conflict in families and the psychological adjustment of …repetti.bol.ucla.edu/jaycox repetti...

Journal of Family Psychology1993, Vol. 7, No. 3, 344-355

Copyright 1993 by the American Psychological Association, Inc.0893-3200/93/S3.OO

Conflict in Families and the Psychological Adjustment ofPreadolescent Children

Lisa H. Jaycox and Rena L. Repetti

We examined the cross-sectional association between conflict in families and childpsychological adjustment in 72 4th-5th graders. Multiple informants (parents, children,and teachers) assessed conflict and anger in the social climate of the home, maritaldiscord, negative emotional tone in the parent-child relationship, and child adjustment.As predicted, child adjustment was more strongly related to family conflict than tomarital discord. There was a stronger association between family conflict and malad-justment in girls. Moreover, the association between a general climate of conflict at homeand child maladjustment was independent of anger and discord in the marital or parent-child relationships. During the study of the effects of interpersonal conflict at home, itappears to be important to identify the locus of anger and aggression. Findings suggestthat researchers should distinguish between a general climate of conflict in the familyand interparental discord.

We examined anger and conflict in familiesand related it to preadolescent children's psy-chological adjustment. The term conflict is usedhere to describe behaviors between family mem-bers that range from verbal disagreements andcriticism to acts of physical aggression. Differenttheoretical perspectives, such as attachmenttheory, social learning theory, and cognitivetheory, agree that children are placed at in-creased risk for the development of behavioraland emotional problems by child-rearing envi-ronments in which interpersonal relations arecharacterized by anger and conflict (cf. Grych &Fincham, 1990; Margolin, 1981; Marvin &Stewart, 1990). It is therefore not surprising thatadolescents' descriptions of high levels of familyconflict have been associated with self-reporteddistress and psychological symptoms in recent

We would like to thank the principals, teachers, par-ents, and children who participated in the study. Weare also grateful to Andy Christensen, Steve Lopez,Gayla Margolin, and three anonymous reviewers fortheir helpful comments on an earlier draft of this ar-ticle. Rena L. Repetti's participation in the prepara-tion of this article was supported by a FIRST awardR29-48595 from the National Institute of MentalHealth.

Correspondence concerning this article should beaddressed to either Lisa H. Jaycox, Department ofPsychology, University of Pennsylvania, 3815 WalnutStreet, Philadelphia, Pennsylvania 19104, or Rena L.Repetti, Department of Psychology, University ofCalifornia, 405 Hilgard Avenue, Los Angeles, Califor-nia 90024-1563.

cross-sectional studies (Kleinman, Handal,Enos, Searight, & Ross, 1989; Kurdek & Sin-clair, 1988; Slater & Haber, 1984) and in ananalysis of 1 -year longitudinal data (Mechanic &Hansell, 1989). In fact, empirical evidence sug-gests that ongoing conflict at home has a greaterimpact on adolescent distress and symptomsthan does parental divorce (Borrine, Handal,Brown, & Searight, 1991; Kurdek & Sinclair,1988; Mechanic & Hansell, 1989; Slater &Haber, 1984). Family conflict was also linked tolower self-esteem among adolescents in onecross-sectional study (Slater & Haber, 1984), butnot in the longitudinal study (Mechanic &Hansell, 1989).

Although most of the literature on family con-flict has focused on adolescents, several studieshave used samples with wide age ranges, frompreschoolers to adolescents. In a study of 6-18-year-olds, parents' reports of conflict at homewere associated with children's self-reported ag-gression (Weidner, Hutt, Conner, & Mendell,1992). In another study of chronically ill 4-14-year-old children, parents' reports of family con-flict correlated with their ratings of less sociabil-ity and more problems of undercontrol amongthe children (Kronenberger & Thompson,1990). However, there was not a significant as-sociation between parents' ratings of conflictand children's internalizing symptoms. A studyof child psychiatric inpatients (aged 6-13 years)indicated that the children who attempted or

344

CONFLICT IN FAMILIES AND CHILD ADJUSTMENT 345

thought about suicide described their families asbeing higher in conflict than did the nonsuicidalchildren (Asarnow, 1992). Unfortunately, noneof these studies examined the effects of familyconflict separately for the different age groups.Thus, although there is empirical evidence of aneffect of family conflict on adolescent adjust-ment, the effect on younger children has notbeen established.

Differentiating Family and MaritalConflict

Related to the literature on family conflict is alarge body of research focusing specifically ondiscord between parents (cf. Grych & Fincham,1990). Although the strength of the associationvaries from study to study, a recent meta-analysis suggests that the association betweeninterparental conflict and behavioral problemsis fairly weak but detectable for boys and thatthere is no reliable association for girls (Reid &Crisafulli, 1990). Among the significant find-ings, however, it is impossible to determine theextent to which results that have been attributedspecifically to conflict between parents are actu-ally caused by a pervasive climate of anger anddiscord in the home.

Marital theorists have long suspected that, inaddition to the adverse effects of direct observa-tions of discord between parents, children mayalso suffer because of the impact that maritalproblems have on the parent-child relationship(Christensen & Margolin, 1988; Margolin,1981). Parents in a discordant marriage may be-come withdrawn or hostile toward their childrenor they may attempt to gain the alliance of thechild against the other parent. Christensen andMargolin (1988) have documented the spread ofmarital conflict to the parent-child relationshipamong families in which the parents are in anunhappy marriage. Moreover, other empiricalevidence confirms that parenting behaviors,such as a lack of warmth, rejection, and emo-tional withdrawal, mediate at least part of theassociation between marital discord and childmaladjustment (Belsky, 1985; Burman, John, &Margolin, 1987; Conger et al., 1992; Fauber,Forehand, Thomas, & Wierson, 1990; Peterson& Zill, 1986; Tschann, Johnston, Kline, &Wallerstein, 1989). If, as these studies indicate,interparental conflict often does not remain con-fined within the marital relationship, then re-

ports of marital conflict may, in many cases,also reflect discord throughout the family.

