 for both scales at all time periods. Children with scores equal to or higher than 60 (i.e., clinical cutoff; Achenbach, 1992) on either scale at one or more time periods were classified as negatively adapted.

Protective Factors

Maternal positive parenting. The nurturing (20 items; e.g., ‘‘I read to my child at bedtime’’) and discipline (30 items; e.g., ‘‘I send my child to bed as a punishment’’) subscales of the Parent Behav- ior Checklist (PBC; Fox, 1994) were used to assess maternal parenting. Responses are scored on a 4-point scale. For the present study, the discipline (reverse coded; higher scores reflect less harsh discipline) and nurturing scores were summed to create a composite positive parenting variable (50–200); higher scores indicate more consistent discipline, less punishment, and more nurturing parenting. High internal consistency has been reported for the Nurturing (a = .82) and Disci- pline (a = .91) scales (Fox, 1994). For the present study, positive parenting scores ranged from 142 to 199 and showed good internal consistency (a = .79 at age 2, a = .77 at age 3, a = .74 at age 4).

566 Martinez-Torteya, Bogat, von Eye, and Levendosky

Child temperament. Temperament characteristics were assessed using the Carey Temperament Scales (97-item Toddler Temperament Scale at age 2, Fullard, McDevitt, & Carey, 1984; 100-item Behavioral Style Questionnaire at ages 3 and 4, McDevitt & Carey, 1978). Mothers rated their child’s attitudes and behavior on a 6-point scale. For this study, the rhythmicity, adaptability, approach, intensity, and mood scales were added to create an easy temperament variable (5–30; Saylor, Boyce, & Price, 2003). Adequate internal consistency has been reported for this composite (a = .87; Bogat et al., 2006). For the present study, scores ranged from 15 to 24 and showed good internal consistency (a = .85 at age 2, a = .85 at age 3, a = .81 at age 4). A dichotomous variable was also created based on the median score (19): Children whose score was greater than the median were classified as having easy temperament; those with scores lower than or equal to the median were classified as difficult.

Child cognitive ability. This was the only measure given at just one time point (age 4). Children’s cog- nitive ability was measured using the Peabody Pic- ture Vocabulary Test–Third Edition (PPVT–III; Dunn & Dunn, 1997). The PPVT–III is a 204-item measure of verbal ability for children ages 2.5 and up. Children are asked to select which of four black-and-white illustrations best represents a word read aloud by the interviewer. Age-normed scores (40–160) have a mean of 100 (SD = 10), and higher scores reflect better cognitive functioning. Good internal consistency (as = .93–.98.) and test–retest reliability (rs = .89–.97) have been reported (Dunn & Dunn, 1997). For the present sample, scores ranged from 55 to 132.

Risk Factors

Maternal depression. Mother’s symptoms of depression were assessed using the 21-item self- report Beck Depression Inventory (BDI; Beck, Men- delson, Mock, & Erbaugh, 1961). Participants select the best self-description from four evaluative state- ments with values from 0 to 3. A total score (0–63) is obtained; higher scores reflect more severe symp- toms. Good internal consistency (a = .86; Beck, Steer, & Garbin, 1988) has been reported. For the present study, scores ranged from 0 to 27 (a = .87 at age 2, a = .90 at age 3, a = .90 at age 4). A dichotomous score was used (Beck et al., 1988); scores greater than or equal to 10 reflect mild to severe depression, whereas scores less than 10 indicate no depression.

Stressful negative life events. Maternal stress asso- ciated with negative life changes was assessed using the 49-item Life Experiences Survey (LES; Sarason, Johnson, & Siegel, 1978). Respondents rated the occurrence and impact of 46 specific events (e.g., ‘‘Death of a close family member’’) during the last year on a 7-point scale, ranging from extremely negative ()3) to extremely positive (3). For the present study all the negative item ratings were summed ()1 to )3). Scores ranged from 0 to )147; lower scores indicate a greater degree of stress. Good test–retest reliability has been reported (r = .88; Sarason et al., 1978). For the present study, scores ranged from 0 to )25 (a = .60 at age 2, a = .68 at age 3, a = .71 at age 4).

Income and race. Participants reported their mean monthly family income at each time period, and the three scores were averaged into a total income score, ranging from $400 to $11,333. Mothers identified their child’s race (White ⁄ Caucasian, Black ⁄ African American, biracial ⁄ mixed, Native American, Asian American ⁄ Pacific Islander, and Latino ⁄ Hispanic ⁄ Chicano). Because of the small number of children in the last three groups as well as the large number of children in the biracial group, the racial grouping White or non-White was used in the analyses.

Procedures

Participants were recruited from a Midwestern state through flyers distributed to organizations serving women as well as flyers posted in the community. Interested pregnant women contacted the project office and were screened by a research assistant to determine eligibility: (a) 18 to 40 years old, (b) able to understand English well enough to complete interviews and questionnaires, and (c) involved in a romantic relationship for at least 6 weeks during pregnancy. Items 6 to 14 of the Conflict Tactics Scale (CTS; Straus, 1979) were administered during the telephone interview, after approximately half of the original sample had been recruited and interviewed (n = 96), in order to ensure that about 50% of the final sample experienced DV during pregnancy. Women were categorized as experiencing DV if they indicated that any of the CTS events occurred during pregnancy. One hundred and sixty-one women were excluded because they did not meet age, relationship status, or DV criteria; there were no demographic differences between these excluded women and participants. The final participants were

Child Resilience to Domestic Violence 567

demographically representative of the area where they were recruited.

All women were first interviewed for the larger study during their last trimester of pregnancy. For the present research, interviews occurred at the child’s second, third, and fourth birthdays. Mother and child were interviewed at the project offices and mothers completed all measures, except the child’s cognitive ability test. Interviews were con- ducted by graduate student project staff members and trained undergraduate research assistants. The DV questionnaires were administered last, to ensure that interviewers were blind to the woman’s abuse status for as long as possible. Interviews took about 2 to 3 hr to complete. Women were paid for their participation, and children received an age-appropriate gift.

Results

Missing Data and Data Reduction

Missing data were imputed using the Hot Deck method (LISREL; Jöreskog & Sörbom, 1996). One case could not be imputed through this procedure and was imputed using the expectation-maximiza- tion algorithm (EM algorithm; SYSTAT 11, 2004). Overall, only 8% of all data points were imputed. Missing data estimates for all the variables were based on total DV, maternal depression, and income at pregnancy, as well as child’s temperament at 2 months old. All transformations, composites, and dichotomizations of continuous scores were conducted after imputation.

