a prospective study of the effects of marital status

A Prospective Study of the Effects of Marital Status and Family Relations on

Young Children's Adjustment Among African American and Caucasian Families

Introduction

Within the last two decades, the divorce rate in the United States has increased substantially. Since 1958, when there were 2.1 divorces per 1,000 couples, a gradual increase in the number of divorces has occurred, peaking at 5.3 per 1,000 couples in 1979 and 1981 (Glick & Lin, 1986), and stabilizing at 4.7 as of 1990 (U.S. Department of Health & Human Services, 1990). According to projections based on 1990 census data, 40% of all children can expect to live in a single-parent household because of divorce before the age of 16 (Cherlin, 1992). Based on the increase in divorces over the last three decades and the prospect that a similarly high number will occur in the future, it remains imperative to understand the effects of divorce on children's adjustment.

As research on divorce and children's adaptation has accumulated, there has been a gradual shift in emphasis from family structure to family process (Emery, Hetherington, & DiLalla, 1984; Emery, 1988). That is, events that accompany marital dissolution, rather than the event of divorce per se have been identified as potentially more salient correlates of children's adjustment (Amato & Rezac, 1994; Camara & Resnick, 1988; Emery, 1988; Fauber, Forehand, Thomas, & Wierson, 1990; Haurin, 1992). Longitudinal investigations of divorced families (Hetherington, Cox, & Cox, 1979, 1985; Wallerstein, 1991; Zill, Morrison, & Coiro, 1993) have provided particularly strong support for this focus on family process. However, in longitudinal studies in which data collection was begun following the parental separation, changes in family process may erroneously be attributed to the parental separation and are subject to the biasing errors of retrospective reporting. In such studies, parents may blame the child's problematic behavior on the divorce rather than circumstances that existed prior to the parental separation.

Prospective longitudinal studies begun prior to the parental separation can therefore be useful in addressing two issues. First, they allow researchers to examine if child adjustment problems were evident prior to parental separation. Second, if problematic child behavior began prior to the separation, factors that might be responsible for its development can be explored. Two family process factors have consistently been postulated to influence child behavior in both divorced and two-partner families: parental conflict and parenting.

Interparental conflict

Although for many years clinicians have postulated an association between parental conflict and maladjustment in children (Baruch & Wilcox, 1944; Minuchin; 1974), empirical attention to the effects of parental discord on children has increased only in the last two decades. From these recent controlled studies and from earlier reports of "broken" families, interparental conflict has been consistently identified as a major correlate of behavior problems in children across a wide array of family structures and settings (for reviews see Davies & Cummings, 1994; Erel & Burman, 1995; Grych & Fincham, 1990), including divorced and separated families (Hetherington et al., 1979). There is some evidence to suggest that parental conflict is the most salient influence on children's adjustment to divorce. In a recent meta-analysis, Amato and Keith (1991) compared the relative efficacy of three process variables (parental absence, economic disadvantage, and parental conflict) to mediate the effects of divorce on children's adjustment. Although moderate effect sizes were found for both parental absence and economic disadvantage, parental conflict accounted for more of the negative consequences of divorce.

Parenting

Childrearing practices also have been found to affect children's adaptation across family structure and developmental period (Baumrind, 1973; Gardner, 1989; Patterson, 1982; see Rothbaum & Weisz, 1994 for review). Post-divorce disruptions in childrearing activities are so frequent and dramatic initially that Wallerstein and Kelly (1980) termed post-divorce a time of "diminished parenting." Strained affectional relationships and difficulty providing a consistent family authority structure are two of the most commonly cited issues (Emery, 1988; Haurin, 1992). However, without prospective data begun prior to the separation, it is unclear if pre-divorce childrearing styles may account for post-divorce child adjustment. One aspect of disrupted parenting commonly seen in divorced families, parental hostility and rejection, has been associated repeatedly with child externalizing problems among school-age children and adolescents (see Loeber & Dishion, 1983; McCord, McCord, & Zola, 1959). Although less is known about its effects among very young children, in a sub-sample of children from the present data set, we found that rejecting parenting at age 2 predicted externalizing behavior problems at age 3.5 (Shaw et al., 1998).