To avoid a misspecification of effects, we in-cluded separate measures of family conflict andmarital conflict. The distinction between the en-tire family system and a subsystem within thefamily is similar to Repetti's (1987) descriptionof common and individual social environments.In our study, the common social environmentwas the social climate that was shared by allfamily members. We used the family's self-reported behavioral norms for expressing angerand coping with conflict as an indicator of thedegree of conflict in the common social environ-ment of the home. Parents and children de-scribed the extent to which interactions amongfamily members tended to be openly angry, hos-tile, and aggressive. The individual social envi-ronment of primary interest was the separatesocial space surrounding the parents (i.e., thesocial climate of their marriage). Marital discordwas indicated by the parents' reported use ofaggression to resolve marital disputes and bydissatisfaction with their marriage.

Because most researchers of marital conflicthave not distinguished between the family andthe marriage as units of analysis, they have notbeen able to address whether discord betweenparents is particularly stressful for a child ormore stressful than living in a home in whichinterpersonal conflict and open expressions ofanger characterize all family relationships. In anexception to that trend in the literature, Kurdekand Sinclair (1988) asked adolescents to de-scribe their perceptions of both interparentalconflict and family conflict; they found that ameasure of their psychological adjustment wasabout equally correlated with the two ratings(rs = .37 and .32, ps < .01, respectively).

There are at least two reasons to expect thatthe impact of a general family atmosphere ofanger and discord would be greater than the spe-cific impact of interparental conflict, particu-larly when parents' perceptions of discord,rather than children's perceptions, are used. Onebasis for this prediction is an expected differencein the frequency with which the child would beexposed to anger and conflict in the home. Itwould be impossible for children to escape ageneral social climate of anger and conflict intheir home. By contrast, parents in a troubledmarriage often strive to maintain the image of aharmonious relationship and may succeed in

346 LISA H. JAYCOX AND RENA L. REPETTI

keeping many of their dissatisfactions and angrydisputes hidden from their children. Althoughthey may accurately describe their marriage asunhappy and discordant when completing aconfidential questionnaire, that does not neces-sarily imply that their children observe frequentopen conflict between them. This point was sup-ported by a recent study in which it was reportedthat there were modest correlations between thedescriptions of marital conflict provided by pre-adolescent children and their parents (rs = .30and .39, ps < .01, respectively; Grych, Seid, &Fincham, 1992). More important, measures ofthe children's psychological adjustment weremuch more strongly related to their own reportsof the frequency and intensity of interparentalconflict than to their parents' reports.

A second basis for our prediction is an ex-pected difference in the extent to which the childwould become involved in angry and confiictiveinteractions. Among families in which open ex-pressions of anger and aggression are the norm,children not only observe fighting among differ-ent family members, but they would presumablyalso become involved in angry conflicts withtheir parents and siblings. Some evidence sug-gests that this may be especially true for daugh-ters (Vuchinich, Emery, & Cassidy, 1988).

Greater exposure to and involvement in hos-tile and aggressive interactions at home providechildren with fewer opportunities to learn adap-tive ways of handling anger and resolving dis-putes. For instance, researchers have found thatchildren exposed to harsh discipline practices athome tend to develop maladaptive social infor-mation processing patterns, which mediate theimpact of harsh discipline on subsequent aggres-sive behavior at school (Weiss, Dodge, Bates, &Pettit, 1992). In this type of home environment,children may develop more negative self-perceptions and come to view the world as ahostile place. Thus, these children may be at anincreased risk for low self-esteem and both in-ternalizing and externalizing behavior prob-lems. Interestingly, marital theorists have iden-tified frequency and child involvement as twodimensions of marital conflict that may be criti-cal to an understanding of the impact of maritalconflict on children's adjustment (Fincham &Osborne, 1993). Because of more frequent expo-sure to and direct involvement in aggressive in-teractions, we expected that a child's self-percep-tion and behavioral adjustment would be more

strongly associated with ratings of a general cli-mate of conflict in the family than with ratingsof conflict within the marital relationship per se.

The Study

In this correlational study we focused on an-ger and conflict in the homes of preadolescentchildren and related it to the children's psycho-logical adjustment. We studied third- throughfifth-grade children in an attempt to replicateand extend reliable findings from the literatureon adolescence to a younger age group: 8-11-year-olds. Two hypotheses were tested. The firsthypothesis predicted that children's and parents'reports of higher levels of family conflict wouldbe associated with three indicators of poor psy-chological adjustment among the children: morebehavioral problems at home and at school andlower perceived self-competence. Second, childadjustment was expected to correlate morestrongly with ratings of family conflict than withparents' reports of marital conflict.

Most of the previously reported associationsbetween family conflict and child or adolescentadjustment were based on ratings obtained froma single source (either child or parent). We usedmultiple informants (mother, father, child, andteacher) in this study to limit problems associ-ated with confounding attributable to respon-dent biases. Therefore, we highlight resultsthat were based on "cross-source" correlations,which were based on measures completed bydifferent informants. The advantage of usingmultiple informants in this type of study is bal-anced by modest levels of agreement betweenindependent raters of child adjustment (Achen-bach, McConaughy, & Howell, 1987; Hinshaw,Han, Erhardt, & Huber, 1992), in part reflectingthe high degree of situational specificity of childbehavior problems. In short, we addressed re-spondent bias problems through the use of mul-tiple informants while acknowledging limits onthe expected level of agreement among differentsources.

Our analyses also addressed the possible con-founding role of angry parent-child interactionsin the association between family conflict andchild maladjustment. The family conflict scalethat we used asked family members to describethe general norms for expressions of anger andaggressive behavior in their home. We won-dered whether the subjects in our study primar-


ily had the parent-child relationship in mindwhen they were responding to the family conflictquestions; we also wanted to rule out the possi-bility that any family conflict results actuallyrepresented no more than the well-documentedassociation between parent-child conflict andchild behavior problems. In particular, wewanted to control for the obvious impact thatchild aggressiveness has on conflict with parentsand other family members. Therefore, we exam-ined the association between family conflict andchild adjustment problems after controlling forparents' ratings of a negative emotional tone inthe parent-child relationship.