Correlations between missingness dichotomous variables (missing data = 1, complete data = 0) and the original variables were either nonsignificant or small (r < .30), indicating a nonsystematic pattern of attrition. Additionally, when participants were classified into complete data versus missing data groups, there were no significant differences in total DV exposure, maternal depression, or income at pregnancy. Therefore, the imputed data set was used in all analyses.

Most predictor variables (except for race and cognitive ability) were measured at ages 2, 3, and 4, and all showed relative stability over time (rs = .29–.66). To attain more parsimonious models and avoid multicollinearity, average scores for all variables were used in analyses. The averaged predictors showed small- to medium-size associa- tions among themselves (rs = .10–.54) and with children’s internalizing and externalizing behav- iors (rs = .08–.57).

Hypothesis Testing

DV Exposure and the Odds of Resilience

To address the first hypothesis, DV and adapta- tion were cross-classified to obtain four groups of children: (a) resilient: exposed to DV and displayed positive adaptation (n = 62), (b) nonresilient: exposed to DV and displayed negative adaptation (n = 51), (c) competent: not exposed to DV and dis- played positive adaptation (n = 63), and (d) vulnera- ble: never exposed to DV and displayed negative adaptation (n = 14; see Table 1). Fifty-four percent of DV-exposed children displayed resilience, whereas 82% of nonexposed children showed positive adaptation. Odds ratio (OR) analysis of the four group frequencies showed a strong main effect for DV exposure, indicating that DV-exposed chil- dren were 3.7 times (95% confidence interval [CI] = 1.86, 7.36, d = .72) more likely to develop behavioral or emotional problems.

Adaptation and DV duration (zero to three time periods of exposure) were cross-classified to deter- mine the link between prolonged exposure and the likelihood of positive adaptation. Eight groups of children were obtained (see Table 1): four with positive adaptation (not exposed to DV, n = 63; exposed at one time period, n = 30; exposed at two time periods, n = 16; and exposed at three time periods, n = 16); and four with negative adaptation (not exposed to DV, n = 14; exposed at one time period, n = 18; exposed at two time periods, n = 14; and exposed at three time periods, n = 19). Positive adapation was achieved by 62% of children exposed during one time period, 55% exposed dur- ing two time periods, and 45% exposed during three time periods. OR analysis showed that children in all the DV-exposed groups were significantly more likely to show externalizing or

Table 1

Group Frequencies: Cross Classification of Adaptation by Domestic

Violence (DV) Exposure and DV Duration

Positive

adaptation

Negative

adaptation

No DV exposure 63 14

DV exposure 62 51

DV duration (No.

of time periods)

DV duration (No.

of time periods)

1 2 3 1 2 3

30 16 16 18 14 19

568 Martinez-Torteya, Bogat, von Eye, and Levendosky

internalizing problems, with effect sizes ranging from medium to large (one time exposure, OR = 2.7, 95% CI = 1.19, 6.15, d = .55; two times, OR = 3.94, 95% CI = 1.57, 9.90, d = .76; and three times, OR = 5.34, 95% CI = 2.12, 12.91, d = .93). However, among DV-exposed children, longer duration of DV exposure did not significantly increase the likelihood of negative adaptation (one vs. two times, OR = 1.46, 95% CI = .58, 3.68; two vs. three times, OR = 1.36, 95% CI = .51, 3.61; and one vs. three times, OR = 1.98, 95% CI = .82, 4.79).

The cross-classification of Adaptation · DV Duration suggested that the number of positively adapted children diminished with prolonged DV exposure, whereas the number of children with internalizing or externalizing problems remained fairly constant (see Table 2). Log-linear modeling was used to elucidate this trajectory. First, a main effects model (DV Duration · Adaptation) was esti- mated. The model showed significant lack of fit, LR v2(3, N = 190) = 16.98, p = .00, and indicated that a linear trajectory does not provide an accurate repre- sentation of the groups’ observed frequencies. Alternatively, a nonstandard model (Mair & von Eye, 2007) including three variables (the main effect of adaptation and independent DV duration main effects for positively adapted and negatively adapted children) was estimated, with high power (.93) to detect medium-sized effects (Gpower; Erdfelder, Faul, & Buchner, 1996).

The model showed adequate fit, LR v2(4, N = 190) = 3.64, p = .46, and Pearson v2(4, N = 190) = 3.69,

p = .45, and revealed significant main effects of DV duration for positively adapted children (z = )5.91, p = .00), but not for negatively adapted children (z = 0.61, p = .53). Thus, the number of positively adapted children significantly decreased and the number of negatively adapted children remained constant as the number of exposure periods increased.

Protective and Risk Factors

Multinomial logistic regression (Hosmer & Lem- eshow, 1989) was conducted (SPSS 14.0) to examine the second research hypothesis. Group status (resil- ient, nonresilient, competent, or vulnerable) was predicted using a main effects model with the pro- tective (i.e., positive parenting, easy temperament, and cognitive ability) and risk factors (i.e., maternal depression, stressful life events, low income, and race) entered in the first step, using the resilient group as the comparison group. Power for detect- ing medium-sized effects in this analysis was high (.95). The model was also estimated using family size as a covariate, to control for its influence on socioeconomic status (SES), and results remained unchanged.

The main effects model showed significantly better fit than the null model, LR v2(21, N = 190) = 131.24, p = .00, and the overall model fit was excellent, LR v2(546, N = 190) = 353.89, p = 1.00, and Pearson v2(546, N = 190) = 559.27, p = .34. Taken together, the main effects of all pre- dictors explained a large amount of variance in

Table 2

Descriptive Statistics by Group

Resilient

(n = 62)

Non resilient

(n = 51)

Competent

(n = 63)

Vulnerable

(n = 14)

Mean (SD) Mean (SD) Mean (SD) Mean (SD)

Race 47% White

53% non-White

35% White

65% non-White

60% White

40% non-White

36% White

64% non-White

Income 2281.10 (1130.49) 1862.71 (1193.22) 3266.18 (2328.31) 2891.27(2650.82)

Maternal Depression 5.68 (4.73) 9.90 (6.03) 3.16 (2.79) 10.40 (8.73)

Negative Life Events )6.21 (4.35) )7.83 (4.83) )3.08 (3.22) )5.31 (3.60) Positive Parenting 172.22 (10.50) 168.18 (11.95) 180.03 (11.67) 174.86 (13.99)

Easy Temperament 19.43 (1.43) 17.86 (1.32) 20.00 (1.51) 18.10 (1.41)

Cognitive Ability 93.35 (18.25) 90.94 (19.63) 101.32 (18.21) 95.14 (16.34)

DV Frequency 7.05 (11.04) 5.88 (6.18)

DV Duration 1.78 (0.83) 2.02 (0.86)

Internalizing 41.76 (5.36) 50.62 (5.35) 39.98 (4.81) 50.19 (5.46)

Externalizing 45.12 (5.50) 55.58 (5.68) 42.40 (5.03) 54.69 (6.15)

Note. Domestic violence (DV) Frequency and DV duration were not calculated for the non-exposed groups.

Child Resilience to Domestic Violence 569

group membership (Nagelkerke R2 = .54); how- ever, only negative life events, LR v2(3, N = 190) = 368.03, p = .00; maternal depression, LR v2(3, N = 190) = 377.69, p = .00; and child’s easy temperament LR v2(3, N = 190) = 394.33, p < .00, emerged as significant predictors (see Table 3). Pre- dictors showed small- to medium-sized effects. Children who exhibited higher levels of easy tem- perament (OR = .39, 95% CI = .26, .58, d = .52) and whose mothers reported less depression (OR = 1.14, 95% CI = 1.03, 1.25, d = .07) were more likely to be classified resilient than nonresilient. Additionally, resilient children experienced more stressful life events (OR = 1.22, 95% CI = 1.07, 1.40, d = .11) than competent nonexposed children. The three predictors also significantly discriminated resilient children from their vulnerable peers: Chil- dren in the vulnerable group experienced fewer stressful life events (OR = 1.26, 95% CI = 1.03, 1.55, d = .13), had more difficult temperaments (OR = .37, 95% CI = .21, .65, d = .55), and had mothers with higher levels of depression (OR = 1.26, 95% CI = 1.10, 1.44, d = .13). Impor- tantly, this group of children was not exposed to DV, but did encounter other environmental risks or vulnerabilities, such as more depressed mothers and more difficult temperaments.

DV duration and frequency could not be incorporated in the previous model because the nonexposed groups had zero variance. Therefore, binomial logistic regression was conducted only with DV-exposed children (n = 113). Race, income, positive parenting, easy temperament, cognitive ability, maternal depression, stressful life events, DV frequency, and DV duration were entered simultaneously in the first step of the model. Power for detecting medium-sized effects was adequate (.86). The main effects model

showed significantly better fit than the null model, LR v2(9, n = 113) = 47.22, p = .00, and had overall good fit, Hosmer–Lemeshow v2(8, n = 113) = 12.93, p = .11, explaining a large percentage of the variance (Nagelkerke R2 = .46). Maternal depression (OR = 1.14, 95% CI = 1.02, 1.28, d = .07) and easy temperament (OR = 0.46, 95% CI = 0.31, 0.68, d = .43) emerged again as significant predictors of resilience, with small to medium effect sizes. Income, race, DV duration, or average frequency of DV did not (see Table 4).

Configurations of DV Exposure, Protective, and Risk Factors

To test the association between specific trajectories of DV exposure (e.g., early vs. late exposure, chronic vs. intermittent exposure) and resilience, prediction configural frequency analysis

Table 3

Multinomial (Four Groups) Logistic Regression Predicting Resilience

Nonresilient Competent Vulnerable

B OR (CI) B OR (CI) B OR (CI)

Race .60 1.83 (0.69–4.84) ).11 .90 (0.38–2.13) 1.12 3.07 (.65–14.52) Income .00 1.00 (0.99–1.00) .00 1.00 (1.00–1.00) .00 1.00 (1.00–1.00)

Maternal depression .13* 1.14 (1.03–1.25) ).02 .98 (0.86–1.11) .23* 1.26 (1.10–1.44) Negative life events .02 1.02 (0.92–1.13) .20* 1.22 (1.07–1.40) .23* 1.26 (1.03–1.55)

Positive parenting .01 1.01 (0.97–1.06) .03 1.03 (0.99–1.08) .04 1.04 (0.98–1.11)

Easy temperament ).94* .39 (0.26–0.58) .07 1.07 (0.81–1.42) )1.00* .37 (0.21–0.65) Cognitive ability .02 1.02 (0.99–1.05) .01 1.01 (0.99–1.03) .02 1.02 (0.98–1.07)

Note. LR v2(546, N = 190) = 353.89, p = 1.00; Pearson v2(546, N = 190) = 559.27, p = .34; Nagelkerke R2 = .54; reference group: resilient. *p < .05.

Table 4

Binary (Two Group) Logistic Regression Predicting Resilience

Wald p OR CI

Race 2.05 .15 0.47 0.17–1.32

Income 2.57 .11 1.00 0.99–1.00

Depression 5.48 .02 1.14* 1.02–1.28

Negative life events 0.20 .66 1.03 0.92–1.15

Positive parenting 0.12 .73 1.01 0.96–1.06

Easy temperament 15.57 .00 0.46* 0.31–.68

Cognitive ability 1.10 .29 1.02 0.99–1.01

DV frequency 2.92 .09 0.94 0.88–1.01

DV duration 1.37 .24 1.46 0.77–2.74

Note. LR v2(9, n = 113) = 12.93, p = .11; Negelkerke R2 = .46; reference group: resilient. OR = odds ratio; CI = confidence interval; DV = domestic violence. *p < .05.

570 Martinez-Torteya, Bogat, von Eye, and Levendosky

(PCFA; von Eye, 2002) was used. PCFA is a cate- gorical data analysis technique that identifies types and antitypes, which indicate a relation between a specific configuration of predictors and the criterion (von Eye & Bogat, 2005). Types are those configura- tions that occur more often than expected by chance; antitypes are those that occur less often than expected by chance. The base model for PCFA is saturated both within the predictors and the criterion variables; thus, the first- and second-order autocorrelations between repeated measures (e.g., DV at ages 2, 3, and 4) are accounted for in the analysis.

DV status at ages 2, 3, and 4 was used to create predictor configurations (1 = no exposure, 2 = DV exposure); total adaptation was used as the criterion (1 = positive across all of the time peri- ods, 2 = negative at any time period). A 2 (DV at age 2) · 2 (DV at age 3) · 2 (DV at age 4) · 2 (adaptation) cross-classification yielded 16 differ- ent configurations; for example, the configuration 2221 describes a child who was exposed to DV during all three time periods and maintained positive adaptation. The Lehmacher Test with Bonferroni correction was used, which is appropriate for product-multinomial sampling. Deviations from the model (types or antitypes) were significant if p < .003. Power to detect medium effect sizes was adequate (.88).

The base model was not a good fit for the pat- tern of cell frequencies, LR v2(7, N = 190) = 23.23, p = .00; that is, the results cannot be accurately explained by the main effects or associations among the predictors; thus, types and ⁄ or antitypes are expected to emerge. Results indicated two types (1111 and 2222) and two antitypes (1112 and 2221; see Table 5). The first type, 1111, fo = 63, fe = 50.66, p = .00, represents children who were not exposed to violence at any time period and who exhibit positive adaptation. Its reciprocal anti- type, 1112, fo = 14, fe = 26.34, p = .00, portrays chil- dren who were never exposed to violence and developed negative adaptation. The second type, 2222, fo = 19, fe = 11.97, p = .00, represents children who were exposed to DV at all time periods and display negative adaptation, and its reciprocal an- titype, 2221, fo = 16, fe = 23.03, p = .00, describes children who were continuously exposed to DV and maintained resilience. Taken together, these configurations indicate that chronic DV exposure predicts the development of internalizing or exter- nalizing problems. As expected, children who were never exposed to DV are more likely to dis- play positive behavioral outcomes. However, no

other specific patterns of DV exposure (e.g., early exposure) showed specific associations with posi- tive or negative behavioral and emotional out- comes.

To identify characteristic configurations of pro- tective factors, risk factors, and DV exposure that are associated with resilience, a main effects confi- gural frequency analysis (CFA) model was tested among the DV-exposed participants (n = 113). CFA also identifies types and antitypes among variables, but the variables are not specified as predictors and criteria as in PCFA (von Eye, 2002). A first-order CFA base model assumes that the variables are not associated among themselves, and it accounts only for their individual main effects. Significant devia- tions from the model (types or antitypes) are obtained based on the comparison of observed and estimated frequencies and indicate second-order interactions among the variables.

Only the predictors that achieved significance in previous models were included in this analysis. Four categorical variables were included: maternal depression (1 = absent, 2 = present), easy tempera- ment (1 = low, 2 = high), DV duration (one to three time periods), and adaptation (1 = positive, 2 = neg- ative). These variables were cross-classified to obtain 24 patterns that represented children’s profile of risk ⁄ protective factors, duration of DV

Table 5

Prediction CFA: DV at Ages 2, 3, and 4, Predicting Adaptation

Predictors Criteria Prediction CFA

DV

2

DV

3

DV

4 Adaptation foijkl feijkl zijkl pijkl

N N N + 63 50.66 3.83 T .000

N N N ) 14 26.34 )3.83 A .000 N N Y + 13 10.53 1.36 .087

N N Y ) 3 5.47 )1.36 .087 N Y N + 6 5.92 0.06 .477

N Y N ) 3 3.08 )0.06 .477 N Y Y + 3 3.95 )0.83 .204 N Y Y ) 3 2.05 0.83 .204 Y N N + 11 15.13 )1.93 .027 Y N N ) 12 7.87 1.93 .027 Y N Y + 9 9.87 )0.49 .312 Y N Y ) 6 5.13 0.49 .312 Y Y N + 4 5.92 )1.38 .084 Y Y N ) 5 3.08 1.38 .084 Y Y Y + 16 23.03 )2.76 A .003 Y Y Y ) 19 11.97 2.76 T .003

Note. LR v2(7, N = 190) = 23.226, p = .002; Lehmacher test with Bonferroni-adjusted p = .003; fo = observed frequency; fe = expected frequency; T = Type; A = Antitype.

Child Resilience to Domestic Violence 571

exposure, and adaptation. For example, the pattern 1111 represents children whose mothers are not depressed, who show low levels of easy tempera- ment, who experienced DV at one time period, and who exhibit positive adaptation. The Lehm- acher test with Bonferroni correction was used to protect from Type I errors (p < .002). Because of the large number of configurations, this analysis’ power estimate (.43) was lower than desirable; thus, results can be considered a conservative estimate of the possible associations between the variables. Accordingly, types and antitypes that constitute smaller deviations from the main effects model might not have been identified as significant.

The main effects model was not a good fit for the pattern of cell frequencies, LR v2(18, n = 113) = 57.55, p = .00, which indicates local associations among the variables. Three types emerged: 1211, 1221, and 2132 (Table 6). The first configuration 1211, fo = 15, fe = 7.44, p = .00, describes a group of resilient children whose mothers are not depressed, who have easy temperaments, and who were

exposed to DV only during one time period. The second configuration, 1221, fo = 10, fe = 4.65, p = .00, describes a similar group of children who experienced DV at two time periods. Taken together, these two types indicate that absence of maternal depression and presence of child easy temperament are associated with resilience, but only among children with one or two time periods of DV exposure. The last configuration, 2132, fo = 9, fe = 2.65, p = .00, describes the group of nonresilient children whose mothers report high levels of depression, display difficult temperaments, and were exposed to DV during three time periods. This pattern suggests that chronic DV in the household is associated with maternal depression, more diffi- cult child temperament, and negative adaptation.

Discussion

The present study explored resilience among DV-exposed young children. The study identified a group of DV-exposed children who maintained

Table 6

CFA With DV-Exposed-Only Children: Depression, Temperament, DV Duration, and Adaptation

Cell index CFA

Depression Temperament Duration Adaptation foijkl feijkl zijkl pijkl

N ) 1 + 13 11.67 .52 .303 N ) 1 ) 7 9.60 1.07 .142 N ) 2 + 5 7.29 1.06 .145 N ) 2 ) 6 6.00 .00 .500 N ) 3 + 3 8.51 2.40 .008 N ) 3 ) 10 7.00 1.39 .082 N + 1 + 15 7.44 3.40 T .000

N + 1 ) 4 6.12 1.02 .154 N + 2 + 10 4.65 2.89 T .002

N + 2 ) 3 3.83 .48 .315 N + 3 + 6 5.43 .29 .385

N + 3 ) 0 4.46 2.44 .007 Y ) 1 + 2 4.41 1.33 .093 Y ) 1 ) 6 3.63 1.41 .080 Y ) 2 + 0 2.76 1.84 .033 Y ) 2 ) 3 2.27 .53 .300 Y ) 3 + 5 3.22 1.12 .132 Y ) 3 ) 9 2.65 4.31 T .000 Y + 1 + 0 2.81 1.85 .032

Y + 1 ) 1 2.31 .94 .173 Y + 2 + 1 1.76 .62 .269

Y + 2 ) 2 1.45 .50 .312 Y + 3 + 2 2.05 .04 .485

Y + 3 ) 0 1.69 1.40 .082

Note. LR v2(18, n = 113) = 57.554, p = .000; Lehmacher test with Bonferroni-adjusted p = .002. CFA = confirmatory factor analysis; DV = domestic violence; fo = observed frequency; fe = expected frequency; T = type.

572 Martinez-Torteya, Bogat, von Eye, and Levendosky

positive adaptation throughout ages 2–4. Fifty-four percent of DV-exposed children in this community sample were classified as resilient, which suggests that positive adaptation in the face of adversity is common and embedded in the processes of normal development (Masten, 2001). Results are consistent with previous studies with DV-exposed shelter- residing children (Grych et al., 2000; Hughes & Luke, 1998). However, DV exposure confers an almost fourfold increase (OR = 3.7) in the likelihood of internalizing or externalizing prob- lems. Similar results have been reported in previ- ous studies (e.g., Sternberg et al., 2006).

Examination of duration, frequency, and trajecto- ries of DV exposure as predictors of children’s behavioral ⁄ emotional outcomes provided mixed results. One-unit increases in DV duration were not associated with higher likelihood of internalizing and externalizing symptoms, as indicated by the odds-ratio analysis and the binary logistic regres- sion results. Similarly, frequency of DV exposure did not predict resilience, indicating that the main- tenance of positive adaptation is not a direct result of exposure to lower levels of adversity. However, log-linear modeling techniques and person- centered analyses (CFA) suggest a complex relation between DV duration and resilience: Prolonged exposure prevented the development of positive adaptation rather than directly increasing the likeli- hood of negative outcomes. That is, the number of positively adapted children decreased as the num- ber of time periods of exposure increased, but increases in DV duration were not associated with greater numbers of negatively adapted children. Findings underscore the heterogeneity of outcomes displayed by DV-exposed children and the impor- tance of using person oriented methods; a main effects model, which aggregates all DV-exposed children in the same group, did not provide an accurate explanation of the relation between duration of exposure and children’s adaptation.

Additionally, exploration of DV trajectories as predictors of resilience revealed that constant expo- sure to DV predicted the development of internaliz- ing or externalizing problems, whereas no specific configuration of exposure (e.g., single exposure, early exposure) was associated with resilience. Findings suggest that the experience of those chil- dren who are continuously under stress is qualita- tively different from that of children exposed to intermittent DV. Children who experience intermit- tent DV exposure might benefit from periods of less stress and of relatively higher quality family func- tioning. As pointed out by previous research with

high-risk populations, when children face continu- ous and severe environmental stressors, it is unlikely that they will be able to sustain resilient adaptation over time (Luthar & Zelazo, 2003). Children were only assessed during the early childhood period, and results should not be generalized to middle childhood or adolescence. Future research should compare the effects of exposure at these different developmental periods to determine whether there are potential discontinuities in children’s adaptation.

Findings identified specific individual and fam- ily factors that predicted children’s group status (resilient, nonresilient, competent, or vulnerable). The main effects model, which included income, race, positive parenting, child easy temperament, child cognitive ability, maternal depression, and stressful life events, was a good predictor of chil- dren’s group status. The combination of all predic- tors provided a better fit for the data than models that only included an isolated variable, underscor- ing the multiply-determined nature of children’s behavior. Overall, resilient children had more pro- tective factors and lower levels of risk as compared to their nonresilient peers; they differed from com- petent nonexposed children only in their higher levels of stressful life events. The similarity between both groups of positively adapted children is consistent with previous findings from longitudi- nal, person-oriented research (e.g., Masten et al., 1999). Interestingly, resilient children were exposed to lower levels of maternal depression and had eas- ier temperaments as compared to vulnerable chil- dren, which suggests that the vulnerable children were under considerable amounts of stress and had few resources to manage their environments, despite not being exposed to DV. Consistent with previous research (e.g., Masten et al., 1999), only a small percentage of nonexposed children were clas- sified as vulnerable.

Maternal depression and child’s easy tempera- ment emerged as significant predictors of resilience. These findings are consistent with previous research with other high-risk populations (e.g., Tschann et al., 1996), including DV-exposed chil- dren (Hughes & Luke, 1998). Mothers with good mental health are more likely to model appropriate responses to stressful events and help their children achieve healthy emotion regulation. On the other hand, depressed mothers are more likely to provide a role model of dysregulated responses to stress. DV-exposed children with easy temperaments are also more likely to display positive adaptation, as their low reactivity, high adaptability, and positive

Child Resilience to Domestic Violence 573

mood might promote the development of self- regulatory competence (Olson et al., 2002) and elicit positive responses from their caregivers and other adults (Rutter & Quinton, 1984).

Surprisingly, positive parenting, child cognitive ability, stressful life events, income, and race did not significantly predict children’s adaptation. Pre- vious research has shown that parental warmth and support predict positive adaptation (e.g., Katz & Gottman, 1997). However, the effects of positive parenting might have less impact in a chaotic envi- ronment, such as that of DV households, which may require stricter parenting to foster competence (see Baldwin, Baldwin, & Cole, 1990; Levendosky & Graham-Bermann, 2000). Measurement issues might also account for some of the discrepancies with previous studies. The present study used self- report measures, and a social desirability bias might have influenced maternal reports. Addition- ally, cognitive ability was not significantly associ- ated with resilience, which is inconsistent with some previous findings (e.g., Tiet et al., 1998). Inconsistencies with previous results might be asso- ciated with measurement differences, as the current research used the PPVT–III, which is strongly dependent on verbal cognitive ability.

In this research, exposure to stressful life events did not discriminate between resilient and nonresil- ient children; however, DV-exposed children (both resilient and nonresilient) had more stressful life events than did nonexposed children (both compe- tent and vulnerable). The association between DV exposure and other stressful life events has been previously reported (Eby, 2004), reflecting the dis- organized environments faced by abused women and their children. Additionally, the same pattern (worse outcomes for DV-exposed children) was also observed for positive parenting and cognitive ability, although differences did not achieve statisti- cal significance. This pattern is consistent with pre- vious findings on abused women’s parenting (e.g., Levendosky & Graham-Bermann, 2000) and DV- exposed children’s verbal cognitive ability (e.g., Huth-Bocks et al., 2001). Results suggest that these resources might be reduced by DV exposure, even within the group of children that achieve positive behavioral adaptation.

Racial background was not a significant predictor of resilience. However, in this research children were classified as White or non-White, which might have obscured the influence of belonging to a particular racial group (i.e., Black, Latino, Asian American, Native American, and biracial children were all aggregated in the non-White group).

Moreover, although race showed nonsignificant main effects in the present study, its moderating effects on other protective and risk factors require further exploration.

Consistent with contemporary conceptualizations of development and resilience (Masten, 2001), results suggest that resilience is associated with specific configurations of adversity (DV exposure), protective, and risk factors. Among DV-exposed children, the combination of child easy tempera- ment and absence of maternal depression was asso- ciated with positive outcomes after one or two time periods of exposure. The association between easy temperament and absence of maternal depression reflects a competent mother–child dyad in which the young child effectively obtains what she or he needs from a responsive mother and is protected against the effects of intermittent DV. Conversely, a distinguishable group exposed to cumulative adversity was identified. More children than expected by chance exhibited co-occurring maternal depression, negative temperament, chronic DV, and negative behavioral adaptation. This configuration suggests that chronic DV is associated with other risk factors at the individual and family levels, and the interaction of chronic adversity and impover- ished resources disrupts the patterns of normal adaptation.

There are some limitations to this study. First, all predictors were assessed through maternal reports, except for child’s cognitive ability. Although there is some evidence that depressed mothers do not consistently overreport their children’s behavioral problems (Richters, 1992), results should be inter- preted with caution because maternal reports might have been influenced by maternal characteristics or social desirability biases. Unfortunately, because of the young age of the child participants, it would have been difficult and unreliable to obtain their own reports of DV exposure, environmental characteristics, or behavior ⁄ well-being. Future research would benefit from a multi-informant, multimethod approach.

Additionally, despite the high retention rate of this research, excluded participants had signifi- cantly higher levels of depression, more frequent DV exposure, and lower family income during pregnancy. Therefore, results might not generalize to women who are experiencing frequent violence and high levels of adversity.

The definition of resilience for the present study was based on behavioral and emotional outcomes. Measures of positive adaptation (e.g., achieve- ment of developmental milestones, quality of peer

574 Martinez-Torteya, Bogat, von Eye, and Levendosky

relations, school achievement) would provide a comprehensive assessment of resilience. Moreover, all predictors were measured as continuous vari- ables, such that high levels of a given predictor might be considered protective (e.g., positive par- enting), whereas low levels of the same predictor reflect risk (e.g., harsh parenting). This approach does not allow for identification of purely protec- tive mechanisms and might not accurately repre- sent the effects of variables that only exert influence at one end of the continuum or that have nonlinear associations with adaptation (Luthar et al., 2006).

Lastly, the sample size placed a constraint on the statistical analyses. For example, it was not possible to test for differences associated with the develop- mental period in which the protection ⁄ risk occurred or the different patterns (increasing vs. decreasing) of exposure or resources. Similarly, testing for interactions between risk ⁄ protection and specific racial backgrounds was not possible within the current sample. Additionally, this research was not able to examine trajectories of behavioral out- comes and the effects of protective factors on changes (e.g., continuity or discontinuity) in adap- tation. Growth curve modeling techniques or group-based semiparametric models are especially suited to address these important questions. Longi- tudinal studies with large samples of DV-exposed children are needed.

In summary, the present article contributes to the current understanding of the effects of DV exposure on children’s development as well as the characteristics of young children who maintain resilience in the context of intimate partner violence directed toward their mothers. Findings indicate that resilience in the face of DV exposure does occur and is associated with a configuration of co- occurring protection and lower risk, particularly when DV exposure is limited. Future research should address the stability of DV-exposed chil- dren’s positive adaptation as well as the role of these individual and family characteristics as pro- tective or risk factors at different developmental stages.

References

Achenbach, T. M. (1991). Manual for the Child Behavior Checklist ⁄ 4–18 and 1991 Profile, Burlington: University of Vermont.

Achenbach, T. M. (1992). Manual for the Child Behavior Checklist ⁄ 2–3 and 1992 Profile, Burlington: University of Vermont.

Baldwin, A. L., Baldwin, C., & Cole, R. E. (1990). Stress- resistant families and stress-resistant children. In J. E. Rolf, A. S. Masten, D. Cicchetti, K. H. Nuechterlein, & S. Weintraub (Eds.), Risk and protective factors in the development of psychopathology (pp. 257–280). New York: Cambridge University Press.

Beck, A. T., Mendelson, M., Mock, J., & Erbaugh, J. (1961). An inventory for measuring depression. Archives of General Psychiatry, 4, 561–571.

Beck, A. T., Steer, R. A., & Garbin, M. G. (1988). Psycho- metric properties of the Beck Depression Inventory: Twenty-five years of evaluation. Clinical Psychology Review, 8, 77–100.

Bergman, L. R., & Magnusson, D. (1997). A person- oriented approach in research on developmental psychopathology. Development and Psychopathology, 9, 291–319.

Bogat, G. A., DeJonghe, E. S., Levendosky, A. A., David- son, W. S., & von Eye, A. (2006). Trauma symptoms among infants who witness domestic violence toward their mothers. Child Abuse & Neglect: The International Journal, 30, 109–125.

Bogat, G. A., Levendosky, A. A., & Davidson, W. S. (1999). Understanding the intergenerational transmission of violence: From pregnancy through the first year of life. Rockville, MD: National Institutes of Justice.

Bogat, G. A., Levendosky, A. A., & von Eye, A. (2005). The future of research on intimate partner violence: Person-oriented and variable-oriented perspectives. American Journal of Community Psychology, 36, 49–70.