Unfortunately, a dearth of studies has been designed with the explicit intent of providing a prospective view of children and divorce. However, it is possible to use studies of normal child development as prospective investigations of the consequences of divorce for children. A few investigative teams have re-analyzed their longitudinal data sets to examine children's antecedent patterns of adjustment to divorce, including the effects of pre-divorce child adjustment and family process factors.

Pre-divorce Behavioral Adjustment in Children from Divorced Families

Block, Block, and Gjerde (1986) re-examined data from their longitudinal study of normal development. They found that, as many as 11 years prior to their parents' divorce, children from eventually-separating families showed more behavior problems than children from always-married families. Analyses for sex differences revealed earlier and more consistent differences for boys than girls. Boys whose parents would later separate demonstrated a pattern of problematic behaviors at ages 3, 4 and 7, including a lack of impulse control, stubbornness, and restlessness. Girls from eventually-separating families only had increased behavior problems at age 4 when compared to their peers who would remain in two-parent homes.

Two other prospective studies with large nationally representative samples also found the effects of divorce to be sharply reduced once earlier behavior problems were taken into account statistically (Cherlin et al., 1991). The British National Child Development Study included all children born in England, Scotland, and Wales during a week in 1958 (initial number of mothers interviewed was 17,414), with assessments for the divorce analysis conducted when children were ages 7 and 11. The United States National Survey of Children was a random-sample survey of 2,279 children ages 7 to 11 with follow-ups 5 years later when children were 11 to 16 years old. In both studies, reductions in post-divorce differences were stronger for boys than girls, and pre-divorce conditions still accounted for significant variance in girls' adjustment following divorce. For boys, the proportion of variance in behavior problems explained by divorce was reduced approximately 50% to a level in which differences were no longer statistically significant.

In contrast, a few recent studies have found no differences in behavioral adjustment among to-be-divorced and always-married families (Forehand, Armistead, & David, 1997; Shaw, Emery, & Tuer, 1993). For example, Forehand and colleagues (1997) found that adolescents from to-be-divorced families functioned similarly to those in always-married families, but better than those who had already divorced. These findings contradict the results of previous studies, suggesting that perhaps differences can be attributed to parental divorce and its concomitant disruption of family processes. However, it should be noted that in these two studies post-divorce differences were either modest or nonsignificant.

The Effects of Pre-divorce Family Process Issues: Parental Conflict and Parenting

Fewer investigators have examined the effects of pre-divorce parental conflict and childrearing practices on pre- or post-divorce child adjustment. However, Block, Block, and Gjerde (1988) found parents of to-be-divorced children to be characterized by high levels of interparental conflict and acrimonious relations between parents and sons. These problems were interpreted as being responsible for the behavior problems found among children prior to the divorce. Similar findings were reported by Shaw and colleagues (1993), using prospective data from the New York Longitudinal Study (Chess & Thomas, 1984; Thomas & Chess, 1977). Shaw and colleagues found that pre-divorce levels of rejecting parenting and interparental conflict distinguished to-be-divorced families from always-married families, and that effects of rejecting parenting were stronger for families with sons than families with daughters.

Family Structure, Ethnicity, and Child Adaptation

Many researchers have reported that relations between family structure and children's adjustment have been less consistent among African American than Caucasian families. Typically, investigators have found that Caucasian children from single-mother homes are at risk for behavior problems, most notably delinquent behavior (Offord, Sullivan, Allen, & Abrams, 1979); however, findings have been mixed for African Americans, with some investigators reporting significant positive associations between single-mother family structure and children's externalizing problems (Dornbusch et al., 1985), others reporting no significant relations (Zimmerman, Salem, & Maton, 1995), and some reporting significantly lower rates of conduct difficulties among single-mother African American families (Farnworth, 1984; Montare & Boone, 1980).

One possible explanation for the more consistent findings observed among Caucasian families concerns the extent to which a two-parent, biologic family structure typifies the family unit among Caucasians and African Americans. According to recent census data for family groups with children under 18, 77% of Caucasian families and 37% of African American families consisted of two-parent, biologic families (U.S. Bureau of the Census, 1992). An important consideration that has been neglected by most researchers concerns whether or not a single-parent family structure has resulted from marital dissolution. In two epidemiologic samples, Duncan, Brooks-Gunn, and Klebanov (1994) and McLeod and Shanahan (1993) have discovered that stable single-parent status was not associated with child externalizing behavior controlling for chronic poverty; however, divorce remained a significant predictor of externalizing behavior among elementary school children. One might speculate that observed ethnic differences in relations between family structure and children's antisocial behavior reflect the higher proportion of always-single mothers, as opposed to divorced single-mothers, among African Americans compared to Caucasians.

Due to a relatively large percentage of single parents in our sample (i.e., always-single and divorced), we were able to examine the effects of different types of single-parent status on children's adjustment, and compare this adaptation to those in different types of two-parent families (e.g., always-married, to-be-divorced). Adaptation of children from always-single and divorcing families can be examined both prior to and following the parental separation in the divorcing families. Single-parent status due to divorce is expected to have more deleterious effects on child adjustment both prior to and following the parental separation because of disturbances in family process that precede and follow divorce (e.g., parental conflict, disruptions in parenting). However, we also expected the effects of pre- and post-divorce marital status to be more pronounced among Caucasian rather than African American children due to the greater prevalence of two-parent families among Caucasian families. That is, single-parent status may be accepted as more normative among African American children regardless of whether or not a divorce has occurred (M. Wilson, personal communication, June 25, 1997).

In sum, our goals were to (1) investigate the pre- and post-divorce adjustment of a cohort of young, low-income boys in comparison to boys from always two-parent and always-single families; (2) examine whether or not ethnicity moderated relations between marital status and child behavior; and (3) test whether differences in child adjustment between marital groups could be accounted for by previous child adjustment, parental conflict, childrearing characteristics, and/or sociodemographic factors. We hypothesized that children in pre- and post-divorce environments would evidence higher rates of behavior problems than children from two-parent families, and that these effects would be more pronounced among Caucasian than African American families. Further, we postulated that to-be-divorced families would be characterized by higher levels of parental conflict, marital dissatisfaction, and rejecting parenting than always two-parent families, and that these characteristics of the pre-divorce environment would account for higher levels of behavior problems among children from divorced homes.

Method

Participants

The source for subject recruitment was Allegheny County's Women, Infants, and Children Program (WIC), which provides nutritional aid to low SES families in the Pittsburgh metropolitan area. Three-hundred ten subjects were recruited from WIC sites throughout the Pittsburgh metropolitan area over the course of two years. The sample represents a second larger cohort of families recruited from WIC, with no overlap of participants from a smaller cohort of boys and girls (e.g., Shaw, Keenan, & Vondra, 1994). Since the intent of the original investigation was to examine the developmental precursors of antisocial behavior, only boys were recruited, and all target children needed to have another sibling living in the home (i.e., lack of family space also has been shown to increase risk for externalizing problems). Subjects were recruited when target children were between 6 and 17 months of age. When infants were 18 months old, mothers ranged in age from 17 to 43 years, with a mean age of 28. Fifty-four per cent of participants were Caucasian, 40% were African American, and 6% were other (e.g., Hispanic, Asian, or biracial). Sampling of marital status was not restricted due to the considerable relationship instability within the sample. At the 18-month visit, 64% were either married or living together, 8% were divorced, 28% were always single, and 1% were other (e.g., widowed, single when the child was born but living together with partner, widowed). