Adolescents' Cognitive and Emotional Responses to Marital Hostility

Child Development, May/June 2007, Volume 78, Number 3, Pages 775 – 789

Adolescents’ Cognitive and Emotional Responses to Marital Hostility

Cheryl Buehler and Garrett Lange Karen L. Franck University of North Carolina University of Tennessee

Early adolescents’ (11 – 14 years) responses to marital hostility were examined in a sample of 416 families. The cognitive-contextual perspective and emotional security hypothesis guided the study and 9 adolescent re- sponses were identified. Prospective associations were examined in several structural equation models that included adolescent problems as outcomes. Self-blame and perceived threat uniquely mediated the association between Year 1 marital hostility and Year 3 adolescent externalizing problems (po.05). Self-blame, lower con- structive representations, internalization of feelings, avoidance, and emotional dysregulation uniquely medi- ated the association between Year 1 marital hostility and Year 3 internalizing problems. Specific cognitive and emotionally based responses are important to understanding how martial hostility affects youth and need to be considered within an integrated model.

Marital hostility places children and adolescents marital hostility. The study examines early adoles- at risk for concurrent and prospective psychosocial cents’ responses to marital hostility in a sample of difficulties (Buehler et al., 1997; Davies, Harold, 416 families using a three-wave, prospective, multi- Goeke-Morey, & Cummings, 2002; Doyle & Mark- method research design. iewicz, 2005; Grych & Fincham, 1990). In addition to Early adolescence is an important juncture for this disrupting parenting (Buehler, Benson, & Gerard, examination because youth are transforming their 2006) and children’s psychophysiological function- relationships with parents and peers (Steinberg, ing (Katz, 2001), recent advances in understanding 2001). Marital hostility that occurs during this the explanatory mechanisms of this risk factor have developmental transition might create additional focused on two complementary explanations: (a) the demands on the youth because they often are role of children’s potentially problematic cognitive compelled to devote psychological resources to appraisals (Grych & Fincham, 1993) and (b) the role processing parents’ disputes. This diversion of psy- of children’s emotional insecurity (Davies et al., chological resources creates potential vulnerabilities 2002). because youth in early adolescence are also experi- The present study addressed these two explana- encing changes in physical development, schools, tions by examining the conjoint and integrative roles and social networks. This is a demanding period of of cognitive appraisals and emotional insecurity development and the concurrent experience of during early adolescence. A conjoint explanation marital hostility inhibits youth from finding refuge suggests that when considered in the same model, within the family domain. Call and Mortimer (2001) problematic cognitive appraisals and emotional in- suggest that the family domain is a potentially im- security regarding marital hostility each have unique portant ‘‘arena of comfort’’ for adolescents under- associations with adolescent problem behavior. An going a multitude of change and marital hostility integrative explanation suggests that specific aspects might endanger some of the salutary benefits of of the cognitive and emotional response systems family life. Early adolescence is also an important overlap, and that once this overlap is considered, time to examine youths’ perceptions regarding cognitively and emotionally oriented responses are marital hostility because they are exploring their both needed to understand youths’ processing of own identity as a partner in close friendships and romantic pairings (Parke et al., 2001). Youths’ per- ceptions regarding parents’ interactions might shape This research was supported by a grant from The National In- their relational templates that serve as interpretive stitute of Mental Health, R01-MH59248. We thank the staff of the schemas for developing and maintaining close rela- Family Life Project for their unending contributions to this work tionships during adolescence and early adulthood and the youth, parents, teachers, and school administrators who (Linder & Collins, 2005). made this research possible. Correspondence concerning this article should be addressed to Cheryl Buehler, Human Development and Family Studies, UNCG, Greensboro, NC 27402-6170. Electronic mail may be sent to r 2007 by the Society for Research in Child Development, Inc. [email protected]. All rights reserved. 0009-3920/2007/7803-0008 776 Buehler, Lange, and Franck