Bradley, R. H., & Corwyn, R. F. (2002). Socioeconomic status and child development. Annual Review of Psychol- ogy, 53, 371–399.

Buckner, J. C., Mezzacappa, E., & Beardslee, W. R. (2003). Characteristics of resilient youths living in poverty: The role of self-regulatory processes. Development and Psy- chopathology, 15, 139–162.

Cole, P. M., Zahn-Waxler, C., Fox, N. A., Usher, B. A., & Welsh, J. D. (1996). Individual differences in emotion regulation and behavior problems in preschool chil- dren. Journal of Abnormal Psychology, 105, 518–529.

Compas, B. E., Connor-Smith, J., & Jaser, S. S. (2004). Temperament, stress reactivity, and coping: Implica- tions for depression in childhood and adolescence. Journal of Clinical Child and Adolescent Psychology, 33, 21–31.

Davies, P. T., Winter, M. A., & Cicchetti, D. (2006). The implications of emotional security theory for under- standing and treating childhood psychopathology. Development and Psychopathology, 18, 707–735.

DeJonghe, E. S., Bogat, G. A., Levendosky, A. A., von Eye, A., & Davidson, W. S. (2005). Infant exposure to domestic violence predicts heightened sensitivity to adult verbal conflict. Infant Mental Health Journal, 26, 268–281.

Dunn, L. M., & Dunn, L. M. (1997). Peabody Picture Vocab- ulary Test (3rd ed.). Circle Pines, MN: American Guid- ance Service.

Child Resilience to Domestic Violence 575

Eby, K. K. (2004). Exploring the stressors of low-income women with abusive partners: Understanding their needs and developing effective community responses. Journal of Family Violence, 19, 221–232.

Erdfelder, E., Faul, F., & Buchner, A. (1996). GPOWER: A general power analysis program. Behavior Research Methods, Instruments, & Computers, 28, 1–11.

Fox, R. A. (1994). Parent Behavior Checklist, Brandon, VT: Clinical Psychology Publishing.

Fullard, W., McDevitt, S. C., & Carey, W. B. (1984). Assessing temperament in one- to three-year-old chil- dren. Journal of Pediatric Psychology, 9, 205–217.

Garmezy, N. (1993). Risk and resilience. In D. C. Funder, R. D. Parke, C. Tomlinson-Keasey & K. Widaman (Eds.), Studying lives through time: Personality and devel- opment (pp. 377–398). Washington, DC: American Psy- chological Association.

Gerard, J. M., & Buehler, C. (2004). Cumulative environ- mental risk and youth maladjustment: The role of youth attributes. Child Development, 75, 1832–1849.

Grych, J. H., Jouriles, E. N., Swank, P. R., McDonald, R., & Norwood, W. D. (2000). Patterns of adjustment among children of battered women. Journal of Consult- ing and Clinical Psychology, 68, 84–94.

Hosmer, D. W., & Lemeshow, S. (1989). Applied logistic regression. New York: Wiley.

Hughes, H. M., & Luke, D. A. (1998). Heterogeneity in adjustment among children of battered women. in G. W. Holden, R. Geffner, & E. N. Jouriles (Eds.), Children exposed to marital violence: Theory, research, and applied issues (pp. 185–221). Washington, DC: American Psy- chological Association.

Huth-Bocks, A. C., Levendosky, A. A., & Semel, M. A. (2001). The direct and indirect effects of domestic vio- lence on young children’s intellectual functioning. Jour- nal of Family Violence, 16, 269–290.

Jaffee, S. R. (2005). Family violence and parent psycho- pathology: Implications for children’s socioemotional development and resilience. In S. Goldstein & R. B. Brooks (Eds.), Handbook of resilience in children (pp. 397–400). New York: Kluwer Academic ⁄ Plenum Publishers.

Jaffee, S. R., Caspi, A., Moffitt, T. E., Polo-Tomas, M., & Taylor, A. (2007). Individual, family, and neighborhood factors distinguish resilient from non-resilient mal- treated children: A cumulative stressors model. Child Abuse and Neglect, 31, 231–253.

Jöreskog, K. G., & Sörbom, D. (1996) LISREL 8: User’s ref- erence guide. Chicago: Scientific Software International.

Katz, L. F., & Gottman, J. M. (1997). Buffering children from marital conflict and dissolution. Journal of Clinical Child Psychology, 26, 157–171.

Kim-Cohen, J., Moffitt, T. E., Caspi, A., & Taylor, A. (2004). Genetic and environmental processes in young children’s resilience and risk to socioeconomic depriva- tion. Child Development, 75, 651–668.

Kitzmann, K. M., Gaylord, N. K., Holt, A. R., & Kenny, E. D. (2003). Child witnesses to domestic violence: A

meta-analytic review. Journal of Consulting and Clinical Psychology, 71, 339–352.

Kliewer, W., Cunningham, J. N., Diehl, R., Parrish, K. A. Walker., J. M., Atiyeh, C., et al. (2004). Violence expo- sure and adjustment in inner-city youth: Child and caregiver emotion regulation skill caregiver-child rela- tionship quality, and neighborhood cohesion as protec- tive factors. Journal of Clinical Child and Adolescent Psychology, 33, 477–487.

Kolbo, J. R. (1996). Risk and resilience among children exposed to family violence. Violence and Victims, 11, 113–128.

Leech, S. L., Larkby, C. A., Day, R., & Day, N. L. (2006). Predictors and correlates of high levels of depression and anxiety symptoms among children at age 10. Jour- nal of the American Academy of Child and Adolescent Psy- chiatry, 45, 223–230.

Levendosky, A. A., Bogat, G. A., Davidson, W. S., & von Eye, A. (2000). Risk and protective factors for domestic vio- lence. Atlanta, GA: Centers for Disease Control, Injury Prevention Center.

Levendosky, A. A., Bogat, G. A., & von Eye, A. (2007). New directions for research on intimate partner vio- lence and children. European Psychologist, 12, 1–5.

Levendosky, A. A., & Graham-Bermann, S. A. (2000). Trauma and parenting in battered women: An addition to an ecological model of parenting. Journal of Aggres- sion, Maltreatment & Trauma, 3, 25–35.

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, 275–287.

Levendosky, A. A., Leahy, K. L., Bogat, G. A., Davidson, W. S., & von Eye, A. (2006). Domestic violence, maternal parenting, maternal mental health, and infant externalizing behavior. Journal of Family Psychology, 20, 544–552.

