Topic: divorce and its impact on children

the_impact_of_divorce_on_children.pdf

Home >Social Science homework help >Sociology homework help >Topic: divorce and its impact on children

The Impact of Divorce on Children Author(s): David H. Demo and Alan C. Acock Source: Journal of Marriage and Family, Vol. 50, No. 3 (Aug., 1988), pp. 619-648 Published by: National Council on Family Relations Stable URL: https://www.jstor.org/stable/352634 Accessed: 25-04-2019 06:19 UTC

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide

range of content in a trusted digital archive. We use information technology and tools to increase productivity and

facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at

https://about.jstor.org/terms

National Council on Family Relations is collaborating with JSTOR to digitize, preserve and extend access to Journal of Marriage and Family

This content downloaded from 132.174.254.81 on Thu, 25 Apr 2019 06:19:15 UTC All use subject to https://about.jstor.org/terms

DAVID H. DEMO Virginia Polytechnic Institute and State University

ALAN C. ACOCK Louisiana State University

The Impact of Divorce on Children

With the acceleration of the divorce rate from the mid-1960s to the early 1980s, the number of non- traditional families (such as single-parent families and reconstituted families) have increased relative to intact, first-time nuclear families. This article reviews empirical evidence addressing the rela- tionship between divorce, family composition, and children's well-being. Although not entirely consistent, the pattern of empirical findings sug- gests that children's emotional adjustment, gender-role orientation, and antisocial behavior are affected by family structure, whereas other dimensions of well-being are unaffected. But the review indicates that these findings should be in- terpreted with caution because of the meth- odological deficiencies of many of the studies on which these findings are based. Several variables, including the level of family conflict, may be cen- tral variables mediating the effect offamily struc- ture on children.

High divorce rates in the United States over the past 20 years have resulted in numerous changes in American family life, with perhaps the most important consequences bearing on children whose families were disrupted. In 1970, 12% of American families with children under age 18 were headed by single parents. By 1984, one- fourth of American families and nearly 60% of

Department of Sociology, Virginia Tech, Blacksburg, VA 24061.

*Department of Sociology and Center for Life Cycle and Population Studies, Louisiana State University, Baton Rouge, LA.70803.

black families were headed by single parents (see Table 1). Millions of other children live in two- parent but reconstituted families, separated from at least one biological parent. In fact, Fursten- berg, Nord, Peterson, and Zill's recent analysis (1983) indicates that less than two-thirds of American children live with both biological parents.

A number of studies use recent social and

demographic trends to predict children's future living arrangements, and while these predictions vary, the consensus is that most youth will spend some time prior to age 18 in a single-parent household (Bumpass, 1984, 1985; Furstenberg et al., 1983; Hofferth, 1985, 1986; Norton and Glick, 1986). Hofferth (1985) suggests that the percentage of black youth who will live with one parent for some period of time prior to age 18 may be as high as 94%, while for white children the corresponding figure is 70%. Norton and Glick's (1986) analysis yields a lower estimate but still projects that 60%o of American children will live in a single-parent family before reaching age 18.

These trends in family composition have major implications for the life course of children and their well-being. The purpose of this article is to review and assess recent empirical evidence on the impact of divorce on children, concentrating on studies of nonclinical populations published in the last decade. We also direct attention to a number

of important theoretical and methodological con- siderations in the study of family structure and youthful well-being. We begin by briefly describ- ing some of the theoretical propositions and assumptions that guide research in this area.

Journal of Marriage and the Family 50 (August 1988): 619-648 619

This content downloaded from 132.174.254.81 on Thu, 25 Apr 2019 06:19:15 UTC All use subject to https://about.jstor.org/terms

Journal of Marriage and the Family

TABLE 1. NUMBER AND PERCENTAGE OF FAMILIES WITH CHILDREN UNDER AGE 18, 1970-1984 (IN THOUSANDS)

Families 1970 Totala 1980 Totala 1984 Totala 1984 Whites 1984 Blacks

Total, children under 18 29,631 32,150 33,246 27,508 4,744

Two-parent 25,823 25,231 24,701 21,978 1,934 One-parent 3,808 6,920 8,544 5,529 2,809 Mother-headed 3,415 6,230 7,599 4,766 2,652 Never-married 248 1,063 2,102 729 1,332 Spouse absent 1,377 1,743 1,762 1,120 597 Separated 962 1,483 1,505 929 544 Other 415 260 257 191 53

Divorced 1,109 2,721 3,174 2,516 593 Widowed 682 703 561 401 128

Father-headed 393 692 945 763 157 Never-married 22 63 166 109 52 Spouse absent 247a 181 187 139 39 Divorced na 340 496 445 44 Widowed 124 107 97 69 23

1970 Total % 1980 Total % 1984 Total %o 1984 White %o 1984 Black %

Two-parent 87.1 78.5 74.3 79.9 40.8 One-parent 12.9 21.5 25.7 20.1 59.2 Mother-headed 11.5 19.4 22.9 17.3 55.9 Never-married 0.8 3.3 6.3 2.7 28.1 Spouse absent 4.6 5.4 5.3 4.1 12.6 Separated 3.2 4.6 4.5 3.4 11.5 Other 1.4 0.8 0.8 0.7 1.1

Divorced 3.7 8.5 9.5 9.1 12.5 Widowed 2.3 2.2 1.7 1.5 2.7

Father-headed 1.3 2.2 2.8 2.8 3.3 Never-married 0.1 0.2 0.5 0.4 1.1 Spouse absent 0.8 0.6 0.6 0.5 0.8 Divorced na 1.1 1.5 1.6 0.9 Widowed 0.4 0.3 0.3 0.3 0.5

Source: U.S. Bureau of the Census, Population Reports, Series P-20, No. 138. aTotals include other races not shown separately in the table. bThe 1970 figure includes families headed by divorced men.

THEORETICAL UNDERPINNINGS

Consistent with the Freudian assumption that a two-parent group constitutes the minimal unit for appropriate sex-typed identification (Freud, 1925/1961), anthropologists, sociologists, and social psychologists have long maintained the necessity of such a group for normal child development. Representative of structural- functional theorizing, Parsons and Bales (1955: 16-17) argued that one of the basic functions of the family is to serve as a stable, organically in- tegrated "factory" in which human personalities are formed.