Theoretical Foundation framework, the emotional security hypothesis sug- gests that youths’ responses involve both emotion Cognitive Appraisals and cognition. Davies and Cummings (1994) theorize The potential role of cognitive appraisals in that ‘‘when conflicts are appraised as destructive and shaping children’s responses to witnessing marital reflective of significant marital disharmony, chil- hostility is detailed theoretically in the cognitive- dren’s negative emotional arousal becomes elevated contextual framework (Grych & Fincham, 1990). In and they are motivated to act to decrease their their original formulations, Grych and Fincham use feelings of emotional insecurity’’ (p. 389). work by Campos, Campos, and Barrett (1989) to In their elaboration of the emotional security hy- define marital hostility as a ‘‘significant event’’ for pothesis, Davies and Cummings (1998) outline how children and youth. Defining marital hostility as a this process mechanism is consistent with a func- significant event suggests that hostility between tionalist perspective of emotion that considers feel- parents is relevant to youth, involves emotional ings, response motivations, and appraisals (Campos, communication among significant others, and is Mumme, Kermoian, & Campos, 1994). They propose arousing (Campos et al., 1989). Grych and Fincham three process components: emotional reactivity, propose that youths’ evaluation of the significance behavioral regulation, and internal representations of interparental hostility involves both emotion and (Davies & Cummings, 1994, 1998). Emotional reac- cognition. Various emotional responses shape the tivity focuses on the feeling manifestation of insecu- appraisals of threat and prompt coping behavior. rityFspecifically fear and distress. Regulation Appraisals also shape affect and coping behaviors. focuses on the behavioral manifestation of insecurity We suggest that coping, in this case, often deals with by including avoidant and involvement coping be- emotional arousal, trying to reduce the marital hos- haviors. Internal representations focus on the ap- tility itself, and/or minimizing exposure to the hos- praisal manifestation of insecurity by including tility (Lazarus & Folkman, 1984). youths’ concerns of parental separation and the Cognitive responses involve appraisals of the spillover of hostility into parent – child relations. hostile interactions (e.g., how bad is it, will parents Theoretically, these three components are interde- contain it, whose fault is it) and estimations of one’s pendent, but distinct, indicators of youths’ need for ability to handle a potentially stressful situation. The emotional security within the context of interparen- cognitive-contextual framework suggests that three tal hostility. types of cognitions are particularly salient: perceived threat, self-blame, and coping efficacy. Conceptually, A Comparative Framework perceived threat focuses on specific fears and wor- ries regarding parents getting hurt, parents separat- Clearly, these two perspectives regarding how ing, and the spillover of hostility into parent – child marital hostility affects youth are complementary. relations. Self-blame focuses on attributions regard- Both view emotion and cognition as being interde- ing the youths’ cause of or responsibility for the pendent, integral elements of youths’ responses to conflict and parental distress. Coping efficacy fo- marital hostility. For example, Grych and Fincham cuses on the youths’ appraisals of their ability to stop (1990) acknowledge the role of emotion-focused re- parents’ hostile interactions or manage their own sponses by theorizing that ‘‘affect is involved in distress. Theoretically, these three appraisals are in- evaluating the significance of the conflict and in terrelated and Grych and Fincham (1990) speculate guiding subsequent behavior’’ (p. 241). Davies et al. that they shape coping responses, particularly (2002) detail some of the interdependencies when whether to get involved in the dispute. stating ‘‘the emotional security hypothesis posits that preserving a sense of security is an important goal that organizes a child’s emotional experiences (e.g., Emotional Security fear), action tendencies (e.g., withdraw, intervene), The emotional security hypothesis suggests that and appraisals of self and interpersonal relationships youths’ responses to marital hostility focus on the (e.g., perceptions of threat to the self)’’ (p. 6). Both implications of the acrimony for their own emotional begin their theorizing regarding the effects of the security (Davies & Cummings, 1994). This hypothe- negative aspects of marital conflict (e.g., hostile, sis proposes that the attainment of emotional secu- frequent, unresolved) when youth are distressed, rity is shaped, in part, by parents’ interactions with aware, and upset. Both describe perceived threat as a one another and the quality of the marital relation- motivational force that compels a behavioral re- ship. Consistent with the cognitive-contextual sponse. Both also define threat in terms of broader Adolescents’ Responses to Marital Hostility 777 implications for concerns regarding marital insta- between the prefrontal cortex that is associated with bility and parent – child difficulties. Both consider attentional processes and working memory and the coping as an integral aspect of the response system. subcortical limbic structures that are associated with A major difference between the two perspectives emotion. Bell and Wolfe (2004) also provide a recent is that the cognitive-contextual framework accentu- summary of developmental neuroscience studies ates the appraisal aspects of the process mechanisms suggesting that the neural mechanisms influencing (Grych & Fincham, 1990) and the emotional security emotional regulation might be the same as those that hypothesis highlights the defining role of emotions influence volitional cognitive processes. and emotion-focused responses (Davies & Cum- mings, 1994). This difference in emphasis centers, in Empirical Contributions part, on the motivational aspects of arousal and threat. The emotional security hypothesis stresses Cognitive Appraisals the importance of preserving emotional security in the face of a potentially disruptive stressor in the Cross-sectional research has demonstrated that family system. The cognitive-contextual perspective the cognitive appraisals of perceived threat, self- stresses the importance of meaning making in blame, and coping efficacy help explain the associ- shaping youths’ immediate and longer term re- ation between marital hostility and youth problem sponses to marital hostility. As stated by Davies et al. behavior, particularly internalizing problems (2002), ‘‘The cognitive-contextual framework differs (Dadds, Atkinson, Turner, Blums, & Lendich, 1999; from the emotional security hypothesis in placing Gerard, Buehler, Franck, & Anderson, 2005; Grych, heavier emphasis on understanding how the cogni- Fincham, Jouriles, & McDonald, 2000). The theoreti- tive dimensions of the child’s appraisals shape the cal assertion that these three appraisals are related, impact of conflict on child adjustment’’ (p. 17). but distinct elements of the process mechanism has Theoretically, the idea of a complex, integrated been supported across these studies. Perceived cognitive and emotional response system is not threat seems to be a particularly salient aspect of this unique to the literature on youths’ responses to marital process mechanism (Gerard et al., 2005). conflict. Recent researchers of development in early There also is evidence that cognitive appraisals and middle childhood have argued that emotion and partially mediate the prospective association be- cognition are inextricably linked and function to- tween marital hostility and early adolescent problem gether in a dynamic manner to organize and direct behavior (Grych, Harold, & Miles, 2003). Grych et al. information processing and action execution (Bell & studied 298 two-parent families in the United King- Wolfe, 2004). Circumstantial evidence for this view is dom (youth average age was 11.67 at Time 1 (T1) and found in the recent literature showing systematic 91% had two biological parents). At T1, they as- relations between emotional and cognitive processes sessed youths’ and parents’ perceptions of marital and developmental outcomes. For example, research conflict, youth conflict appraisals of threat and self- by Nigg and Huang-Pollock (2003) on early child- blame, and youth problem behavior. A year later hood psychopathology has shown that young chil- they assessed conflict appraisals and youth problem dren’s adjustment difficulties are often characterized behavior. Youth and parent reports of marital conflict by parallel deficits in both cognitive and emotion were analyzed in separate models, as were youth processes. Moreover, Blair (2002) elucidated the re- internalizing and externalizing problem behavior. ciprocal nature of cognitive and emotion processes Perceived threat partially mediated the 1-year, pro- for children’s behavioral self-regulation such that spective association between youths’ and parents’ emotion plays a functional role in organizing and reports of marital conflict and youth internalizing directing cognition. He also argued that negative problems. Conversely, self-blame partially mediated emotionality has a disruptive role for adults’ higher the 1-year, prospective association between youths’ order cognitive processes of attention, memory, ex- and parents’ reports of marital conflict and exter- ecutive function, and goal-directed problem-solving nalizing problems. activity (also see Mathews & Wells, 1999; Mogg & These findings suggest that the conflict appraisals Bradley, 1999). Perhaps the most compelling argu- of threat and self-blame helped explain part of the ment for examining emotion and cognition together prospective association between marital hostility comes from the neuroscience literature. Blair (2002) and future problem behavior during early adoles- describes numerous neuroanatomical investigations cence. They also suggest some specialized effects in (e.g., Derryberry & Tucker, 1994) that have provided that perceived threat was associated with internal- evidence of neural pathway interconnections izing problems and self-blame with externalizing 778 Buehler, Lange, and Franck problems. However, the examination was limited in conflict and adolescent problem behavior were as- that internalizing and externalizing problems were sessed at T1 and cognitive appraisals, emotional in- analyzed in separate models, and therefore, unique security, and problem behavior were assessed 1 year associations could not be isolated. In addition, the later. Youth emotional insecurity completely medi- measure of perceived threat combined the theoretical ated the association between interparental conflict components of perceived threat and coping efficacy, and increases in internalizing and externalizing precluding an empirical examination of the unique problems over the 1 year. Self-blame was not a me- contributions of threat and efficacy appraisals. diator because T1 interparental conflict was not as- sociated with T2 self-blame. Perceived threat was not a mediator because T2 perceived threat was not as- Emotional Security sociated with changes in T2 adolescent problems. As with cognitive appraisals, components of Although these findings suggest that emotional emotional insecurity mediate the contemporaneous insecurity might be a unique mediator when con- association between marital hostility and adolescent sidered conjointly with cognitive appraisals, three problem behavior (Davies & Cummings, 1998). The aspects of the study limit this conclusion. First, al- theoretical contention that the components are in- though emotional insecurity was modeled as a latent terrelated but distinct aspects of emotional insecurity construct, perceived threat and self-blame were in- was supported. Emotional reactivity and negative cluded as separate manifest variables. This might internal representations were particularly salient have fragmented the role of cognitive appraisals components and associations with both internalizing when contrasting the unique contributions of ap- and externalizing problems were partially explained. praisals and emotional insecurity. Second, internal- These findings demonstrated some specialized and izing and externalizing problems were modeled unique associations because both internalizing and separately and so the unique and specialized pro- externalizing problems were included in the same cesses for explaining interparental conflict processes analytic model. could not be addressed. Finally, some of the over- There also is evidence that emotional insecurity lapping items from the measures of cognitive ap- partially mediates the prospective association be- praisals and emotional security were not taken into tween marital hostility and early adolescent problem account. behavior (Harold, Shelton, Goeke-Morey, & Cum- The present study builds on these studies of U.K. mings, 2004). Harold and associates studied 181 families by (a) examining these theoretical proposi- families in the United Kingdom the (average age of tions in a sample of families from the United States, youth was 11.65 years and 87% lived with both bio- (b) including the cognitive appraisal of coping effi- logical parents). Marital conflict (parent reports) and cacy, (c) examining internalizing and externalizing emotional reactivity, behavioral regulation, and in- problem behavior in the same model to control for ternal representations (youth interview) were as- their moderate association and identify specialized sessed at T1. One year later, emotional security about and unique associations, (d) treating indicators of parenting and adolescent problem behavior was as- cognitive appraisals and emotional insecurity simi- sessed. Emotional reactivity and behavioral regula- larly in the analytic model, (e) including observer tion mediated the prospective association between ratings of marital hostility to help reduce shared marital discord and adolescent externalizing prob- method variance, and (f) including three waves of lems, controlling for emotional security about data so that marital hostility, youth responses, and parenting. Negative internal representations were adolescent problem behavior are assessed at differ- associated positively with marital discord and youth ent time periods. In addition to these contributions, internalizing and externalizing problems through the present study strengthens this literature by ex- emotional insecurity about parenting. amining youths’ independent and interrelated emo- tional and cognitive responses to marital hostility. This is done by examining the unique and overlap- Conjoint Examinations ping aspects in the self-report measurement of Using the previously described sample from the youths’ cognitive appraisals and emotional insecu- United Kingdom, Davies et al. (2002) examined the rity. A major contribution of this work is that theo- conjoint, prospective associations among interpa- retical models are examined that identify the rental conflict, perceived threat appraisals, self- mediating functions of specific, explanatory cogni- blame appraisals, perceptions of emotional insecu- tive, and emotional responses to parents’ marital rity, and adolescent problem behavior. Interparental hostility. Adolescents’ Responses to Marital Hostility 779

Four research questions are addressed as follows: all study variables (contact corresponding author for statistical details). (1) Do cognitive appraisals uniquely mediate the Youth ranged in age from 11 to 14 years prospective associations between marital (M 5 11.86, SD 5 .69). There were 211 daughters hostility and adolescent problem behavior? (51%). Ninety-one percent of the families were Eu- (2) Do indicators of emotional insecurity ropean American and 3% were African American. uniquely mediate the prospective associations Our African American sample underrepresented the between marital hostility and adolescent percentage of married African American couples problem behavior? with their own children younger than 18 in the (3) Does a conjoint model of adolescents’ emo- county (4.5%) and in the United States (7.8%) (U.S. tional and cognitive responses mediate the Census Bureau, 2000a, Table PCT27 of SF4). The prospective associations between marital median level of parents’ education in this sample hostility and adolescent problem behavior? was 2 years of college, similar to that of European (4) Does an integrated emotional and cognitive American adults in the county who were older than process mechanism mediate the prospective 24 (county mean category was some college; U.S. associations between marital hostility and Census Bureau, 2000b, Table P148A of SF4). The adolescent problem behavior? median level of 2001 household income for families in this study was approximately $70,000, which was higher than the median 1999 income for married- Method couple families in the county ($59,548, U.S. Census Bureau, 2000c, Table PCT40 of SF3). Sampling Procedures and Characteristics The sample used for this study was taken from a Data Collection Procedures larger study of the effects of marital conflict on the transition from childhood into adolescence. For the Youth completed a questionnaire during school. larger study, sixth-grade youth in 13 middle schools They had as much time as needed to finish, and in a large county in the Southeastern United States several trained researchers were available to answer during the 2001 school year were invited to partici- questions. One of the youth’s teachers also com- pate. Children in sixth grade were selected because pleted a questionnaire that focused on the child’s they are beginning the transition from childhood into behavior. Family members were mailed a question- adolescence. Ninety-six percent of the teachers par- naire and asked to complete each independently. The ticipated. Youth were given their invitation in the completed questionnaires were collected during a form of a letter during homeroom to take home to home visit. Parents and youth completed another parents. Two additional direct mailings to parents brief questionnaire during the home visit. This sec- were carried out. Approximately 71% of the families ond questionnaire contained the most sensitive in- returned the consent form and 80% of the youth re- formation (e.g., assessment of marital hostility) and a ceived parental permission to participate. This sam- researcher’s presence ensured privacy. ple was representative of families in the county on Family members participated in two observa- race, parents’ marital status, and family poverty tional tasks during the home visit that were video- status (contact the corresponding author for details taped and coded later by trained observers. The first using census information). was a 20-min problem-solving activity and mother, Eligible families from the larger sample were in- father, and youth participated. Each person com- vited to participate in the present study based on the pleted the Issues Checklist before the interaction task following criteria: (a) parents married or long-term (Conger et al., 1992), and based on this information, cohabitants and (b) no stepchildren in the familyFin the home visitors selected several salient topics. or out of the home. Four hundred and sixteen par- Participants were asked to elaborate the issue, ents agreed to participate in the study (37% response identify who is usually involved, and suggest pos- rate). The primary reasons given for nonparticipa- sible solutions. The second 20-min task included tion included time constraints and an unwillingness only the wife and husband and focused on the for one or more family members to be videotaped. marital relationship, daily interaction patterns, and Using information from the initial youth survey for coparenting. Within each participating family, selection analyses, eligible participating families different coders rated the interaction from the two were similar to eligible nonparticipating families on tasks to minimize coder carryover effects. 780 Buehler, Lange, and Franck

For the longitudinal component of this study, both wives and husbands. For the observational family members completed questionnaires two more measure, coders rated wife’s behavior toward hus- times at yearly intervals. There were 366 participat- band and husband’s behavior toward wife during ing families at Year 2 (Y2) (88% retention) and 340 the two interaction tasks. The following scales were families at Y3 (82% retention of Y1 families). Attri- used from the Iowa Family Interaction Rating scales: tion analyses using MANOVA were conducted using hostility, angry coercion, verbal attack, and antisocial the Y1 data and there were no differences between (Melby et al., 1993). Verbal attack includes critical, the retained and attrited families on any of the study demeaning, global comments that transcend specific variables (contact corresponding author for statisti- situations. Angry coercion includes control attempts cal details). that are hostile in nature. Hostility is a composite scale that includes situationally specific criticisms and sarcasm. Antisocial is a composite negativity Measures rating that also includes uncooperativeness and The analytic models included (a) Y1 marital hos- rudeness. In addition, two rating scales were devel- tility and Y1 adolescent internalizing and external- oped for this study: personal attack and yelling. izing problem behavior, (b) Y2 problematic youth Personal attack includes global criticisms that are emotional and cognitive conflict responses, and (c) directed toward the partner’s character. Yelling in- Y3 adolescent internalizing and externalizing prob- cludes intense, expressed negative affect. Behavior is lem behavior. rated using a 1 (not present)to9(mostly characteristic) Adolescent problem behavior. Youth reports of response format. Cronbach’s a was .85 for the ob- problem behavior were measured using the Child served rating composite. Twenty percent of the tasks Behavior Checklist – Youth Self-Report (CBCL – YSR, were selected randomly to be coded by a second Achenbach, 1991b). This measure consisted of a se- coder and the average agreement across raters was ries of statements that might describe the youth .79. Interrater reliability was assessed by calculating during the previous 6 months. The response format single-item intraclass correlation coefficients was 0 (not true), 1 (somewhat or sometimes true), and 2 (ICCs) based on a one-way random effects ANOVA (very true or often true). Thirty items measured ex- (Melby & Conger, 2001). The average ICC for ternalizing problems (Y1 a 5 .85; Y3 a 5 .89), and 31 this composite measure was .51, which is adequate items measured internalizing problems (Y1 a 5 .88; for these rating scales and comparable to other Y3 a 5 .89). Examples of externalizing items were ‘‘I studies that have used IFIRS ratings (Melby & lie or cheat’’ and ‘‘I disobey at school.’’ Examples of Conger, 2001). internalizing items were ‘‘I feel worthless or inferi- Youths’ marital conflict responses. Youth completed or’’ and ‘‘I am unhappy, sad, or depressed.’’ Raw two measures that assess perceptions of or responses scores were used (Achenbach, 1991b). to parents’ disputes. The Children’s Perceptions of As a second indicator of adolescent externalizing Interparental Conflict scale (CPIC) includes three problems, a teacher also completed the externalizing cognitive appraisal subscales (Grych, Seid, & Finc- subscale of the Teacher Report Form each year in ham, 1992) and uses a 3-point response format. The March or April (Achenbach, 1991a). Cronbach’s a perceived threat subscale has 6 items (e.g., When my was .95 for Y1 and .91 for Y3. Youth also completed parents argue I’m afraid that something bad will the short version of the Children’s Depression happen; Y2 a 5 .85). The coping efficacy subscale has Inventory as a second indicator of adolescent inter- 6 items (e.g., When my parents argue they don’t nalizing problems (Kovacs, 1992). This includes 10 listen to anything I say; Y2 a 5 .67). A higher score items that ask about feelings and behavior during the represented perceived efficacy. The self-blame sub- previous 2 weeks (e.g., frequency of sadness; Y1 scale has 5 items (e.g., It’s usually my fault when my a 5 .72; Y3 a 5 .79). parents argue; Y2 a 5 .86). Marital hostility. At Y1, spouses completed an 18- Youth completed the Security in the Interparental item measure of marital hostility (13 items from the Subsystem scale to assess emotional security associ- verbal and physical aggression subscales of the ated with marital conflict (SIS; Davies, Forman, Rasi, Conflicts and Problem Solving Strategies question- & Stevens, 2002). Items use a 4-point response for- naire [Kerig, 1996], and 5 items from Buehler et al.’s, mat. The emotional reactivity subscale has 12 items 1998 measure of overt conflict). Sample items were ‘‘I (e.g., When my parents argue I feel upset; Y2 a 5 .87). tell my spouse to shut up,’’ ‘‘I slap my spouse,’’ and The behavioral regulation subscale has 12 items (e.g., ‘‘I criticize my spouse.’’ The response format ranged When my parents have an argument I try to solve the from 1 (never)to5(always). Cronbach’s a was .89 for problem for them; Y2 a 5 .84). The internal repre- Adolescents’ Responses to Marital Hostility 781 sentation subscale has 12 items (e.g., When my ciated uniquely with Y3 adolescent externalizing parents have an argument I worry about my family’s (.25, po.01) and internalizing problem behavior (.23, future; Y2 a 5 .80). po.01).