Method

Sample and Procedure

Families were recruited through letters sent home toparents of third-, fourth-, and fifth-grade children atfive parochial schools in two metropolitan areas inPennsylvania and New Jersey. Children were inter-viewed at school, and questionnaires were completedby parents at home. In most cases, the child and parentmeasures were obtained within 2 weeks of each other.Teachers completed classroom behavior rating scaleswithin 6 weeks of the initial child interview.

Of the approximately 250 families contacted, 81families (32%) agreed to participate. Similarly low re-sponse rates have been reported in other studies inwhich families were recruited through schools (cf.Mechanic & Hansell, 1989). Nine families with onlyone parent living at home were dropped from theanalyses. The final sample of 72 children (32 boys and40 girls) ranged in age from 8 to 11 years (M = 9.2, SD= 1). Parent questionnaires were returned for 67 chil-dren (93%). In 59 cases, questionnaires were returnedfrom both parents; in 8 cases only mothers returnedthe questionnaire. Teacher ratings (n = 16) were ob-tained for all 72 children.

The participants were predominantly White (94%),Catholic (91%) families with an average yearly familyincome of about $55,000 in 1989-1990. Family sizeranged from one to five children; the majority of thefamilies (81%) had two or three children living athome. In all cases, two parents lived at home with thechild, including four cases of blended families inwhich a biological parent and a stepparent lived withthe child.

Measures

Family conflict. The nine-item Conflict subscaleof the Family Environment Scale (FES; Moos &Moos, 1981) was used to assess perceptions of familyconflict. This is a well-accepted measure with demon-strated validity and reliability (Touliatos, Perlmutter,& Straus, 1990). It was used in all but one of thestudies of family conflict reviewed earlier. A high

score indicates more openly expressed anger, aggres-sion, and conflict among family members (e.g.,"Family members sometimes get so upset they throwthings," "We fight a lot in our family"). Parents weregiven a 4-point response scale (mostly true, some-what true, somewhat false, and mostly false) that wasbased on feedback from an earlier study that showedthat the simple true-false response option was unsat-isfying for adults to use (Cronbach's alphas = .83 and.78 for mothers and fathers, respectively). When weconverted responses from the 4-point scale to theusual 2-point true-false scale for comparison pur-poses, the scores for the families in our study (moth-ers, M = 3.12, SD = 2.3; fathers, M = 3.04, SD = 2.2)were well within the norms from a national sample(M = 3.31, SD = 1.9; Moos & Moos, 1981). Therewas a significant correlation between mothers' andfathers' scores on the Conflict subscale of the FES(r = .39, p < .01), which was similar to parent agree-ment results from other studies (Weidner et ah,1992). In addition, the patterns of correlations be-tween mothers' and fathers' separate Conflict scoresand the other variables in this study were highly simi-lar. Therefore, for simplicity of presentation, we aver-aged mothers' and fathers' perceptions of family con-flict to obtain a mean parental score for familyconflict.

To assess preadolescent children's perceptions offamily conflict, we modified the FES Conflict itemsto match the format used in Harter's (1985) Self-Perception Profile for Children. Children indicatedwhich of two families was more like their family andwhether that family was "really" like their family oronly "sort of" like their family (e.g., "In some fami-lies, family members fight a lot, but in other families,family members don't fight much at all. Which ofthese families is most like your family? Is this reallytrue for your family or only sort of true for your fam-ily?"). There was an acceptable level of internal con-sistency for this modified scale (Cronbach's alpha =.80). As shown in Table 1, children's perceptions offamily conflict agreed with their parents' perceptionsto roughly the same degree that the mothers' and fa-thers' perceptions agreed with each other (r = .38,

Marital discord. Marital aggression and maritalsatisfaction were used as two indicators of maritaldiscord. First, parents rated their use of verbal aggres-sion during marital disputes on the Verbal Aggres-sion subscale of the Conflict Tactics Scale (Straus,1979). The Verbal Aggression subscale includes fouritems (e.g., "Insulted or swore at the other one") thatrespondents rate on a 7-point scale (0 = never, 6 =more than 20 times) to indicate how often they usedeach tactic in the past year (Cronbach's alphas = .70and .80 for mothers and fathers, respectively). ThePhysical Aggression subscale was not used because ofanticipated objections from the schools. Partners'separate reports of their use of verbal aggression dur-ing marital disputes were significantly correlated (r =.38, p < .01). As with the measure of family conflict,we averaged mothers' and fathers' scores to obtain amean marital aggression score. Compared with na-tional survey data, the overall Verbal Aggression


Table 1Pearson Correlations Between Conflict and Psychological Adjustment for Both Sexes

Variable

Family conflict measures1. Child's report of family

conflict2. Parents' average report of

family conflictMarital discord measures

3. Parents' average report ofmarital satisfaction

4. Parents' average report ofverbal aggression withspouse

Child adjustment measures5. Child's perceived

self-competence6. Parents' average report of

internalizing behavior7. Parents' average report of

externalizing behavior8. Teacher's report of behavior

problems

n

72

67

67

67

72

72

72

72

1

—

.38**

-.16

.07

-.58**

.22

.35*

.03

2

—

-.51**

.45**

-.30*

.32*

.55**

.34*

3

-.51**

.14

-.07

-.29*

-.22

4

_

.04

.11

.20

.09

5

—

-.17

-.33*

-.13

6

—

.75**

-.02

7 8

—

.18 —* p < .05. ** p < .05 with Bonferroni correction.

score for couples in this study (M = 7.0, SD = 4.3)was at the 65th percentile (Straus, 1979), indicatingslightly higher than average conflict between parentsin this study.