Similarly, social learning theory emphasizes the importance of role models, focusing on parents as the initial and primary reinforcers of child behavior (Bandura and Walters, 1963). Much of the research adopting this perspective centers on parent-child similarities, analyzing the transmission of response patterns and the in-

hibitory or disinhibitory effect of parental models. The presence of the same-sex parent is assumed to be crucial in order for the child to

learn appropriate sex-typed behavior. This assumption is shared by developmental and sym- bolic interactionist theories, various cognitive ap- proaches to socialization, and confluence theory, as well as anthropological theories (Edwards, 1987).

It logically follows that departures from the nuclear family norm are problematic for the child's development, especially for adolescents, inasmuch as this represents a crucial stage in the developmental process. Accordingly, a large body of research literature deals with father absence, the effects of institutionalization, and a host of "deficiencies" in maturation, such as those hav- ing to do with cognitive development, achieve- ment, moral learning, and conformity. This focus has pointed to the crucial importance of both

620

This content downloaded from 132.174.254.81 on Thu, 25 Apr 2019 06:19:15 UTC All use subject to https://about.jstor.org/terms

Impact of Divorce on Children

parents' presence but also has suggested that cer- tain causes for parental absence may accentuate any negative effects. Lynn, for example, asserts (1974: 279):

The research on the relationship between father absence and the general level of the child's ad- justment reveals that the loss of a father for any reason is associated with poor adjustment, but that absence because of separation, divorce, or desertion may have especially adverse effects.

Some researchers suggest even more dire out- comes whenever parental separation, divorce, or desertion occur. Among these are vulnerability to acute psychiatric disturbances, the child's aver- sion to marriage, and proneness to divorce once they do marry (Anthony, 1974). In sum, two general propositions are suggested:

1. Children reared in households where the

two biological parents are not present will exhibit lower levels of well-being than their counterparts in intact nuclear families.

2. The adverse effects on youthful well-being will be especially acute when the cause of parental absence is marital separa- tion, divorce, or desertion.

Divorce and Family Structure

In examining research that addresses these two propositions, it is important to distinguish be- tween studies investigating the effects of family structure and those investigating the effects of divorce. Most studies compare intact units and single-parent families, guided by the assumption that the latter family structure is precipitated by divorce. Of course, this is not always the case. Single-parent families consist of those with parents who have never married, those formed by the permanent separation of parents, and those precipitated by the death of a parent. Simple com- parisons between one- and two-parent families are also suspect in that two-parent families are not monolithic. First-time or nondivorced units differ

from divorced, remarried units in which step- parents are involved. In addition, little recog- nition has been given to the fact that families of different types may exhibit varying levels of insta- bility or conflict, a potentially confounding variable in establishing the effects of family struc- ture. In short, most investigations of the linkage between family structure and youthful well-being have failed to recognize the complexity of present- day families.

While family composition is a critical con- sideration in assessing the impact of divorce on children, we must also examine the unique events, disruptions, and transitions characterizing the divorce process that are not experienced by children and other members of nondivorced

families. In particular, there are significant changes in family composition, parent-child in- teraction, discipline, and socioeconomic cir- cumstances, as well as the emotional reactions that parents and children have to divorce. These events are accompanied by changes in ex- trafamilial relations and social networks, often as a result of stigma attached to divorced parents and their children. Although stepfamilies are beyond the scope of this review, researchers must also distinguish the consequences of divorce from those of remarriage and subsequent changes in family composition (see Ganong and Coleman, 1984, for a review of the emerging literature on reconstituted families and their impact on children).

Bearing in mind these conceptual distinctions, we now move to a systematic review of recent evidence on the impact of divorce on children and adolesecents.

EXISTING RESEARCH

A substantial amount of research has examined

the effects of family structure on children's social and psychological well-being. Many studies docu- ment negative consequences for children whose parents divorce and for those living in single- parent families. But most studies have been con- cerned with limited dimensions of a quite complex problem. Specifically, the research to date has typically (a) examined the effects of divorce or father absence on children, ignoring the effects on adolescents; (b) examined only selected dimen- sions of children's well-being; (c) compared intact units and single-parent families but not recog- nized important variations (e.g., levels of marital instability and conflict) within these structures; and (d) relied on cross-sectional designs to assess developmental processes.

Social and psychological well-being includes aspects of personal adjustment, self-concept, in- terpersonal relationships, antisocial behavior, and cognitive functioning. It should be noted that some of these variables (e.g., personal adjust- ment) have been the subject of voluminous

621

This content downloaded from 132.174.254.81 on Thu, 25 Apr 2019 06:19:15 UTC All use subject to https://about.jstor.org/terms

Journal of Marriage and the Family

research, while others (e.g., interpersonal rela- tions) have received relatively little attention. In Tables 2 to 6 we outline selected studies published since 1975 that were designed to compare the well- being of children and adolescents living in intact families and families disrupted by divorce.'

Personal Adjustment

Personal adjustment is operationalized in various ways by different investigators but includes such variables as self-control, leadership, responsibili- ty, independence, achievement orientation, ag- gressiveness, and gender-role orientation (see Table 2). As we see when examining the 16 studies outlined in Table 2, there are also wide variations in sample size and composition. But the overall pattern of empirical findings suggests temporary deleterious effects of parental divorce on children's adjustment, with these effects most common among young children (Desimone-Luis, O'Mahoney, and Hunt, 1979; Hetherington, Cox, and Cox, 1979; Kurdek, Blisk, and Siesky, 1981; Wallerstein and Kelly, 1975, 1980a).2 Kurdek and Siesky (1980b, c) suggest that older children ad- just more readily because they are more likely to discuss the situation with friends (many of whom have had similar experiences), to understand that they are not personally responsible, to recognize the finality of the situation, to appreciate both parents for their positive qualities, and to recognize beneficial consequences such as the end of parental fighting and improved relations with parents.

On the basis of her review of research con-

ducted between 1970 and 1980, Cashion (1984: 483) concludes: "The evidence is overwhelming that after the initial trauma of divorce, the children are as emotionally well-adjusted in these [female-headed] families as in two-parent families." Investigations of long-term effects (Acock and Kiecolt, 1988; Kulka and Weingarten, 1979) suggest that, when socioeconomic status is controlled, adolescents who have experienced a parental divorce or separation have only slightly lower levels of adult adjustment.