Analytic Strategy Two different sets of longitudinal models were Preliminary Analyses estimated. The first set includes time-ordered data so that the prospective patterning could be examined. Separate cognitive and emotional models. Before These models included Y1 marital hostility, Y2 youth conducting conjoint analyses that included both responses to marital hostility, Y3 adolescent problem cognitive appraisals and emotional insecurity, two behavior, and controls for Y1 adolescent problem independent, time-ordered models were examined. behavior. The second model included Y1 marital The first model focused on youths’ cognitive ap- hostility, a reduced set of youth responses, Y3 ado- praisals. Y1 marital hostility was a latent construct lescent problem behavior, and controls for Y1 ado- with three manifest indicators: wife self-report, lescent problem behavior. This second theoretical husband self-report, and observed ratings. One la- model was examined to determine whether a more tent Y2 youth response construct was specified with parsimonious set of cognitive and emotional vari- three manifest indicators: perceived threat, coping ables could better account for youth responses to efficacy, and self-blame. Y3 adolescent externalizing marital hostility. Data were examined using struc- was a latent construct with two indicators: youth and tural equation modeling (Amos 5) with the signifi- teacher reports. Y3 adolescent internalizing was a cance threshold set at po.05. Error covariances were latent construct with two indicators: youth reports estimated for internalizing and externalizing prob- using the YSR and the CDI. Y1 adolescent external- lems that were assessed using the same form (e.g., izing and internalizing were controlled. Y1 marital youth report of internalizing and externalizing). This hostility was associated positively with Y2 prob- was done because we expected shared method lematic cognitive appraisals (.66, po.01). Y2 cogni- variance when using a given version of the Achen- tive appraisals were associated uniquely with bach assessment (Bollen, 1989; Kenny & Kashy, increases in adolescent externalizing (.26, po.05) and 1992). Model fit was evaluated using the chi-square internalizing problems (.31, po.05). These findings statistic and two fit indexes. A nonsignificant chi- replicated previous research that suggested that square indicated a good model fit. However, because youth cognitive appraisals mediate the prospective of the large sample size, a significant chi-square was association between marital hostility and adolescent expected for most models and additional fit indexes problem behavior (Grych et al., 2003). were examined (Byrne, 2001). Fit indexes such as the The second preliminary model focused on youths’ comparative fit index (CFI) range from 0 to 1.00, with perceptions of emotional insecurity. The same latent a cutoff of .95 or higher indicating a well-fitting constructs were used for marital hostility and ado- model and .90 indicating an adequate fit (Hu & lescent problem behaviors as described in the pre- Bentler, 1999). Values for the root mean square error vious paragraph. One latent Y2 youth response of approximation (RMSEA) below .05 indicate a good construct was specified with three manifest indica- model fit and between .06 and .08 indicate an ade- tors: emotional reactivity, behavior regulation, and quate fit (Browne & Cudeck, 1993). Missing data internal representations. Y1 marital hostility was were treated using a full information maximum associated positively with Y2 youth perceptions of likelihood estimation procedure (FIML). FIML emotional insecurity (.46, po.01). Y2 perceptions of produces estimates that are less biased than do emotional insecurity regarding marital hostility were other procedures such as deleting cases (Acock, associated uniquely with increases in adolescent ex- 2005). ternalizing (.27, po.01) and internalizing problems (.38, po.01). These findings replicated previous re- search that suggested that youth cognitive appraisals Results mediate the prospective association between marital Correlations for all variables at Y1 are shown above hostility and adolescent problems (Harold et al., the diagonal in Table 1. Data below the diagonal are 2004). correlations among the variables across the 3 years. Conjoint model. Two intervening latent constructs All coefficients were in the expected directions. Us- were specified to examine youth responses to marital ing latent constructs, Y1 marital hostility was asso- hostility. Cognitive appraisals had three manifest 8 uhe,Lne n Franck and Lange, Buehler, 782

Table 1 Marital Hostility, Youth Conflict Responses, and Adolescent Problem Behavior: Correlations and Descriptive Statistics