The second indicator of marital discord was par-ents' scores on the Dyadic Adjustment Scale (DAS), a32-item valid and reliable measure of marital adjust-ment (Spanier, 1976; Touliatos et al., 1990). Mothers'and fathers' scores were highly correlated (r = .62,p < .001). Like the other family and marital vari-ables, we averaged parents' scores to obtain a meanmarital satisfaction score. The overall marital satis-faction score for couples in this study (M = 111.9,SD - 17.1) was similar to mean scores for marriedcouples in a normative sample (M = 114.8, SD =17.8) and well above scores for divorced couples(M = 70.7, SD = 23.8; Spanier, 1976). There was asignificant negative correlation between parents'marital satisfaction scores and their marital aggres-sion scores (r = - .51, p < .001).

As shown in Table 1, children's perceptions offamily conflict were not related to their parents' re-ports of marital satisfaction (r = -.16) or to maritalaggression (r = .07). However, parents' perceptions offamily conflict were significantly associated with theirreported use of aggression during marital disputes(r = .45> p < .001) and negatively associated withmarital satisfaction (r = - .51, p < .001).

Child psychological adjustment. We argued ear-lier that exposure to and involvement in angry andconflictive family interactions would influence chil-dren's self-perceptions and behavioral adjustment.We therefore assessed three aspects of adjustment inthis study: the child's perception of self and behaviorproblems at home and at school. We used the Self-Perception Profile for Children (Harter, 1985), a 36-

item interview measure, to assess children's percep-tions of their competence in various domains, suchas scholastic, athletic, and social relations (e.g.,"Some kids feel that they are very good at theirschool work, but other kids worry about whether theycan do the school work assigned to them. Which kindof kid is more like you? Is that only sort of true foryou or really true for you?"). Responses to all 36items were summed to create an overall measure ofperceived self-competence.

Parents rated child behavior using the Child Be-havior Checklist (Achenbach, 1978; Achenbach &Edelbrock, 1979). Scores on this 113-item symptomchecklist are based on norms for a child's sex and agegroup (in this case, 6-11-year-olds). Two summaryscores were computed: one for externalizing behaviorproblems, which includes aggressive, hyperactive,and delinquent behavior, and one for internalizingbehavior problems, which includes social withdrawal,somatic complaints, and depressive symptoms. Therewas a high degree of interparental agreement on bothscales (internalizing behavior problems, r = .51, p <.001; externalizing behavior problems, r = .65, p £.001). Mothers' and fathers' data were averaged tocreate a mean parental score for internalizing and ex-ternalizing behavior problems. Most internalizingand externalizing behavior problem scores in thissample were in the normal range, with mean T scoresranging from 52 to 53 for boys and 52 to 55 for girls.As can be seen in Table 1, a child's perception ofhimself or herself as competent was negatively corre-lated with the parents' reports of more externalizingproblems (r = —.33, p < .01) but was not correlatedwith reports of internalizing problems (r = —.17).

Examination of boys' and girls' data separately re-vealed that there were significant sex differences in


the relation between self-reported competence andparents' reports of both internalizing (rs = .14, ns,and -.38, p < .05, for boys and girls, respectively, z =2.11, p < .05) and externalizing (rs = .12, ns, and-.57, p < .01, for boys and girls, respectively, z =3.00, p < .01) behavior problems. In both cases, therewas a significant negative correlation between girls'perceived self-competence and their parents' reportsof behavior problems but nonsignificant correlationsfor boys.

Teachers assessed children's behavior in the class-room using the Conners Teacher Rating Scale (Con-ners, 1969). This 39-item scale assesses a variety ofbehavior problems, such as conduct problems, hyper-activity, and anxious-passive behavior. Responses tothe items are summed to reflect an overall measureof the child's behavioral adjustment at school. Incontrast to the level of agreement between mothersand fathers, teachers' observations of behavior prob-lems at school were not related to the parents' reportsof internalizing and externalizing behavior problems(rs = —.02 and —.18). As noted earlier, low levels ofagreement between raters of child adjustment arecommon (Achenbach et al., 1987; Hinshaw et al.,1992) and, in part, reflect a high degree of situationalspecificity of child behavior problems. As with theparents' ratings, there was a significant negative cor-relation between girls' perceived self-competence andteachers' reports of behavior problems at school (r =—.38, p < .05) but a nonsignificant correlation forboys (r = - .02).

Negative emotional tone in the parent-child rela-tionship. Although our study did not focus onparent-child conflict, the parent questionnaire in-cluded a six-item measure of negative emotional tonein the parent-child relationship. Parents rated howoften feelings of rejection, anger, hostility, tension,and disappointment characterized their interactionswith their child (e.g., "Between us there is a feeling ofanger," "I feel rejected as a parent"). As expected,mothers' descriptions of a negative emotional tone inthe mother-child relationship (Cronbach's alpha =.79) were significantly correlated with their ratings offamily conflict (r = .39, p < .001) and with their rat-ings of child behavior problems (rs = .59 and .36, ps< .001 and .01, for externalizing and internalizingsymptoms, respectively). Similarly, fathers' descrip-tions of a negative emotional tone in the father-childrelationship (Cronbach's alpha = .78) were signifi-cantly correlated with their ratings of family conflict(r = .48, p < .001) and with their ratings of child be-havior problems (rs = .55 and .49, ps < .001, for ex-ternalizing and internalizing symptoms, respectively).

Results

The correlations among the eight main vari-ables in this study are reported in Table 1. Tocontrol for experimentwise error, we computedBonferroni-adjusted probabilities for this 8 x 8Pearson product-moment correlation matrix;these are also reported in Tables 2 and 3. To

address potential problems with respondent bi-ases inflating correlations, we direct the readerto the cross-source correlations for each hypoth-esis. We tested for possible interactions andmain effects of child age and grade level in ouranalyses and found none, which was not surpris-ing given the limited age ranges. Controlling forthe number of siblings also did not change theresults. Thus, the results that are reported herewere not moderated or confounded by child ageor family size.

The first hypothesis predicted that family con-flict would be associated with child maladjust-ment. To test this hypothesis, we computed amultivariate regression with all four child ad-justment measures simultaneously regressed onthe two indicators of family conflict (i.e., chil-dren's ratings and parents' averaged ratings).Overall, there was a significant association be-tween family conflict and child adjustmentproblems (Wilks's lambda = .57), F(4, 62) =11.6, p < .001, with a high canonical correlationof .65 (.59 for boys, .76 for girls).