In two other studies Kinard and Reinherz

(1984, 1986) observed elementary school children in three different family situations (never- disrupted; disrupted prior to starting school; and recently disrupted) and found that children in recently disrupted families suffered pronounced

and multidimensional effects: problems in atten- tiveness at school, lowered academic achievement, withdrawal, dependency, and hostility. While their findings are not definitive, Kinard and Reinherz speculate that either "the effects of parental divorce on children diminish over time; or that the impact of marital disruption is less severe for preschool-age children than for school- age children" (1986: 291). Children's age at the time of disruption may also mediate the impact of these events on other dimensions of their well-

being (e.g., self-esteem or gender-role orientation) and thus will be discussed in greater detail below (also, see Rohrlich, Ranier, Berg-Cross, and Berg- Cross, 1977, for a clinical perspective on the im- pact of divorce on children of different ages). But two variables that critically affect children's ad- justment to divorce are marital discord and children's gender.

Marital discord. A significant pattern in the em- pirical literature is that personal adjustment, like other dimensions of well-being, is not related to family structure but is adversely affected by parental discord (Ellison, 1983; Rosen, 1979). Kurdek and Siesky's (1980b) extensive data on children who had experienced their parents' divorce indicated that, although learning of the divorce and adjusting to the loss of the non- custodial parent were painful, children indicated that these adjustments were preferable to living in conflict. Many studies report that children's ad- justment to divorce is facilitated under conditions of low parental conflict-both prior to and subse- quent to the divorce (Guidubaldi, Cleminshaw, Perry, Nastasi, and Lightel, 1986; Jacobson, 1978; Lowenstein and Koopman, 1978; Porter and O'Leary, 1980; Raschke and Raschke, 1979; Rosen, 1979).

Children's gender. Children's gender may be especially important in mediating the effects of family disruption, as most of the evidence sug- gests that adjustment problems are more severe and last for longer periods of time among boys (Hess and Camara, 1979; Hetherington, 1979; Hetherington, Cox, and Cox, 1978, 1979, 1982; Wallerstein, 1984; Wallerstein and Kelly, 1980b).3 Guidubaldi and Perry (1985) found, controlling for social class, that boys in divorced families manifested significantly more maladaptive symp- toms and behavior problems than boys in intact

622

This content downloaded from 132.174.254.81 on Thu, 25 Apr 2019 06:19:15 UTC All use subject to https://about.jstor.org/terms

TABLE 2. SUMMARY OF RECENT STUDIES ON EFFECTS OF DIVORCE ON PERSONAL ADJUSTMENT OF CHILDREN AND ADOLESCENTS

Description Study of Sample

Grossman et al. (1980)

Guidubaldi and Perry (1985)

Hainline and Feig (1978)

Hess and Camara (1979)

Hetherington, Cox, and Cox (1979)

294 white male and female college students

365 boys and 334 girls in grades 1, 3, and 5 (Tl); grades 2, 3, 5, and 7 at follow-up (T2) (National Association of School Psychologists -Kent State University data)

36 female college stu- dents aged 17-23 (mean age = 18.5); 80% white, 20%7o black

32 white boys and girls aged 9-11

48 white boys and 48 white girls; mean age = 3.9 at Tl and 5.8 at T2

Family Structure of Respondents

262 students from in- tact families; 24 with history of parental divorce and remarriage; 8 from divorced, non- remarried families

341 children with divorced parents and 358 from intact families

12 women whose fathers had died (6 be- fore age 5, 6 between ages 5-11); 12 women whose parents had divorced (6 before age 5, 6 between ages 5-11); 12 women from intact families

16 children in intact families; 16 children in recently divorced families

24 boys and 24 girls from divorced, mother- custody families; 24 boys and 24 girls from nondivorced families

Socioeconomic Dependent Major Background Measures Findings

Middle class EI-ISB, LLCS, LWRISB

SES (defined separately by family income and by educational and occupational classifica- tions) was examined and was instituted as a control in the analyses, but no breakdown by SES was given.

Lower-middle and middle class

All children had 2 working parents, each of whom had at least 2 years of college education.

Middle class

HESB, PAR, locus of control, optimism- pessimism, VTQ, Achenbach Parent and Teacher Rating Scales

BSRI, RRLS, RSRT, RIAS, BSF, RLCS, BMFS

Stress, as measured by parents' ratings on a behavior checklist

Observational measures of children's free play and social interaction; teacher ratings of be- havior; peer nomina- tion measures

Individuals (especially males) whose parents had divo'rced had higher ego-identity achievement scores than individuals in intact families.

At Time 2 (mean of 6.4 years post-divorce) and con- trolling for SES, boys in divorced families had a greater frequency of mal- adaptive symptoms, in- appropriate behavior, and unhappiness. Among girls, the only significant difference was that girls in divorced families had higher internal locus of control.

Women whose fathers were absent were virtually indis- tinguishable from those with fathers present. Early father absence was associated with external locus of control.

Children in divorced families experienced greater stress, as did those who had poor re- lationships with their parents.

Play patterns of boys and girls from divorced families were less socially and cog- nitively mature immediately following divorce; for boys immature play continued into second year post- divorce. Children in di- vorced families were also less happy and more anxious, with these effects again lasting longer for boys.

(Continued on next page)

This content downloaded from 132.174.254.81 on Thu, 25 Apr 2019 06:19:15 UTC All use subject to https://about.jstor.org/terms

TABLE 2. SUMMARY OF RECENT STUDIES ON EFFECTS OF DIVORCE ON PERSONAL ADJUSTMENT OF CHILDREN AND ADOLESCENTS-Continued

Description Family Structure Socioeconomic Dependent Major Study of Sample of Respondents Background Measures Findings Kalter et al. (1984)

Kalter et al. (1985)

Kinard and Reinherz (1984)

Kurdek and Siesky (1980a)

Oshman and Manosevitz (1976)

Parish et al. (1980)

Rosen (1979)

48 white boys and girls in 3rd and 5th grade

84 female college students aged 17-22

202 boys and 192 girls in 3rd grade; 99% white

48 boys and 64 girls in grades 3-8

125 male college students (mean age = 19.8)

158 female and 68 male college students

117 white males and females aged 9-28

Evenly divided between intact and divorced families

42 women with divorced parents; 42 with intact home

38 children in early (preschool) disrupted families; 36 in recently disrupted families; 320 in never-disrupted families

56 children in intact families; 56 children in divorced families

39 with biological fathers present; 47 fatherless; 39 with step- fathers

Middle and upper- middle class

Middle- and upper- middle class students at a "highly selective, expensive college"

Nowicki-Strickland Locus of Control Scale

TAT measuring percep- tions of masculinity- femininity; question- naire measuring life satisfaction

Predominantly working CAAP, PBQ, class (Classes III and SBCL IV on Hollingshead Index)

Middle class, as mea- sured by Hollingshead's Index

3 groups were matched on parental education and family income (presumably middle class).