Variables 1 2 3 4 5 6 7 8 9 10 11 12 13

1. Marital hostilityFMRa F .51 .37 .22 À .31 .08 .22 .18 .16 .18 .11 .14 .13 2. Marital hostilityFFRa .51 F .38 .21 À .31 .11 .21 .15 .18 .14 .13 .15 .16 3. Marital hostilityFORa .37 .38 F .29 À .26 .13 .21 .16 .25 .18 À .06 .18 .16 4. Perceived threatFYRb .24 .18 .21 F À .44 .51 .54 .36 .50 .19 .09 .29 .24 5. Coping efficacyFYRb À .16 À .24 À .15 À .29 F À .20 À .39 À .31 À .35 À .23 À .14 À .29 À .24 6. Self-blameFYRb .12 .11 .07 .44 À .25 F .29 .08 .33 .40 .21 .37 .25 7. Emotional reactivityFYRc .24 .29 .24 .58 À .38 .27 F .64 .42 .30 .08 .40 .28 8. RegulationFYRc .23 .27 .19 .46 À .27 .24 .65 F .12 .14 À .06 .25 .06 9. RepresentationsFYRc .22 .29 .27 .54 À .43 .33 .61 .37 F .27 .17 .33 .32 10. A. ExternalizingFYRd .09 .07 .11 .29 À .19 .44 .23 .15 .24 F .35 .59 .39 11. A. ExternalizingFTRd .08 .12 .03 .08 À .17 .17 .11 .04 .18 .19 F .04 .01 12. A. InternalizingFYRe .07 .09 .13 .49 À .24 .44 .32 .26 .33 .54 .01 F .58 13. A. DepressionFYRe .10 .08 .13 .38 À .24 .36 .27 .20 .30 .37 .16 .61 Mean W1: 1.78 W1: 1.72 W1: 2.26 W2: 1.31 W2: 2.58 W2: 1.22 W2: 1.44 W2: 1.97 W2: 1.48 W1: 9.47; W3: 8.69 W1: 3.1; W3: 2.99 W1: 10.96 W3: 8.18 W1: 1.3; W3: 1.4 SD W1: 0.41 W1: 0.39 W1: 0.89 W2: 0.44 W2: 0.44 W2: 0.37 W2: 0.45 W2: 0.58 W2: 0.59 W1: 5.98; W3: 7.36 W1: 5.89;W3: 5.35 W1: 7.5; W3: 7.44 W1: 2.2; W3: 2.6 Range W1: 1 – 3.2 W1: 1 – 3.1 W1: 1 – 5.7 W2: 1 – 3 W2: 1 – 3 W2: 1 – 3 W2: 1 – 3.3 W2: 1 – 3.5 W2: 1 – 4 W1: 0 – 41; W3: 0 – 55 W1: 0 – 46; W3: 0 – 33 W1: 0 – 45; W3: 0 – 51 W1: 0 – 15; W3: 0–16 a 0.89 0.89 0.81 W2: 0.86 W2: 0.67 W2: 0.86 W2: 0.87 W2: 0.84 W2: 0.80 W1: 0.85; W3: 0.89 W1: 0.87; W3: 0.84 W1: 0.88; W3: 0.89 W1: 0.72; W3: 0.79

Note. FR 5 father report; MR 5 mother report; TR 5 teacher report; YR 5 youth report. Correlations for Wave 1 data are above the diagonal and for the prospective analyses below the diagonal. Bolded correlations are significant at po.05. aMarital hostility. bCognitive appraisals. cEmotional security. dAdolescent externalizing. eAdolescent internalizing. Adolescents’ Responses to Marital Hostility 783 indicators: perceived threat, coping efficacy, and self- conducts exploratory factor analysis by asking for a blame. Emotional insecurity also had three manifest possible number of factors and then provides a series indicators: emotional reactivity, internal representa- of factor analyses, one for each of the preselected tions, and behavioral regulation. As with the above number of factors. The RMSEA statistic is provided analyses, marital hostility was measured at Y1 and for each factor analysis, and as described above, an adolescent problem behavior was measured at Y3 RMSEA o.05 indicates a good fit. We requested with Y1 problems controlled. 1 – 10 factors. Once an acceptable factor solution This conjoint model did not provide useful infor- was determined, this factor structure was confirmed mation because the correlation between Y2 cognitive using the Y2 data. appraisals and Y2 emotional insecurity was too high The RMSEA fell below .05 with 7 factors (.047). (.90). This created multicollinearity that led to inac- Thus, we examined the factor structure for 7 factors, curate estimates for the associations between Y2 8 factors, 9 factors, and 10 factors. The 9-factor so- youth responses and Y3 adolescent problems. The lution was selected (RMSEA 5 .032). This solution inaccurate estimates included artificially inflated had item loadings that made sense conceptually, had coefficients between Y2 cognitive appraisals and Y3 more primary factor loadings >.40, and had fewer adolescent problems and inverse associations be- split factor loadings (a difference in primary and tween Y2 emotional insecurity and Y3 adolescent secondary factor loadings of .20 was desired). Ten problems. None of these strong paths were statisti- items were eliminated due to low primary loadings cally significant. The estimation problems for asso- (o.40) or due to strong loadings on the primary and ciations with adolescent problems that resulted from secondary factors (suggesting low discriminant va- the lack of discrimination between cognitive ap- lidity). Forty-six items remained. The 9 factors are praisals and emotional insecurity did not affect the emotional dysregulation (11 items), perceived threat associations for the front half of the model. Y1 mar- to self and family (4 items), constructive represen- ital hostility was uniquely associated with both tations (4) (reverse coded in the model), coping ef- problematic cognitive appraisals (.46, po.01) and ficacy (4 items), self-blame (7 items), avoidance (5 emotional insecurity (.49, po.01). items), behavioral dysregulation (3 items), internal- ization of feelings (4 items), and involvement (4 items). The primary factor loadings were high (78% An Integrated Youth Response System: Measurement above .60), and the differences between primary and The results from the conjoint examination of the secondary loadings were large for most items (con- intervening roles of cognitive appraisals and emo- tact corresponding author for detailed table). Using tional insecurity suggested that youth responses help Y2 data, these 46 items were analyzed with confir- explain the association between marital hostility and matory factor analysis and 96% of the items had adolescent problem behavior because the associa- factor loadings above .60. tions between marital hostility and adolescent problems became nonsignificant. However, the findings also indicated that some of the explanatory An Integrated Youth Response System: Mediating effects of cognitive appraisals and emotional inse- Processes curity are overlapping and redundant. The high correlation between the constructs of cognitive ap- Most of the youth response variables were mod- praisals and emotional insecurity suggested that the erately correlated (Table 2). The strongest associa- CPIC and SIS do not have adequate discriminant tions were among emotional dysregulation and validity. Thus, an integrated conceptualization of perceived threat to self or family members, emo- youth cognitive and emotional responses was con- tional dysregulation and involvement of parental sidered by conducting an exploratory factor analysis disputes, avoidance and internalizing feelings, and of the Y1 items from the CPIC and the SIS. MPLUS perceived threat and self-blame. A preliminary was used. The items were treated as categorical analysis was conducted in which one latent construct variables because the frequency distributions sug- was specified that represented Y2 adolescent cogni- gested an ordinal rather interval level of measure- tive and emotional responses to marital hostility. As ment. A promax rotation was used because this is an hypothesized, Y1 marital hostility was associated oblique rotation that allows for nonorthogonal fac- positively with Y2 problematic youth responses (.55, tors. A nonorthogonal factor structure is a reasonable po.01). Both externalizing (.66, po.01) and inter- assumption, given youth responses are correlated nalizing problems (.43, po.01) were stable over the (Davies et al., 2002; Grych et al., 1992). MPLUS 2-year period. Y2 youth responses were associated 784 Buehler, Lange, and Franck

Table 2 Correlations Among Variables Representing Adolescents’ Year 2 (Y2) Responses to Marital Hostility

Y2 adolescent response variables 2 3 4 5 6 7 8 9

1. Emotional dysregulation .45 .14 À .28 .16 .29 .34 .38 .43 2. Threat to self/family .27 À .27 .45 .04 .15 .11 .13 3. Lower constructive representations À .15 .25 À .26 .16 À .10 À .20 4. Coping efficacy À .13 À .26 À .13 À .16 .00 5. Self-blame À .06 .35 .09 À .05 6. Avoidance .04 .40 .21 7. Behavioral dysregulation .19 .16 8. Internalizing feelings .21