Luthar, S. S., Cicchetti, D., & Becker, B. (2000). The con- struct of resilience: A critical evaluation and guidelines for future work. Child Development, 71, 543–562.

Luthar, S. S., Sawyer, J. A., & Brown, P. J. (2006). Concep- tual issues in the studies of resilience—Past, present, and future research. Annals of the New York Academy of Sciences, 1095, 105–115.

Luthar, S. S., & Zelazo, L. B. (2003). Research on resil- ience: An integrative review. In S. S. Luthar (Ed.), Resilience and risk: Adaptation in the context of childhood adversities (pp. 510–549). New York: Cambridge Univer- sity Press.

Mair, P., & von Eye, A. (2007). Application scenarios for nonstandard log-linear models. Psychological Methods, 12, 139–156.

Malcarne, V. L., Hamilton, N. A., Ingram, R. E., & Taylor, L. (2000). Correlates of distress in children at risk for affective disorder: Exploring predictors in the offspring of depressed and nondepressed mothers. Journal of Affective Disorders, 59, 243–251.

576 Martinez-Torteya, Bogat, von Eye, and Levendosky

Marshall, L. L. (1992). Development of the Severity of Violence Against Women Scales. Journal of Family Vio- lence, 7, 103–121.

Masten, A. S. (2001). Ordinary magic: Resilience processes in development. American Psychologist, 56, 227–238.

Masten, A. S., Hubbard, J. J., Gest, S. D., Tellegen, A., Garmezy, N., & Ramirez, M. (1999). Competence in the context of adversity: Pathways to resilience and malad- aptation from childhood to late adolescence. Develop- ment and Psychopathology, 11, 143–169.

Masten, A. S., & Obradovic, J. (2006). Competence and resilience in development. Annals of the New York Acad- emy of Sciences, 1094, 13–27.

McDevitt, S. C., & Carey, W. B. (1978). The measurement of temperament in 3–7 year old children. Journal of Child Psychology and Psychiatry, 19, 245–253.

McLoyd, V. C. (1998). Socioeconomic disadvantage and child development. American Psychologist, 53, 185–204.

Olson, S. L., Bates, J. E., Sandy, J. M., & Schilling, E. M. (2002). Early developmental precursors of impulsive and inattentive behavior: From infancy to middle child- hood. Journal of Child Psychology and Psychiatry, 43, 435– 448.

Richters, J. E. (1992). Depressed mothers as informants about their children: A critical review of the evidence for distortion. Psychological Bulletin, 112, 485–499.

Rutter, M. (1979). Protective factors in children’s responses to stress and disadvantage. in M. W. Kent & J. E. Rolf (Eds.), Primary prevention of psychopathology: Vol. 3. Social competence in children (pp. 49–74). Hanover, NH: University Press of New England.

Rutter, M. (2006). Implications of resilience concepts for scientific understanding. Annals of the New York Acad- emy of Sciences, 1094, 1–12.

Rutter, M., & Quinton, D. (1984). Parental psychiatric dis- order: Effects on children. Psychological Medicine, 14, 853–880.

Sarason, I. G., Johnson, J. H., & Siegel, J. M. (1978). Assessing the impact of life changes: Development of the Life Experiences Survey. Journal of Consulting and Clinical Psychology, 46, 932–946.

Saylor, C. F., Boyce, G. C., & Price, C. (2003). Early pre- dictors of school-age behavior problems and social skills in children with intraventricular hemorrhage (IVH) and ⁄ or extremely low birth weight (ELBW). Child Psychiatry & Human Development, 33, 175–192.

Shannon, K. E., Beauchaine, T. P., Brenner, S. L., Neuhaus, E., & Gatzke-Kopp, L. (2007). Familial and temperamental predictors of resilience in children at

risk for conduct disorder and depression. Development and Psychopathology, 19, 701–727.

Smith, J., & & Prior, M. (1995). Temperament and stress resilience in school-age children: A within-families study. Journal of the American Academy of Child & Adoles- cent Psychiatry, 34, 168–179.

Steinhausen, H. C., & Metzke, C. W. (2001). Risk, com- pensatory, vulnerability, and protective factors influ- encing mental health in adolescence. Journal of Youth and Adolescence, 30, 259–280.

Sternberg, K. J., Baradaran, L. P., Abbott, C. B., Lamb, M. E., & Guterman, E. E. (2006). Type of violence, age, and gender differences in the effects of family violence on children’s behavior problems: A mega-analysis. Developmental Review, 26, 89–112.

Straus, M. A. (1979). Measuring intrafamily conflict and violence: The Conflict Tactics (CT) Scales. Journal of Marriage & the Family, 41, 75–88.

Straus, M. A., Hamby, S. L., Boney-McCoy, S., & Sugar- man, D. B. (1996). The revised Conflict Tactics Scales (CTS2)—Development and preliminary psychometric data. Journal of Family Issues, 17, 283–316.

Thomas, A., & Chess, S. (1985). The behavioral study of temperament. In J. Strelau, F. H. Farley, & A. Gale (Eds.), The biological bases of personality and behavior: Vol. 1. Theories, measurement techniques, and development (pp. 213–225). New York: Hemisphere Publishing ⁄ Harper & Row.

Tiet, Q. Q., Bird, H. R., Davies, M., Hoven, C., Cohen, P., Jensen, P. S., et al. (1998). Adverse life events and resil- ience. Journal of the American Academy of Child & Adoles- cent Psychiatry, 37, 1191–1200.

Tschann, J. M., Kaiser, P., Chesney, M. A., Alkon, A., & Boyce, G. C. (1996). Resilience and risk among preschool children: Family functioning, temperament, and behavior problems. Journal of the American Academy of Child & Adolescent Psychiatry, 35, 184–192.

von Eye, A. (2002). Configural frequency analysis: Methods, models, and applications. Mahwah, NJ: Erlbaum.

von Eye, A., & Bogat, G. A. (2005). Logistic regression and prediction configural frequency analysis: A com- parison. Psychology Science, 47, 326–341.

Wyman, P. A., Cowen, E. L., Work, W. C., Hoyt-Meyers, L., Magnus, K. B., & Fagen, D. B. (1999). Caregiving and developmental factors differentiating young at-risk urban children showing resilient versus stress-affected outcomes: A replication and extension. Child Develop- ment, 70, 645–659.