Intact and fatherless No information given families

92 individuals whose Middle class parents had divorced and 25 whose families were intact

Modified version of BSRI

EIS

Rokesch Values Survey

TAT, ISB, and an in- depth clinical interview

Children in divorced families had higher internal locus of control.

No significant differences in life satisfaction or in satis- faction with dating. How- ever, women in divorced families were more likely to hold negative views of both men and women.

Children in recently dis- rupted families had signifi- cantly more problems with attention at school, with- drawal, and dependency. No differences on anxiety.

Children in divorced families were significantly more androgynous than children in intact families, who tended to be traditionally sex-typed.

Males with fathers (and stepfathers) present scored significantly higher on psychosocial development than males in fatherless families.

Individuals from intact and fatherless families were similar on most values, but individuals in divorced families were more likely to value self-respect, mature love, and being polite. No differences between in- tact and divorced group. Poor adjustment was associated with parental conflict preceding and during the divorce.

(Continued on next page)

C-4

o Q

A-

tlt

1z'

This content downloaded from 132.174.254.81 on Thu, 25 Apr 2019 06:19:15 UTC All use subject to https://about.jstor.org/terms

;)

TABLE 2. SUMMARY OF RECENT STUDIES ON EFFECTS OF DIVORCE ON PERSONAL ADJUSTMENT OF CHILDREN AND ADOLESCENTS-Continued

Description Family Structure Socioeconomic Dependent Major Study of Sample of Respondents Background Measures Findings

Santrock 120 boys in 5th and 60 boys in intact Predominantly lower A variety of behavioral Controlling for relevant var- (1975) 6th grade families; 20 in early- class. Father-present measures, including iables, there were no differ-

divorced (boys younger and father-absent Kohlberg measures of ences on behavioral than age 6) families; 20 groups were matched moral judgment, plus measures of moral develop- in late-divorced (boys on SES, race, school, teacher ratings of ment, but teachers judged between 6-10) families; and other background social conscience (e.g., father-absent boys to have 20 in father-deceased characteristics. self-discipline, trust) lower levels of social (boys between 6-10) conscience. families

Santrock and 33 boys and 27 girls 20 children in father- Groups were matched A variety of projective, Overall, there was no main Warshak (1979) aged 6-11 custody families; 20 in on SES, family size, self-report, and obser- effect for family structure,

mother-custody and sibling status. vational measures, in- but children living with families; 20 in intact cluding the ABIC opposite-sex parents were families not as well adjusted.

Wyman et al. 268 children in 98 children with Divorce sample "com- STAIC A-Trait Children in divorced group (1985) grades 4-6 divorced parents; 170 parable to the intact Scale used to measure had significantly higher

from intact families sample in sociodemo- trait anxiety anxiety. graphic background"

Young and 98 female college 16 daughters of un- Middle class MFF, ACL, SII Daughters who had lost Parish (1977) students aged 17-22 married widows; 12 of their fathers (due to divorce

remarried widows; 17 or death) and whose of unmarried divorcees; mothers had not remarried 26 of remarried were more insecure than divorcees; 27 with daughters in father-present father present families. No differences on

impulsivity-reflexivity.

Note: ABIC = Adaptive Behavior Inventory for Children; ACL = Adjective Check List; BMFS = Broverman Masculinity-Femininity Scale; BSF = Bendig Short Form of the Manifest Anxiety Scale; BSRI = Bem Sex Role Inventory; CAAP = Child and Adolescent Adjustment Profile; EI-ISB = Marcia Ego-Identity Incomplete Sentence Blank; EIS = Ego Identity Scale; HESB = Hahnemann Elementary School Behavior Rating Scales; ISB = Incomplete Sentence Blank; LLCS = Levenson's Locus of Control Scale; LWRISB = Loevinger, Wessler, and Redmore Incomplete Sentence Blank; MFF = Matching Familiar Figures Test; PAR = Sells and Roff Peer Acceptance-Rejection Rating; PBQ = Preschool Behavior Questionnaire; RIAS = Rubin's Intercourse Acceptability Scale; RLCS = Rotter Locus of Control Scale; RRLS = Rubin's Romantic Love Scale; RSRT = Rubin's Sex Role Traditionalism Scale; SBCL = Simmons Behavior Checklist; SII = Security-Insecurity Inventory; STAIC A-Trait Scale = State Trait Anxiety Inventory for Children; TAT = Thematic Apperception Test; VTQ = Vineland Teacher Questionnaire.

This content downloaded from 132.174.254.81 on Thu, 25 Apr 2019 06:19:15 UTC All use subject to https://about.jstor.org/terms

Journal of Marriage and the Family

families. Girls differed only on the dimension of locus of control; girls in divorced households scored significantly higher than their counterparts in intact households.

One explanation for boys' greater difficulties in adjusting to parental divorce is that typical postdivorce living arrangements are quite dif- ferent for them than for girls. While custodial mothers provide girls with same-sex role models, most boys have to adjust to living without same- sex parents. In examining boys and girls living in intact families and in different custodial ar-

rangements, Santrock and Warshak (1979) found that few effects could be attributed to family structure per se, but that children living with opposite-six parents (mother-custody boys and father-custody girls) were not as well adjusted on measures of competent social behavior. While father custody is rare, this study illustrates the im- portance of examining variations in postdivorce family structures (and specifically the combina- tion of parent's gender and child's gender) for estimating the effects of divorce on children.

Along related lines, a number of researchers have examined gender-role orientation and, specifically, the relation of father absence to boys' personality development. Most of the evidence in- dicates that boys without adult male role models demonstrate more feminine behavior (Biller, 1976; Herzog and Sudia, 1973; Lamb, 1977a), ex- cept in lower-class families (Biller, 1981b). A variety of studies have shown that fathers in- fluence children's gender role development to be more traditional because, compared to mothers, they more routinely differentiate between masculine and feminine behaviors and encourage greater conformity to conventional gender roles (Biller, 1981a; Biller and Davids, 1973; Bronfen- brenner, 1961; Heilbrun, 1965; Lamb, 1977b; Noller, 1978). Lamb (1977a) argues that because gender identity is usually developed by age 3 and because family influences are central to this proc- ess, the effects of father absence on gender- appropriate behavior may be most pronounced among boys who are very young (ages 5 and under) at the time of family disruption. Beyond early childhood, gender roles are largely estab- lished and children experience increasingly diverse extrafamilial social contexts and relationships that bear on their development.4 But it should be reiterated that these effects have been attributed

to father absence and thus would be expected to

occur among boys in all female-headed families, not simply those that have experienced divorce.

The claim has also been made that boys' ad- justment problems are often compounded by custodial mothers' denigrating the masculinity of absent fathers, an occurrence that is particularly likely in black matriarchal families (Biller and Davids, 1973). The assumption here is that boys are trying to be masculine without the benefit of the same-sex role model and that the absent role

model is portrayed as undesirable. However, most of the research on boys' adjustment fails to con- sider the quality or quantity of father-child con- tact or the availability of alternative male role models (e.g., foster father, grandfather, big brother, other male relatives, coach, friend, etc.), which makes it difficult to assess the impact of changing family structure on boys' behavior. There are also limitations imposed by concep- tualizing and measuring masculinity-femininity as a bipolar construct (Bem, 1974; Constantinople, 1973; Worell, 1978), and there is evidence that boys and girls in father-absent families are better described as androgynous (Kurdek and Siesky, 1980a).

Positive outcomes of divorce. While much of the literature on divorce and children seems

ideologically driven and biased toward emphasiz- ing negative effects on children (Edwards, 1987; Raschke and Raschke, 1979), the tendency of children in single-parent families to display more androgynous behavior may be interpreted as a beneficial effect. Because of father absence, children in female-headed families are not

pressured as strongly as their counterparts in two- parent families to conform to traditional gender roles. These children frequently assume a variety of domestic responsibilities to compensate for the absent parent (Weiss, 1979), thereby broadening their skills and competencies and their definitions of gender-appropriate behavior. Divorced parents also must broaden their behavioral patterns to meet increased parenting responsibilities, thereby providing more androgynous role models. Kurdek and Siesky (1980a: 250) give the illustration that custodial mothers often "find themselves needing to acquire and demonstrate a greater degree of dominance, assertiveness, and independence while custodial fathers may find themselves in situations eliciting high degrees of warmth, nur- turance, and tenderness."

626

This content downloaded from 132.174.254.81 on Thu, 25 Apr 2019 06:19:15 UTC All use subject to https://about.jstor.org/terms

Impact of Divorce on Children

Aside from becoming more androgynous, adolescents living in single-parent families are characterized by greater maturity, feelings of ef- ficacy, and an internal locus of control (Guidubaldi and Perry, 1985; Kalter, Alpern, Spence, and Plunkett, 1984; Wallerstein and Kel- ly, 1974; Weiss, 1979). For adolescent girls this maturity stems partly from the status and respon- sibilities they acquire in peer and confidant rela- tionships with custodial mothers.5

Finally, the relationship between family struc- ture and personal adjustment (and other dimen- sions of well-being) must be viewed as reciprocal. The child's psychological state prior to changes in family structure is an important element in the child's ability to adjust to new situations and rela- tionships. There is evidence (Kurdek et al., 1981) that children and adolescents with an internal

locus of control and a high level of interpersonal reasoning adjust more easily to their parents' divorce and that children's divorce adjustment is related to their more global personal adjustment.

Self-Concept

In Table 3 we summarize studies examining the impact of divorce on children's self-concept. A series of studies by Parish and his collaborators indicates that children in divorced, non-remarried families have lower self-esteem than children in

intact families (Parish and Dostal, 1980; Parish and Taylor, 1979; Young and Parish, 1977). Measuring children's self-evaluations in 1979 and again in 1982, Parish and Wigle (1985) demonstrated that children whose family struc- ture was intact throughout the study had the highest self-evaluations, while those whose parents divorced in the intervening years ex- perienced declining self-evaluations, and those whose parents were divorced throughout the 3-year period apparently adjusted to their new situations and reported higher self-evaluations than they had previously. As is the case for most research on children of divorce, however, the studies conducted by Parish and his associates did not investigate pre- or postdivorce levels of family conflict.

Marital discord. The bulk of evidence summar-

ized in Table 3 is consistent with the findings on personal adjustment; that is, family structure is unrelated to children's self-esteem (Feldman and

Feldman, 1975; Kinard and Reinherz, 1984; Parish, 1981; Parish, Dostal, and Parish, 1981), but parental discord is negatively related (Amato, 1986; Berg and Kelly, 1979; Cooper, Holman, and Braithwaite, 1983; Long, 1986; Raschke and Raschke, 1979; Slater and Haber, 1984). Because this conclusion is based on diverse samples of boys and girls of different ages in different living arrangements, the failure to obtain effects of family structure suggests either that family com- position really does not matter for children's self- concept or that family structure alone is an insuf- ficient index of familial relations. Further, these studies suggest that divorce per se does not adversely affect children's self-concept. Cashion's (1984) review of the literature indicates that children living in single-parent families suffer no losses to self-esteem, except in situations where the child's family situation is stigmatized (Rosenberg, 1979). Cautioning that considerably more research is needed before firm conclusions

can be drawn, Long (1986: 26) suggests that future work investigate "Hetherington's (1979) idea that a stable home in which parents are divorced is better for a child than is a 'conflict-

ridden' home where both parents are present." Although countless studies have examined

global self-esteem, two critical limitations characterize this body of research: (a) nearly all of these studies are cross.sectional, which restricts the assessment of developmental change and stability in self-concept; and (b) little is known about the various dimensions of self-concept (e.g., self-efficacy, nonevaluative self- descriptions) other than overall self-esteem. It is necessary, therefore, to examine different dimen- sions of self-concept as they change over time and as they relate to different structures and patterns of family interaction.