Note. A significant correlation is bolded. Estimates are for the response variables shown in Figure 1. with increases in adolescent externalizing (.28, po.01) Two Y2 youth responses mediated the association and internalizing problems (.47, po.01). between Y1 marital hostility and changes in adoles- The next set of analyses examined the unique ef- cent externalizing problems: self-blame and per- fects of the nine youth response variables. Given ceived threat (Figure 1). Contrary to the hypothesis, there were moderate correlations among the youth perceived threat was associated with decreases in response variables (Table 2), the purpose of this adolescent externalizing problems. Five Y2 youth analysis was to see whether there was a set of spe- responses mediated the association between cific, unique responses that helped explain the pro- Y1 marital hostility and increases in adolescent spective association between marital hostility and internalizing problems: emotional dysregulation, adolescent problem behavior. self-blame, lower constructive representations,

Year 2 Year 1 Year 3

Emotional dysregulation Y1 Internalizing .33 .04.14 YR Threat YR-Dep .95 .40 Lower constructive .59 .35 representations –.18

.19-.13 Internalizing.74 .19 MR Coping efficacy –.46 FR .69 .32.23 Self-blame .52.29 .69 .18 Marital Hostility .11.02 .26 Avoidance

.56 .28 Behavioral dysregulation OR .16 .27 Externalizing

Internalization .36 feelings .56 .95

Involvement Y1 TR Externalizing YR

Figure 1. Autoregressive integrated cognitive and emotional response model. N 5 416; w2 (108) 5 312.08; comparative fit index 5 .91; root mean square error of approximation 5 .067. MR 5 mother report; FR 5 father report; OR 5 observer rating; TR 5 teacher report. Nonsig- nificant coefficients are indicated by dotted lines. Internalizing and externalizing: .49, po.01. Year 1 (Y1) marital hostility and Y1 ado- lescent internalizing: .26, po.01. Y1 marital hostility and Y1 adolescent externalizing: .22, po.01. Direct effects were not significant. Correlations among mediating variables are in Table 2. Adolescents’ Responses to Marital Hostility 785

YR YR-Dep. Y1 Internalizing FR Obs MR .95 .60 Y2 Self- blame .43 .69 .29 .56 .69

.59 .34 Y3 .32 Internalizing Y1 Marital .50 Y2 .36 Y2 L. Hostility Upset Constructive Representations .20 .70 .60 .69 .5

.59 Y3 Emo. Dys. Beh. Dys. Y2 Externalizing Withdraw

Threat .59 .63 .95 .36 –.53 .63 Y1 Externalizing Avoid Efficacy TR YR Intern. Feelings