To control for the possible confounding influ-ence of conflict in the parent-child relationship,we performed another multivariate regression inwhich the four child adjustment outcome mea-sures were simultaneously regressed onto thetwo measures of family conflict (i.e., child rat-ings and parents' averaged ratings) and the twomeasures of negative emotional tone in themother-child and father-child relationships.The independent effects of the two types of vari-ables (i.e., family conflict and parent-child nega-tive emotional tone) were examined. After con-trolling for negative emotional tone in theparent-child relationship, we found that theassociation between family conflict and childmaladjustment remained significant (Wilks'slambda = .68), F(4, 53) = 6.24, p < .001, with ahigh canonical correlation of .57 (.45 for boys,.77 for girls). Similarly, after controlling for thetwo family conflict variables, we found a signifi-cant association between the parents' ratings of anegative emotional tone in their relationshipswith their child and the child's maladjustment(Wilks's lambda = .64), F(4, 53) = 7.47, p <.001, with a high canonical correlation of .60(.51 for boys, .70 for girls). In short, althoughfeelings of anger, hostility, and rejection in theparent-child relationship were associated withchild maladjustment, they did not account for


Table 2Pearson Correlations Between Conflict and Psychological Adjustment for Girls Only

Variable « 1 2 3 4 5 6


conflict2. Parents' average report of

family conflictMarital discord measures

3. Parents' average report ofmarital satisfaction

4. Parents' average report ofverbal aggression withspouse


self-competence6. Parents' average report of

internalizing behavior7. Parents' average report of

externalizing behavior8. Teacher's report of behavior

problems

40 —

35 .49* —

35 -.10 -.55** —

35 -.09 .36* -.54** —

40 -.68** -

35 .37*

35 .39*

40 .16

59**

48*

67**

25

.16

-.15

-.35*

-.12

.09

.15

.21

-.01

—

-.38*

-.57**

-.38*

—

.80**

.01

—

.28* p < .05. ** p < .05 with Bonferroni correction.

the strong association found between ratings offamily conflict and child maladjustment.

Follow-up multiple regressions in which eachof the four child adjustment outcome measureswere analyzed separately revealed a similar pat-tern of results. Of particular interest was the ro-bustness of the association between family con-flict and children's externalizing behaviorproblems. Parents' ratings of externalizing prob-lems remained significantly associated withfamily conflict (whether rated by parents or chil-dren) in regressions that controlled for a nega-tive emotional tone during mother-child andfather-child interactions.

The robust association between the two fam-ily conflict variables and the set of child adjust-ment outcomes seemed to warrant further ex-ploration of the individual correlations. Table 1shows that all eight of the correlations betweentwo measures of family conflict and four mea-sures of psychological adjustment were in thepredicted direction and that six of the correla-tions reached statistical significance. Parents'reports of family conflict were significantly re-lated to all four measures of child adjustment.Children's perceptions of family conflict wererelated to their reports of lower self-competenceand to the parents' reports of more externaliz-ing behavior problems. Three of the five cross-source correlations were significant. Children

from homes that were characterized by parentsas being more conflictive perceived themselvesto be less competent, and their teachers de-scribed more behavior problems at school.Children from homes that they themselvescharacterized as more conflictive were de-scribed by their parents as displaying more ex-ternalizing behavior problems. Only two of thecorrelations were significant with the Bonfer-roni correction, and both of these were basedon data from the same informant. Childrenwho described a more conflictive family envi-ronment also reported less perceived self-competence, and parents who described a highlevel of conflict among family members ob-served more externalizing behaviors in theirchild.

The results of separate analyses of girls' andboys' data are presented in Tables 2 and 3. Us-ing only girls' data, there were six statisticallysignificant correlations. Three of them were sig-nificant with the Bonferroni correction, includ-ing one cross-source correlation: a negative as-sociation between parents' ratings of familyconflict and girls' perceived self-competence.Using boys' data, there were three statisticallysignificant correlations and none with the Bon-ferroni correction. Comparisons of the indi-vidual correlation coefficients revealed onlyone significant sex difference. The parent mea-


Table 3Pearson Correlations Between Conflict and Psychological Adjustment for Boys Only

Variable 1


conflict 32 —2. Parents' average report of

family conflict 32Marital discord measures

3. Parents' average report ofmarital satisfaction 32

4. Parents' average report ofverbal aggression withspouse 32


self-competence 326. Parents' average report of

internalizing behavior 327. Parents' average report of

externalizing behavior 328. Teacher's report of behavior

problems 32

.37*

-.31

.39*

-.44*

-.01

.25

.11

-.50*

.52*

.01

.15

.47*

.36*

—

-.47*

.12

.05

-.18

-.33

—

-.02

.04

.22

.10

—

.14

.12

-.02

—

.68** —

-.10 .26

* p < .05. ** p < .05 with Bonferroni correction.

sure of family conflict was correlated with girls'perceived self-competence (r = -.59, p < .01),but not with boys' self-competence (r = —.01, z= 2.61,p<.05).'

The second hypothesis predicted that childadjustment would be more strongly associatedwith family conflict than with marital conflict.There were two measures of marital discord inthis study: parents' averaged reports of maritalsatisfaction and their averaged reports of verbalaggression. In a multivariate regression analysis,we simultaneously regressed all four child ad-justment measures on the two marital discordvariables, revealing a nonsignificant association(Wilks's lambda = .97), F(4, 62) = 0.42 (theoverall canonical correlation was .16; rs = .15and .21 for boys and girls, respectively). Asshown in Table 1, all eight of the correlationsbetween the two measures of marital discord(i.e., satisfaction and aggression) and four mea-sures of child adjustment were fairly low (corre-lations ranged from .04 to .29), and only onereached statistical significance. Parents who re-ported low levels of marital satisfaction also de-scribed their child as having more externalizingbehavior problems. There were no significantsex differences in the pattern of correlations.