Cognitive Functioning

Most of the research relating cognitive function- ing to family structure (summarized in Table 4) has assessed cognitive performance by using stan- dardized intelligence and academic achievement tests or scholastic grade-point averages. Many of these studies find that family conflict and disrup- tion are associated with inhibited cognitive func- tioning (Blanchard and Biller, 1971; Feldman and Feldman, 1975; Hess and Camara, 1979; Kinard and Reinherz, 1986; Kurdek, 1981; Radin, 1981).

627

This content downloaded from 132.174.254.81 on Thu, 25 Apr 2019 06:19:15 UTC All use subject to https://about.jstor.org/terms

TABLE 3. SUMMARY OF RECENT STUDIES ON EFFECTS OF DIVORCE ON SELF-CONCEPT OF CHILDREN AND ADOLESCENTS

Description Family Structure Socioeconomic Dependent Major Study of Sample of Respondents Background Measures Findings Berg and Kelly 57 boys and girls (1979) aged 9-15

258 boys and 209 girls aged 9-12

(1) 40 girls in 3rd and 6th grade

(2) 522 girls aged 11-18 (1972 National Survey of Youth)

(3) 84 female college students aged 17-22

153 white children aged 3-21

202 boys and 192 girls in 3rd grade; 99% white

19 children in "intact- accepted" families; 19 in "intact-rejected" families; 19 whose parents had divorced

Two-parent cohesive families; one-parent co- hesive families; isolated child families; divided families; parent- coalition families

14 girls with divorced parents, 26 living with both natural parents

62 girls with divorced parents, 460 girls from intact households

42 women with divorced parents, 42 with intact home

45 mothers and their 74 children from divorced families, and 44 mothers and their 79 children from in- tact families

38 children in early (preschool) disrupted families; 36 in recently disrupted families; 320 in never-disrupted families

No information given PHCSC Children in intact but conflict-ridden families had

40% professional or managerial; 33% clerical, sales, and skilled occupations; 27% unskilled and semiskilled occupa- tions.

All middle class

CSEI; PHCSC

PCS

Divorced group "repre- Self-esteem sented significantly lower SES households"

Middle- and upper- middle class students at a "highly selective, expensive college" Predivorce annual fam- ily income of $12,000- $40,000

Self-esteem

Self-concept Referents Test (children 3-7); Bills Index of Adjust- ment and Values (chil- dren 10-21)

Predominantly working PHCSC class (Classes III and IV on Hollingshead Index)

significantly lower self- esteem than those in other 2 groups.

Children reporting little family support, whether in one- or two-parent families, had lowest self-esteem.

Among 3rd graders, girls in divorced families had signifi- cantly lower perceived competence. No significant differences among 6th graders.

No significant differences.

In divorced families, mothers' present adjustment predicted children's self- concept. In intact families, mothers' past adjustment predicted children's self- concept.

No significant differences.

Q-

(Continued on next page)

Cooper et al. (1983)

Kalter et al. (1985)a

Kanoy et al. (1984)

Kinard and Reinherz (1984)

This content downloaded from 132.174.254.81 on Thu, 25 Apr 2019 06:19:15 UTC All use subject to https://about.jstor.org/terms

TABLE 3. SUMMARY OF RECENT STUDIES ON EFFECTS OF DIVORCE ON SELF-CONCEPT OF CHILDREN AND ADOLESCENTS-Continued

Description Family Structure Socioeconomic Dependent Major Study of Sample of Respondents Background Measures Findings

Long (1986)

Parish (1981)

Parish et al. (1981)

Parish and Dostal (1980)

Parish and Taylor (1979)

199 predominantly white female college freshmen (mean age = 17.7)

1,409 male and female college students

284 male and female children in grades 5-8

738 boys and girls aged 11-14

406 boys and girls in grades 3-8

150 with intact fam- ilies; 21 in reconstituted families; 26 in single- parent families; 2 in other arrangements

Intact; divorced non- remarried; divorced re- married; death non- remarried; death remarried families

Two family types (in- tact or divorced) and two family dispositions (happy or unhappy)

Intact; divorced and remarried; divorced and non-remarried

347 from intact fami- lies; 44 from divorced, female-headed families; 15 from reconstituted families

Predominantly middle class

No information given

RSE Daughter's self-esteem sig-

PAIC

nificantly related to parental happiness (even with family structure controlled) but not to family structure (even with parental happiness con- trolled).

No significant differences by type of family structure.

No main effects for family type, but children from happy families had signifi- cantly higher self-concepts. Females from happy, intact families had significantly higher self-concepts and males from unhappy, divorced families had significantly lower self- concepts.

Children from intact fami- lies had significantly higher self-evaluations than children in divorced, non- remarried families.

PAIC

PAIC Children whose parents had divorced and whose mothers had not remarried had significantly lower self- concepts than children in intact families.

(Continued on next page)

0 lt>

1*1

(b i

a o

s C) :Z.

;:3

This content downloaded from 132.174.254.81 on Thu, 25 Apr 2019 06:19:15 UTC All use subject to https://about.jstor.org/terms

TABLE 3. SUMMARY OF RECENT STUDIES ON EFFECTS OF DIVORCE ON SELF-CONCEPT OF CHILDREN AND ADOLESCENTS-Continued

Description Family Structure Socioeconomic Dependent Major Study of Sample of Respondents Background Measures Findings

Parish and Wigle 120 public school 4 groups of 30 children No information given PAIC Children in intact-intact (1985) students each, defined by their group had highest self-

family structure at 2 points in time (1979 and 1982): intact-intact; divorced-divorced; intact-divorced; and a control group

o O

evaluations at both measure- ment points (T1 and T2). Children whose parents divorced between T1 and T2 experienced a significant decline in self-evaluations, while self-evaluations of children in divorce-divorce group significantly increased by T2.

289 black and white boys and girls in grades 3, 6, and 8

268 children in grades 4-6

70% living with both parents (including re- constituted); 26% with mother only; 4% with neither parent

98 children with di- vorced parents, 170 from intact families

Unknown, but assumed PHCSC to be mixed, and skewed toward lower SES

Divorce sample "com- PCS parable to the intact sample in sociodemo- graphic background"

No significant differences by family structure, but signifi- cantly lower self- concepts among children reporting high family conflict.

No significant differences in perceived social competence, physical competence, or general self-esteem. Divorced group had significantly lower perceived cognitive competence.

Young and Parish (1977)

98 female college students aged 17-22

16 daughters of un- married widows; 12 of remarried widows; 17 of unmarried divorcees; 26 of remarried di- vorcees; 27 with father present

Middle class ACL Daughters who had lost their fathers (due to divorce or death) and whose mothers had not remarried had significantly more negative self-evaluations.

Note: ACL = Adjective Check List; CSEI = Coopersmith Self-Esteem Inventory; PAIC = Personal Attribute Inventory for Children; PCS = Perceived Competence Scale for Children; PHCSC = Piers-Harris Children's Self-Concept Scale; RSE = Rosenberg Self-Esteem Scale.

aThree separate studies, using three different samples, are presented in Kalter et al. (1985).

Raschke and Raschke (1979)

Wyman et al. (1985)

hr tl

This content downloaded from 132.174.254.81 on Thu, 25 Apr 2019 06:19:15 UTC All use subject to https://about.jstor.org/terms

Impact of Divorce on Children

Two important reviews of research on children in fatherless families produce different conclusions: Herzog and Sudia (1973) conclude that children's school achievement is not affected by father absence, but Shinn (1978) concludes that father absence has a number of detrimental effects on

children's intellectual performance. Basing her conclusions on 30 studies that met reasonable

methodological criteria, Shinn reports that "financial hardship, high levels of anxiety, and in particular, low levels of parent-child interaction are important causes of poor performance among children in single-parent families" (1978: 316). In this section we summarize the differential effects

of family disruption on academic performance by gender and social class and offer some insights as to the mechanisms by which these effects occur.

Children's gender. Some studies suggest that negative effects of family disruption on academic performance are stronger for boys than for girls (Chapman, 1977; Werner and Smith, 1982), but most of the evidence suggests similar effects by gender (Hess and Camara, 1979; Kinard and Reinherz, 1986; Shinn, 1978). While females traditionally outscore males on standardized tests of verbal skills and males outperform females on mathematical skills, males who have experienced family disruption generally score higher on verbal aptitude (Radin, 1981). Thus, the absence of a father may result in a "feminine" orientation toward education (Fowler and Richards, 1978; Herzog and Sudia, 1973). But an important and unresolved question is whether this pattern results from boys acquiring greater verbal skills in mother-headed families or from deficiencies in mathematical skills attributable to father absence.

The latter explanation is supported by evidence showing that father-absent girls are disadvantaged in mathematics (Radin, 1981).

Children's race. There is a limited amount of evidence that father absence is more harmful to the intelligence and academic achievement of black children (Sciara, 1975), especially black males (Biller and Davids, 1973), but most studies show academic achievement among black children to be unaffected by family structure (Hunt and Hunt, 1975, 1977; Shinn, 1978; Solomon, Hirsch, Scheinfeld, and Jackson, 1972). Svanum, Bringle, and McLaughlin (1982) found, controlling for social class, that there are no significant effects of

father absence on cognitive performance for white or black children. Again, these investigations focus on family composition and demonstrate that the effects of family structure on academic performance do not vary as much by race as by social class, but race differences in the impact of divorce remain largely unexplored. As Table 4 il- lustrates, we did not find any studies that com- pared white and black populations of children in divorced and nondivorced families.

Family socioeconomic status. A review by Hether- ington, Camara, and Featherman (1983) underscores the importance of social class as a mediating variable. They note small differences favoring children in two-parent families on stan- dardized tests of intelligence and academic achievement that decrease when socioeconomic

circumstances are controlled. Differences remain, however, on measures of school performance (e.g., grade-point average), with children in one- parent families at a disadvantage. In a study of predominantly white working-class children, Kinard and Reinherz (1986) investigated the im- pact of marital disruption on specific dimensions of school performance. Fourth-graders whose families were recently disrupted (i.e., children whose parents divorced since the children entered school) had lower scores on language aptitude and a composite measure of academic achievement than children in never-disrupted families or families in which disruption had occurred several years earlier. But no group differences were detected in mathematics achievement. When

maternal education was controlled, there were no differences in reading achievement. In fact, maternal education had a stronger effect on school performance than did marital disruption. Differences in teacher assessments of productivity disappeared when gender and maternal education were controlled (Kinard and Reinherz, 1984).

These findings direct attention to a major methodological problem indicated in earlier reviews (Herzog and Sudia, 1973; Shinn, 1978), namely, inadequate attention to the role of social class in moderating the effects of family disrup- tion on children's academic performance. When social class is controlled, children in female- headed families fare no worse than children from

two-parent families on measures of intelligence (Bachman, 1970; Kopf, 1970), academic achieve- ment (Shinn, 1978; Svanum et al., 1982), and

631

This content downloaded from 132.174.254.81 on Thu, 25 Apr 2019 06:19:15 UTC All use subject to https://about.jstor.org/terms

TABLE 4. SUMMARY OF RECENT STUDIES ON EFFECTS OF DIVORCE ON COGNITIVE FUNCTIONING OF CHILDREN AND ADOLESCENTS

Description Family Structure Socioeconomic Dependent Major Study of Sample of Respondents Background Measures Findings Chapman (1977)

Hess and Camara (1979)

Kalter et al. (1985)a

Kinard and Reinherz (1984)

Kinard and Reinherz (1986)

Svanum et al. (1982)

96 white college students aged 17-23 (mean = 18.9)

32 white boys and girls aged 9-11

(1) 40 girls in 3rd and 6th grade

(2) 84 female college students aged 17-22

202 boys and 192 girls in 3rd grade; 99% white

78 girls and 72 boys in 4th grade; 99% white

Representative national sample of 6,109 male and female children aged 6-11 (National Health Examination Survey, Cycle II)

16 males and 16 females in each of 3 groups: father-absent, stepfather present, and intact

16 children in intact families; 16 children in recently divorced families

14 girls with divorced parents, 26 living with both natural parents 42 women with divorced parents, 42 with intact home