Figure 2. Autoregressive reduced cognitive and emotional response model. N 5 416; w2 (128) 5 509.78; comparative fit index 5 .82; root mean square error of approximation 5 .085. MR 5 mother report; FR 5 father report; OR 5 observer rating; TR 5 teacher report. Nonsig- nificant coefficients are indicated by dotted lines. Year 3 (Y3) internalizing and externalizing: .50, po.01. Y1 marital hostility and Y3 internalizing: À .11 (ns). Y1 marital hostility and Y3 externalizing: À .01 (ns). Y1 marital hostility and adolescent internalizing: .26, po.01. Y1 marital hostility and adolescent externalizing: .22, po.01. internalization of feelings, and avoidance. Coping of constructive representations regarding family efficacy, behavioral dysregulation, and involvement were included as manifest indicators. Y1 marital in parents’ disputes were not unique explanatory hostility was associated strongly with Y2 youth up- mechanisms. settedness (note no shared informants) and was as- One final analysis was conducted. The purpose of sociated uniquely with the Y2 withdrawal response. this analysis was to generate an autoregressive, the- As theorized, Y2 upsettedness was associated oretical model of youth cognitive and emotional re- uniquely with self-blame, lower constructive repre- sponses that pared down the nine response variables sentations, and withdrawal, and was not associated to a more parsimonious model. Based on prior the- directly with adolescent problems. Y2 self-blame and orizing (Davies & Cummings, 1994; Grych & Finc- lower levels of constructive representations were ham, 1990), it was hypothesized that youth who associated uniquely with increases in adolescent in- witness marital hostility might become upset be- ternalizing and externalizing problems. Y2 with- cause they perceive the hostility as self-relevant and drawal was associated uniquely only with increases frightening (Figure 2). Informed by the results from in adolescent internalizing problems. Although these the analysis with the nine manifest responses, it was patterns were consistent with previous research and also hypothesized that they might respond to this proposed theory, the model fit was marginal, distress by blaming self, having lower levels of w2 5 509.78 (128), p4.05, CFI 5 .82, RMSEA 5 .085, constructive representations regarding family rela- and fit the data less well than the model that in- tionships, and withdrawing. Based on the previous cluded the nine separate adolescent responses, analyses and theorizing, upsettedness was included Dw2 5 197.70 (20), po.01. as a latent construct with three manifest indicators: emotional dysregulation, behavioral dysregulation, Discussion and perceived threat. Withdrawal was also included as a latent construct with three manifest indicators: Youths’ cognitive and emotional responses to their avoidance, internalization of feelings, and lower parents’ hostile interactions were examined. The levels of coping efficacy. Self-blame and lower levels study was guided theoretically by the cognitive- 786 Buehler, Lange, and Franck contextual model of interparental conflict (Grych & The findings from this study also provide infor- Fincham, 1990) and the emotional security hypoth- mation about specific process mechanisms that ex- esis (Davies et al., 2002). The findings suggested that plain the associations between marital hostility and youth have a variety of specific, unique emotional changes in adolescent problem behavior. Marital and cognitive responses to parents’ marital hostility hostility was associated with increases in adolescent that place them at risk for current internalizing and internalizing problems through five unique response externalizing adjustment difficulties. mechanisms: self-blame, lower levels of constructive The examination of conjoint and integrative in- representations regarding family relationships, in- fluences of interparental conflict-related cognitive ternalizing distress, avoidance, and emotional dys- appraisals and emotional insecurity suggested a regulation. Part of the attribution of self-blame substantial overlap in these cognitive and emotional involves perceived responsibility for parents’ dis- responses. This overlap replicates the strong corre- putes that might result in feelings of guilt and lation found by Davies et al. (2002) and supports the shame. Over time these feelings might contribute to idea of a functionalist approach to emotion (Campos negative self-evaluations. These negative self-evalu- et al., 1994) that undergirds the emotional security ations might threaten self-esteem and foster feelings hypothesis, the appraisal orientation developed by of anxiety and depression (Garber, Robinson, & Grych and Fincham (1990), and perspectives that Valentiner, 1997). The unique pathway through focus on youths’ emotion and stressor-based coping lower levels of constructive representations details behavior (Kerig, 2001; Lazarus & Folkman, 1984). the important role that less positive family schemas From this integrated perspective, youths’ responses might have in increasing feelings of hopelessness to marital hostility include (a) emotional dysregula- and mistrust regarding commitment and intimacy in tion, (b) appraisals of threat to self/family, close relationships. The potential effects of these blame, and efficacy, and (c) coping responses of negative appraisals over time are important because acting out, avoidance, involvement in parental dis- they might shape the relational templates that ado- putes, and internalizing feelings. The findings from lescents’ use when forming and maintaining rela- this study suggested that this set of responses ade- tionships with peers. Unique response process quately characterized early adolescents’ responses to mechanisms also included internalizing distress and marital hostility, and that cognitive and emotional avoidance. Both are inherent in the development responses completely mediated the association be- of generalized internalizing problem behavior. Al- tween marital hostility and future adolescent prob- though related, each had a unique association with lem behavior. increases in generalized internalized problem be- A major finding from this study is that there are a havior. This pattern demonstrates how stressor-spe- fairly large number of theoretically meaningful and cific responses are associated with more generalized unique responses that youth have when exposed to problem syndromes over time. However, it also marital hostility over time. Y1 marital hostility was might be that some of the youth response variables, uniquely associated with eight of the nine response such as internalizing distress, might be confounded variables (each except for involvement). These un- with the assessment of more generalized problem ique associations occurred in spite of moderate-sized behavior. Future research is needed to distinguish associations among some of the responses. The stress-specific responses from generalized problem strongest unique association between marital hostil- behavior syndromes over time. Controlling for the ity and youth responses was an inverse association other eight responses, emotional dysregulation was with coping efficacy. Over time, exposure to marital uniquely associated with increases in internalizing hostility is associated with lower levels of compe- problems. This unique effect might be due, in part, to tency appraisals related to minimizing parents’ ac- the debilitating effects of rumination. Adolescents rimony and less confidence in being able to regulate who are absorbed by experience of witnessing mar- personal upsettedness. The findings also suggest ital hostility are less able to address other develop- that youth respond to marital hostility by avoiding mental tasks (e.g., doing well in school) and attend to the conflict interactions rather than by involving other important relationships with peers and non- themselves in the disputes. This is a positive finding, familial adults. given involvement might lead to triangulation be- Two response mechanisms mediated the pro- tween parents and triangulation or feeling caught spective association between marital hostility and between parents is associated with emotional ad- adolescent externalizing problem behavior. As with justment difficulties (Amato & Afifi, 2006; Grych, internalizing problems, self-blame was a significant Raynor, & Fosco, 2004). linking mechanism. This finding replicates previous Adolescents’ Responses to Marital Hostility 787 research (Grych et al., 2003) and might reflect per- lidity information regarding the nine response vari- ceived or actual responsibility for parents’ disputes ables identified in this study. Future research also (i.e., parents fighting about the youth’s conduct needs to extend beyond a focus on adolescent problems). However, because Y1 adolescent problem problem behavior and include other aspects of de- behavior was controlled, the concurrent association velopment, such as relationships with peers, dating between marital hostility and adolescent conduct partners, and general social competence. problems was accounted for in this study, reducing Within the context of these limitations, the present the likelihood that the significance of self-blame is a findings regarding cognitive appraisals and emo- by-product of ‘‘child effects.’’ Grych et al. speculated tional insecurity document the important role of that self-blame might be associated with adolescent adolescents’ subjective constructions in understand- externalizing problems because youth might involve ing how witnessing marital hostility influences ad- themselves in parents’ disputes using distracting, olescent problem behavior. The study demonstrated disruptive behavior. In this study, self-blame was not that adolescents’ cognitive and emotional responses associated with involvement in disputes during are both needed in a process model that details how seventh grade. In addition, self-blame was associat- marital hostility places youth at risk for internalizing ed strongly with behavioral dysregulation during and externalizing problems. Future theorizing and seventh grade, suggesting a possible linkage be- research needs to consider the temporal stability and tween adolescents’ situationally based aggressive impact of these subjective constructions for other responses to marital hostility and broader-based important developmental outcomes and social rela- current and future externalizing problems (Cum- tionships with peers and intimates. mings, Goeke-Morey, & Papp, 2004). Contrary to the hypothesis, Y2 perceived threat References was associated with decreases in adolescent exter- nalizing problems when controlling for the other Achenbach, T. M. (1991a). Manual for the teachers report form eight youth responses. (The uncontrolled association and 1991 profile. Burlington: Department of Psychiatry, University of Vermont. was positive.) If this unexpected pattern replicates in Achenbach, T. M. (1991b). Manual for the youth self-report future research, one plausible interpretation might and 1991 profile. Burlington: Department of Psychiatry, be what Emery (1988) referred to as ‘‘angel-like’’ University of Vermont. behavioral responses to interparental strife. He Acock, A. C. (1997). Working with missing values. Journal speculated that some children might respond to of Marriage and Family, 67, 1012 – 1028. worry and concern about marital hostility by acting Amato, P. R., & Afifi, T. D. (2006). Feeling caught between good or increasing their prosocial behavior (and parents: Adult children’s relations with parents and minimizing aggressive and delinquent behavior). subjective well-being. Journal of Marriage and Family, 68, Although this study makes several theoretical and 222 – 235. methodological contributions, there also are several Bell, M. A., & Wolfe, C. D. (2004). Emotion and cognition: limitations. The number of minority families was too An intricately bound developmental process. Child De- velopment, 75, 366 – 370. small to allow for moderating tests across racial or Blair, C. (2002). School readiness: Integrating cognition and ethnic groups. Thus, generalizations of these find- emotion in a neurobiological conceptualization of chil- ings to minority families should be made cautiously. dren’s functioning at school entry. American Psychologist, Future research needs to be conducted with larger 57, 111 – 127. samples of ethnic minority families so that both Bollen, K. A. (1989). Structural equations with latent variables. within- and between-group analyses can be con- Oxford, UK: Wiley. ducted. Moreover, youths’ responses to marital Browne, M. W., & Cudeck, R. (1993). Alternative ways of hostility need to be assessed using multiple methods assessing model fit. In K. Bollen & J. Long (Eds.), Testing to help reduce method variance that might be in- structural equation models (pp. 136 – 162). Newbury Park, flating the association among some of the response CA: Sage. variables. Minimizing shared method variance Buehler, C., Anthony, C., Krishnakumar, A., Stone, G., Gerard, J., & Pemberton, S. (1997). Interparental conflict might reduce the intercorrelations and provide a and youth problem behaviors: A meta-analysis. Journal more valid estimate of the unique effects of some of of Child and Family Studies, 6, 233 – 247. the response variables. It is also important to note Buehler, C., Benson, M., & Gerard, J. M. (2006). that the response variables identified were based on Interparental hostility and early adolescent problem youths’ responses to two questionnairesFthe CPIC behavior: The mediating role of specific aspects of and the SIS. As such, the use of additional methods parenting. Journal of Research on Adolescence, 16, 265 – to assess youth responses will provide needed va- 292. 788 Buehler, Lange, and Franck