Additional analyses directly compared thetwo marital discord variables with the two fam-ily conflict measures (i.e., children's reports and

parents' averaged reports) as correlates of childadjustment.2 First, all four adjustment variables

1 The significantly stronger association between par-ents' reports of family conflict and lowered perceivedself-competence in girls occurred for three of the sixdomain-specific subscales on Harter's (1985) Self-Perception Profile for Children: Perceptions of Scho-lastic Competence, Physical Appearance, and GlobalSelf-Worth. On the other hand, there were no sex dif-ferences found for the remaining subscales (i.e., SocialCompetence, Athletic Competence, and BehavioralConduct).

2 The family conflict measure included two itemsassessing physical aggression (i.e., "Family memberssometimes hit each other," "Family members some-times get so upset they throw things"). Because thePhysical Aggression subscale was omitted from themeasure of marital aggression (i.e., the Conflict Tac-tics Scale), a direct comparison of these two variablesmay be problematic. Results suggesting the relativestrength of family conflict could reflect the importanceof the two physical aggression items rather than thesignificance of a general family climate of conflict. Wetherefore constructed an abbreviated family conflictscale omitting the two physical aggression items. Theabbreviated scale was highly correlated with the fullfamily conflict scale (rs = .97 and .96, ps < .001, forchildren and average parent ratings, respectively). Wethen recomputed all of the analyses presented in thisarticle using the abbreviated family conflict scale andfound no changes in the results. Therefore, the physi-cal aggression items on the family conflict scale do notaccount for the results presented in this section or inany other section of this article.


Table 4Changes in R2 Associated With Parents' Reports of Family Conflict and Marital Discord asPredictors of Child Adjustment

R2 change: R2 change:Parents' reports Parents' reports

of family conflict1" of marital discord0Child adjustment

outcome Total R2

Self-competenceParents' ratings of internalizing behavior problemsParents' ratings of externalizing behavior problemsTeacher-reported behavior problems

.13*

.11

.30***

.13*

.10**

.10**

.22***

.08**

.04

.01

.003

.01a Total variance accounted for by all three predictor variables (i.e., parents' reports of family conflict and the twomarital discord variables [parents' reports of marital satisfaction and marital aggression]). b Unique varianceaccounted for by the parents' reports of family conflict after controlling for the two marital discord variables.c Unique variance accounted for by the two marital discord variables after controlling for the parents' reports offamily conflict.*p<.05. **p<.0\. ***/>< .001.

were regressed onto the four conflict variables(i.e., two measures of family conflict and twomeasures of marital discord), revealing a signifi-cant canonical correlation of .43 (Wilks'slambda = .82), ^(4, 62) = 3.44, p < .05 (rs = .45and .52 for boys and girls, respectively). Theindependent contributions of each of the twotypes of conflict variables (i.e., marital discordand family conflict) were examined next. Thetwo family conflict variables remained signifi-cantly associated with child adjustment prob-lems after the level of marital discord had beencontrolled (Wilks's lambda = .56), F(4, 61) =12.0, p < .001 (the overall canonical correlationwas .66; rs = .61 and .76 for boys and girls,respectively). By contrast, marital discord wasnot associated with child adjustment after con-trolling for family conflict (Wilks's lambda =.95), F(4, 61) = 0.77 (the overall canonical cor-relation was .22; rs = .24 and .21 for boys andgirls, respectively).

Hierarchical multiple regressions were alsocomputed in which each of the four child adjust-ment outcome measures were separately re-gressed onto a set of three variables: a singlefamily conflict variable (either the parents' rat-ing or the child's rating) and the two maritaldiscord variables. In the first group of regres-sions, we used parents' ratings as the measure offamily conflict. The results are presented inTable 4. Even after controlling for the maritalvariables, parents' descriptions of a climate ofconflict in the home remained significantly as-sociated with all four measures of child adjust-ment, including the two cross-source associa-tions with children's perceived self-competence

and teachers' reports of behavior problems. Themarital discord variables, on the other hand, didnot make a unique contribution to any of theadjustment measures.

Children's perceptions of family conflict wereused in the second group of regressions. Theresults are presented in Table 5. After control-ling for parents' reports of marital discord, wefound that children's descriptions of a climate ofconflict in the home remained significantly as-sociated with the two measures of child adjust-ment that were significant in the zero-ordercorrelations: perceived self-competence and par-ents' ratings of externalizing behavior problems.Again, the two marital discord variables did notmake a unique contribution to any of the adjust-ment outcomes.

In sum, although family conflict was a reliablecorrelate of child maladjustment, marital satis-faction and marital aggression were not. More-over, family conflict was correlated with childadjustment independent of the level of discordin the parents' marriage and in the parent-childrelationship.

Discussion

The evidence supported the two hypothesesabout the relation between interpersonal conflictat home and child adjustment. First, preadoles-cent children in families that were characterizedby more open expressions of anger and conflictappeared to be less well adjusted. In particular,they were more likely to have poor perceptionsof themselves and to display externalizing be-


Table 5Changes in R2 Associated With Children's Reports of Family Conflict and Parents' Reports ofMarital Discord as Predictors of Child Adjustment

Child adjustmentoutcome

Self-competenceParents' reports of internalizing behavior problemsParents' reports of externalizing behavior problemsTeacher-reported behavior problems

Total R2*

.34***

.06

.18**

.05

R2 change:Children's reportsof family conflict1*

.31***

.05

.09**

.001

R2 change:Parents' reports

of marital discord0

.002

.01

.06

.05a Total variance accounted for by all three predictor variables (i.e., children's reports of family conflict and thetwo marital discord variables [parents' reports of marital satisfaction and marital aggression]). b Uniquevariance accounted for by the children's reports of family conflict after controlling for the two marital discordvariables. c Unique variance accounted for by the two marital discord variables after controlling for thechildren's reports of family conflict.*p< .05 . **p<.01 . ***/><.001.

havior problems at home. Findings from cross-sectional and longitudinal research with adoles-cents was thereby replicated in our sample of8-11-year-old preadolescents.

The link between family conflict and childmaladjustment appeared to be fairly robust. Forexample, analyses that controlled for a negativeemotional tone in the parent-child relationshipsuggested that the family conflict variable wasnot simply a proxy for aversive parent-child in-teractions. Of course, we were not able to deter-mine the extent to which children's behaviorproblems contributed to perceptions of angerand aggression in the family. Although our pri-mary interest was in child adjustment as an out-come of exposure to interpersonal conflict athome, we are mindful of the significant rolethat children play in shaping negative familyinteractions (Christensen, Phillips, Glasgow, &Johnson, 1983).

The cross-source correlations also providedconvincing support for the first hypothesis. Chil-dren's descriptions of family conflict were asso-ciated with parents' reports of externalizing be-havior problems, and parents' descriptions offamily conflict were associated with both teach-ers' reports of problems at school and children'sperceived self-competence. Although not all ofthe cross-source correlations survived the Bon-ferroni correction, there was a reliable correla-tion between parents' reports of family conflictand girls' perceptions of self-competence.

The stronger support among girls for the firsthypothesis, at least with regard to their self-perceptions, is consistent with other research in-dicating that a family emotional climate of anger

has a greater impact on daughters' emotionaldevelopment (Eisenberg et al., 1992). The effectsof family conflict on girls clearly warrant furtherinvestigation. Evidence that daughters are morelikely than sons to intervene in sibling-siblingand parent-sibling conflicts suggests one prom-ising avenue for exploration (Vuchinich et al.,1988).

As expected, family conflict was a much stron-ger correlate of child adjustment than was mari-tal discord. The generally nonsignificant asso-ciation in our study between measures ofmarital discord and child adjustment is not un-usual. According to a recent review, fewer thanhalf of the correlations between interparentalconflict and child behavior problems are signifi-cant in a majority of studies (Reid & Crisafulli,1990). Multiple regressions using parents' de-scriptions of anger in both their marriage andtheir family as predictors of child adjustmentprovided revealing comparisons of family andmarital conflict. On the one hand, the parents'descriptions of a conflictive climate at home wasindependently associated with all four child ad-justment measures. Their report of discord intheir marriage, on the other hand, was not asso-ciated with child maladjustment in any of theanalyses. Unfortunately, because we did not in-clude children's reports of marital conflict, wewere unable to directly compare children's per-ceptions of marital and family conflict.

We certainly do not wish to make the argu-ment that marital conflict is not detrimental tochildren. Our data do suggest, however, that in anonclinic sample general behavioral norms inthe family for acceptable ways of handling anger


and conflict may be more significant for childadjustment than is discord within the parents'marriage. For example, parents in a troubledmarriage may be able to keep their childrenfrom observing their conflicts, thereby shieldingthem from the harmful effects of open maritaldiscord. This possibility was suggested both bythe nonsignificant association between maritaldiscord and child adjustment and by the patternof correlations between parents' and children'sperceptions of the family. As expected, parents'reports of discord in their marriage were corre-lated with their descriptions of a conflictive cli-mate in their home, and parents and childrenagreed in their ratings of family conflict. How-ever, children's perceptions of conflict at homewere independent of their parents' reports ofmarital discord.

To accurately identify the effects of maritalconflict on children's psychological adjustment,it seems critical to assess both the extent towhich (a) a child actually observes marital dis-putes and (b) open expressions of anger and ag-gression are part of the general social climate inthe home. The significant correlations betweenparents' ratings of family conflict and their rat-ings of marital problems (on the basis of mea-sures that are used frequently in this field) high-light the need to assess both dimensions offamily life. When parents describe high levels ofinterparental discord, the reported anger and ag-gression is not necessarily a characteristic of themarital relationship alone. The emphasis onmarital conflict in the literature might havemasked the effects of family systems variablesand limited our understanding of the waysin which anger and conflict at home affectchildren.

The need for a more systematic follow-up tothis study is clear. For example, during the in-terviews, many children indicated that theywere responding to the family conflict itemswith sibling conflict in mind. Furman andBuhrmester (1985) similarly found that childrenreported more conflict with siblings than withparents. It is important to understand how mari-tal, parent-child, and sibling conflicts contributeto family norms for expressions of anger andaggression and how these different categories offamily conflict influence important develop-mental outcomes.

References

Achenbach, T. M. (1978). The child behavior profile:I. Boys aged 6-11. Journal of Consulting and Clini-cal Psychology, 46, 478-488.

Achenbach, T. M, & Edelbrock, C. S. (1979). Thechild behavior profile: II. Boys aged 12-16 and girlsaged 6-11 and 12-16. Journal of Consulting andClinical Psychology, 41, 223-233.

Achenbach, T. M, McConaughy, S. H., & Howell, C.T. (1987). Child/adolescent behavioral and emo-tional problems: Implications of cross-informantcorrelations for situational specificity. PsychologicalBulletin, 101, 213-232.

Asarnow, J. R. (1992). Suicidal ideation and attemptsduring middle childhood: Associations with per-ceived family stress and depression among childpsychiatric inpatients. Journal of Clinical Child Psy-chology, 21, 35-40.

Belsky, J. (1985). The determinants of parenting: Aprocess model. Child Development, 55, 83-96.

Borrine, M. L., Handal, P. J., Brown, N. Y., & Sea-right, H. R. (1991). Family conflict and adolescentadjustment in intact, divorced, and blended fami-lies. Journal of Consulting and Clinical Psychology,59, 753-755.

Burman, B., John, R. S., & Margolin, G. (1987). Ef-fects of marital and parent-child relations on chil-dren's adjustment. Journal of Family Psychology, 1,91-108.

Christensen, A., & Margolin, G. (1988). Conflict andalliance in distressed and non-distressed families. InR. A. Hinde & J. Stevenson-Hinde (Eds.), Relation-ships within families: Mutual influences (pp. 263-282). Oxford, England: Clarendon Press.

Christensen, A., Phillips, S., Glasgow, R. E., &Johnson, S. M. (1983). Parental characteristics andinteractional dysfunction in families with child be-havior problems: A preliminary investigation. Jour-nal of Abnormal Child Psychology, 11, 153-166.

Conger, R. D., Conger, K. J., Elder, G. H., Jr., Lorenz,F. O., Simons, R. L., & Whitbeck, L. B. (1992). Afamily process model of economic hardship and ad-justment in early adolescent boys. Child Develop-ment, 63, 526-541.

Conners, C. K. (1969). A teacher rating scale for use indrug studies with children. American Journal of Psy-chiatry, 126, 152-156.

Eisenberg, N., Fabes, R. A., Carlo, G., Troyer, D.,Speer, A. L., Karbon, M., & Switzer, G. (1992). Therelations of maternal practices and characteristics tochildren's vicarious emotional responsiveness.Child Development, 63, 583-602.

Fauber, R., Forehand, R., Thomas, A. M., & Wierson,M. (1990). A mediational model of the impact ofmarital conflict on adolescent adjustment in intactand divorced families: The role of disrupted parent-ing. Child Development, 61, 1112-1123.

Fincham, F. D., & Osborne, L. N. (1993). Maritalconflict and children: Retrospect and prospect.Clinical Psychology Review, 13, 75-88.

Furman, W., & Buhrmester, D. (1985). Children'sperceptions of the personal relationships in their


social networks. Developmental Psychology, 21,1016-1024.

Grych, J. H., & Fincham, F. D. (1990). Marital conflictand children's adjustment: A cognitive-contextualframework. Psychology Bulletin, 108, 267-290.

Grych, J. H., Seid, M., & Fincham, F. D. (1992). As-sessing marital conflict from the child's perspective:The Children's Perception of Interparental ConflictScale. Child Development, 63, 558-572.

Harter, S. (1985). Manual for the Self-Percept ion Pro-file for Children. Denver, CO: University of Denver.

Hinshaw, S. P., Han, S. S., Erhardt, D., & Huber, A.(1992). Internalizing and externalizing behaviorproblems in preschool children: Correspondenceamong parent and teacher ratings and behavior ob-servations. Journal of Clinical Child Psychology, 21,143-150.

Kleinman, S. L., Handal, P. J., Enos, D., Searight, H.R., & Ross, M. J. (1989). Relationship between per-ceived family climate and adolescent adjustment.Journal of Clinical Child Psychology, 18, 351-359.

Kronenberger, W. G., & Thompson, R. J., Jr. (1990).Dimensions of family functioning in families withchronically ill children: A higher order factor analy-sis of the Family Environment Scale. Journal ofClinical Child Psychology, 19, 380-388.

Kurdek, L. A., & Sinclair, R. J. (1988). Adjustment ofyoung adolescents in two-parent nuclear, stepfather,and mother-custody families. Journal of Consultingand Clinical Psychology, 56, 91-96.

Margolin, G. (1981). The reciprocal relationship be-tween marital and child problems. In J. P. Vincent(Ed.), Advances in family intervention, assessmentand theory (Vol. 2, pp. 131-182). Greenwich, CT:JAI Press.

Marvin, R. S., & Stewart, R. B. (1990). A family sys-tems framework for the study of attachment. In M.T. Greenberg, D. Cicchetti, & E. M. Cummings(Eds.), Attachment in the preschool years: Theory,research, and intervention (pp. 51-86). Chicago:University of Chicago Press.

Mechanic, D., & Hansell, S. (1989). Divorce, familyconflict, and adolescents' well-being. Journal ofHealth and Social Behavior, 30, 105-116.

Moos, R. F., & Moos, B. S. (1981). Family Environ-ment Scale: Manual. Palo Alto, CA: Consulting Psy-chologists Press.

Peterson, J. L., & Zill, N. (1986). Marital disruption,parent-child relationships, and behavior problemsin children. Journal of Marriage and the Family, 48,295-307.

Reid, W. J., & Crisafulli, A. (1990). Marital discordand child behavior problems: A meta-analysis. Jour-nal of Abnormal Child Psychology, 18, 105-117.

Repetti, R. L. (1987). Individual and common com-ponents of the social environment at work and psy-chological well being. Journal of Personality andSocial Psychology, 52, 710-720.

Slater, E. J., & Haber, J. D. (1984). Adolescent adjust-ment following divorce as a function of familial con-flict. Journal of Consulting and Clinical Psychology,52, 920-921.

Spanier, G. B. (1976). Measuring dyadic adjustment:New scales for assessing the quality of marriage andother similar dyads. Journal of Marriage and theFamily, 38, 15-28.

Straus, M. A. (1979). Measuring intra-family conflictand violence: The Conflict Tactics (CT) Scales.Journal of Marriage and the Family, 41, 75-88.

Touliatos, J., Perlmutter, B. F., & Straus, M. A. (1990).Handbook of family measurement techniques. New-bury Park, CA: Sage.

Tschann, J. M., Johnston, J. R., Kline, M., & Waller-stein, J. S. (1989). Family process and children'sfunctioning during divorce. Journal of Marriageand the Family, 51, 431-444.

Vuchinich, S., Emery, R. E., & Cassidy, J. (1988).Family members as third parties in dyadic familyconflict: Strategies, alliances, and outcomes. ChildDevelopment, 59, 1293-1302.

Weidner, G., Hutt, J., Conner, S. L., & Mendell, N. R.(1992). Family stress and coronary risk in children.Psychosomatic Medicine, 54, 471-479.

Weiss, B., Dodge, K. A., Bates, J. E., & Pettit, G. S.(1992). Some consequences of early harsh disci-pline: Child aggression and a maladaptive social in-formation processing style. Child Development, 63,1321-1335.

Received December 19, 1992Revision received May 13, 1993

Accepted May 27, 1993

conflict in families and the psychological adjustment of …repetti.bol.ucla.edu/jaycox repetti...

Documents