38 children in early (preschool) disrupted families; 36 in recently disrupted families; 320 in never-disrupted families

87 children in never- disrupted families; 33 in early (preschool) dis- rupted families; 30 in recently disrupted families

5,493 children in in- tact families; 616 chil- dren in father-absent families

No information given

All children had 2 working parents, each of whom had at least 2 years of college education

All middle class

Field dependence measured by Form A of EFT; SAT

Work effectiveness at school as measured by school records and teacher ratings

SAT

Middle- and upper-class GPA students at a "highly selective, expensive college" Predominantly working CAAP class (Classes III and IV on Hollingshead Index)

Working class (ma- jority in Classes III and IV on the Hollingshead Index)

Representative sample; SES (defined in terms of family income and householder education) used as control variable

PSS, CAAP, SFTAA, CAT-70, parent and teacher ratings of academic achievement

Vocabulary and block design subtests of WISC; arithmetic and reading subtests of WRAT

Intact family males were more field independent and had significantly higher SAT verbal and total scores than males in other 2 groups. Intact family females had lower SAT verbal scores than the other 2 groups. Childen in divorced families were less productive in schoolwork. Present child relationships were also significantly related to children's schoolwork.

No significant differences.

Controlling for gender and maternal education, there were no differences in students' productivity by family type.

Children in recently dis- rupted families had signifi- cantly lower scores on lan- guage, total achievement, and productivity. No group differences in mathematics achievement or in reading achievement with maternal education controlled.

Controlling for SES, no sig- nificant effects of father absence for white or black children.

Note: CAAP = Child and Adolescent Adjustment Profile; CAT-70 = California Achievement Test-70; EFT = Embedded Figures Test; GPA = Grade Point Average; PSS = Preschool Screening System; SAT = Stanford Achievement Test; SFTAA = Short Form Test of Academic Aptitude; WISC = Wechsler Intelligence Scale for Children; WRAT = Wechsler Reading Achievement Test.

aThree separate studies, using three different samples, are presented in Kalter et al. (1985). All three studies are described in Table 3.

i11

I~ ?z I A

This content downloaded from 132.174.254.81 on Thu, 25 Apr 2019 06:19:15 UTC All use subject to https://about.jstor.org/terms

Impact of Divorce on Children

educational attainment (Bachman, O'Malley, and Johnston, 1978).6 Further, Svanum et al. (1982: 143) point out that there are many considerations in deciding whether to control for SES in ex- aminations of cognitive performance. While much of the observed variance in cognitive performance may be attributable to SES, "the role that SES assumes in the underlying processes of father absence and cognitive development is unclear at this stage of research." In order to disentangle the intricate effects of family structure and SES on children's cognitive performance, family researchers need to examine the socioeconomic history of intact families and those in which disruption occurs, to examine the economic resources available to children at

various stages of cognitive development, and to assess changes in economic resources and family relationships that accompany marital disruption.

Family processes. In recent years important in- sights have been gained into the specific processes by which marital disruption may affect children's school performance. First, family disruption alters daily routines and work schedules and im- poses additional demands on adults and children living in single-parent families (Amato, 1987; Furstenberg and Nord, 1985; Hetherington et al., 1983; Weiss, 1979). Most adolescents must assume extra domestic and child care respon- sibilities, and financial conditions require some to work part-time. These burdens result in greater absenteeism, tardiness, and truancy among children in single-parent households (Hether- ington et al., 1983). Second, children in recently disrupted families are prone to experience emo- tional and behavioral problems such as aggres- sion, distractibility, dependency, anxiety, and withdrawal (Hess and Camara, 1979; Kinard and Reinherz, 1984), factors that may help to explain problems in school conduct and the propensity of teachers to label and stereotype children from broken families (Hess and Camara, 1979; Hether- ington et al., 1979, 1983). Third, emotional prob- lems may interfere with study patterns, while demanding schedules reduce the time available for single parents to help with homework. In support of the latter point, Furstenberg and Nord (1985) examined parent-child interaction patterns in dif- ferent family types and found few differences in time spent together in social and recreational ac-

tivities, but found that resident parents in reconstituted and single-parent families were much less likely than parents in intact families to help with homework. In sum, a variety of per- sonal, family, and school processes operate to the detriment of academic performance among children of divorce.

Interpersonal Relationships

Compared to the large bodies of research on per- sonal adjustment, self-concept, and cognitive functioning, relatively few studies have examined interpersonal relations among children and adolescents in different family structures (see Table 5). Generally, investigations have focused on peer relations among children and dating pat- terns among adolescents.

Peer relations. Studies of preschool children (Hetherington et al., 1979) and preadolescents (Santrock, 1975; Wyman, Cowen, Hightower, and Pedro-Carroll, 1985) suggest that children in disrupted families are less sociable: they have fewer close friends, spend less time with friends, and participate in fewer shared activities. Stolberg and Anker (1983) observe that children in families disrupted by divorce exhibit psychopathology in interpersonal relations, often behaving in unusual and inappropriate ways. Other studies suggest that the effects are temporary. Kinard and Reinherz (1984) found no differences in peer rela- tions among children in intact and disrupted families, but those in recently disrupted families displayed greater hostility. Kurdek et al. (1981) conducted a two-year follow-up of children whose parents had divorced and showed that relation- ships with peers improved after the divorce and that personal adjustment was facilitated by op- portunities to discuss experiences with peers, some of whom had similar experiences. However, Guidubaldi and Perry (1985) observed a much dif- ferent pattern: among boys, those from divorced families had greater contact with friends, and among girls there were no differences by family structure.

Dating patterns. Hetherington (1972) reported that adolescent girls whose fathers were absent prior to age 5 had difficulties in heterosexual rela- tions, but Hainline and Feig's (1978) analyses of

633

This content downloaded from 132.174.254.81 on Thu, 25 Apr 2019 06:19:15 UTC All use subject to https://about.jstor.org/terms

TABLE 5. SUMMARY OF RECENT STUDIES ON EFFECTS OF DIVORCE ON INTERPERSONAL RELATIONSHIPS OF CHILDREN AND ADOLESCENTS

Study

Booth et al. (1984)

Guidubaldi and Perry (1985)

Hess and Cam