Buehler, C., Krishnakumar, A., Stone, G., Anthony, C., Self-worth as a mediator. Journal of Adolescent Research, Pemberton, S., Gerard, J., et al. (1998). Interparental 12, 12 – 33. conflict styles and youth problem behaviors: A two- Gerard, J. M., Buehler, C., Franck, K. L., & Anderson, O. sample replication study. Journal of Marriage and the (2005). In the eyes of the beholder: The functional roles Family, 60, 119 – 132. of perceived appraisals associated with interparental Byrne, B. M. (2001). Structural equation modeling with conflict. Journal of Family Psychology, 19, 376 – 384. AMOS: Basic concepts, applications, and programming. Grych, J. H., & Fincham, F. D. (1990). Marital conflict and Mahwah, NJ: Erlbaum. children’s adjustment: A cognitive-contextual frame- Call, K. T., & Mortimer, J. T. (2001). Arenas of comfort in work. Psychological Bulletin, 108, 267 – 290. adolescence: A Study of adjustment in context. Mahwah, NJ: Grych, J. H., & Fincham, F. D. (1993). Children’s appraisals Lawrence Erlbaum. of interparental conflict: Initial investigations of the Campos, J. J., Campos, R. G., & Barrett, K. C. (1989). cognitive-contextual framework. Child Development, 64, Emergent themes in the study of emotional develop- 215 – 230. ment and emotion regulation. Developmental Psychology, Grych, J. H., Fincham, F. D., Jouriles, E. N., & McDonald, R. 25, 394 – 402. N. (2000). Interparental conflict and child adjustment: Campos, J. J., Mumme, D. L., Kermoian, R., & Campos, R. Testing the mediational role of appraisals in the cogni- G. (1994). A functionalist perspective on the nature of tive-contextual framework. Child Development, 71, 1176 – emotion. Monographs of the Society for Research in Child 1193. Development, 59, 284 – 303. Grych, J. H., Harold, G. T., & Miles, C. J. (2003). A pro- Conger, R. D., Conger, K. J., Elder, G. H., Lorenz, F. O., spective investigation of appraisals as mediators of the Simons, R. L., & Whitbeck, L. B. (1992). A family process link between interparental conflict and child adjust- model of economic hardship and adjustment of early ment. Child Development, 74, 1176 – 1193. adolescent boys. Child Development, 63, 526 – 541. Grych, J. H., Raynor, S. R., & Fosco, G. M. (2004). Family Cummings, E. M., Goeke-Morey, M. C., & Papp, L. M. processes that shape the impact of interparental conflict (2004). Everyday marital conflict and child aggression. on adolescents. Development and Psychopathology, 16, Journal of Abnormal Child Psychology, 32, 191 – 202. 649 – 665. Dadds, M. R., Atkinson, E., Turner, C., Blums, G. J., & Grych, J. H., Seid, M., & Fincham, F. D. (1992). Assessing Lendich, B. (1999). Family conflict and child adjustment: marital conflict from the child’s perspective: The Chil- Evidence for a cognitive-contextual model of intergen- dren’s Perception of Interparental Conflict Scale. Child erational transmission. Journal of Family Psychology, 13, Development, 63, 558 – 572. 194 – 208. Harold, G. T., Shelton, K. H., Goeke-Morey, M. C., & Davies, P. T., & Cummings, E. M. (1994). Marital conflict Cummings, E. M. (2004). Marital conflict, child emo- and child adjustment: An emotional security hypothesis. tional security about family relationships and child ad- Psychological Bulletin, 116, 387 – 411. justment. Social Development, 13, 350 – 376. Davies, P. T., & Cummings, E. M. (1998). Exploring chil- Hu, L., & Bentler, P. M. (1999). Cutoff criteria for fit indexes dren’s emotional security as a mediator of the link be- in covariance structure analysis: Conventional criteria tween marital relations and child adjustment. Child versus new alternatives. Structural Equation Modeling, 6, Development, 69, 124 – 139. 1 – 55. Davies, P. T., Forman, E. M., Rasi, J. A., & Stevens, K. I. Katz, L. F. (2001). Physiological processes as mediators (2002). Assessing children’s emotional security in the of the impact of marital conflict on children. In J. H. interparental relationship: The Security in the Interpa- Grych & F. D. Fincham (Eds.), Interparental conflict rental Subsystem Scales. Child Development, 73, 544 – 562. and child development: Theory, research and applications Davies, P. T., Harold, G. T., Goeke-Morey, M. C., & Cum- (pp. 188 – 212). Cambridge, UK: Cambridge University mings, E. M. (2002). Child emotional security and in- Press. terparental conflict. Monographs of the Society for Research Kenny, D. A., & Kashy, D. A. (1992). Analysis of the mul- in Child Development 67(3, Serial No. 270). titrait-multimethod matrix by confirmatory factor anal- Derryberry, D., & Tucker, D. (1994). Motivating the focus of ysis. Psychological Bulletin, 112, 165 – 172. attention. In P. Niedenthal & P. Kitayama (Eds.), The Kerig, P. K. (1996). Assessing the links between interpa- heart’s eye: Emotional influences in perception and attention rental conflict and child adjustment: The conflicts and (pp. 167 – 196). New York: Academic Press. problem-solving scales. Journal of Family Psychology, 10, Doyle, A. B., & Markiewicz, D. (2005). Parenting, marital 454 – 473. conflict and adjustment from early-to mid-adolescence: Kerig, P. K. (2001). Children’s coping with interparental Mediated by adolescent attachment style? Journal of conflict. In F. D. Fincham & J. H. Grych (Eds.), Interpa- Youth and Adolescence, 34, 97 – 110. rental conflict and child development: Theory, research, and Emery, R. E. (1988). Marriage, divorce, and children’s adjust- applications (pp. 213 – 245). New York: Cambridge Uni- ment. Newbury Park, CA: Sage. versity Press. Garber, J., Robinson, N. S., & Valentiner, D. (1997). The Kovacs, M. (1992). Children’s depression inventory manual. relation between parenting and adolescent depression: North Tonawanda, NY: Multi-Health Systems. Adolescents’ Responses to Marital Hostility 789

Lazarus, R. S., & Folkman, S. (1984). Stress, appraisal, and duct disorder/delinquency. In B. B. Lahey, T. E. Moffitt coping. New York: Springer. & A. Caspi (Eds.), Causes of conduct disorder and juvenile Linder, J. R., & Collins, W. A. (2005). Parent and peer delinquency (pp. 227 – 253). NY: Guilford. predictors of physical aggression and conflict manage- Parke, R. D., Kim, M., Flyr, M., McDowell, D. J., Simpkins, ment in romantic relationships in early adulthood. S. D., Killian, C. M., et al. (2001). Managing marital Journal of Family Psychology, 19, 252 – 262. conflict: Links with children’s peer relationships. In J. H. Mathews, G., & Wells, W. (1999). The cognitive science of Grych & F. D. Fincham (Eds.), Interparental conflict and attention and emotion. In T. Dalgleish & M. Power child development: Theory, research and applications (pp. (Eds.), Handbook of cognition and emotion (pp. 171 – 192). 291 – 314). Cambridge, UK: Cambridge University Press. West Sussex, UK: Wiley. Steinberg, L. (2001). We know some things: Parent-ado- Melby, J. N., & Conger, R. D. (2001). The Iowa Family In- lescent relationships in retrospect and prospect. Journal teraction Rating Scales: Instrument summary. In P. K. of Research on Adolescence, 11, 1 – 19. Kerig & K. M. Lindahl (Eds.), Family observational coding U.S. Census Bureau. (2000a). PCT27. Family type by presence systems (pp. 33 – 57). Mahwah, NJ: Erlbaum. of own children under 18 years of age of own children. Fam- Melby, J. N., Conger, R. D., Book, R., Rueter, M., Lucy, L., ilies (total), White alone, Black alone. Retrieved September Repinski, D., et al. (1993). The Iowa Family Interaction 29, 2004, from http://factfinder.census.gov, Summary Rating Scales, 4th ed. Unpublished manuscript, Center for File 4. Family Research in Rural Mental Health, Iowa State U.S. Census Bureau. (2000b). PCT148A. Sex by educational University, Ames. attainment for the population 25 years and over (White Mogg, K., & Bradley, B. (1999). Selective attention and alone). Retrieved September 29, 2004, from http://fact- anxiety: A cognitive-motivational perspective. In T. finder. census.gov, Summary File 3. Dalgleish & M. Power (Eds.), Handbook of cognition and U.S. Census Bureau. (2000c). PCT40. Median family income emotion (pp. 145 – 170). West Sussex, UK: Wiley. in 1999 (dollars) by family type by presence of own children Nigg, J. T., & Huang-Pollock, C. L. An early-onset model of under 18 years. Retrieved September 3, 2005, from the role of executive functions and intelligence in con- http://factfinder.census.gov, Summary File 3.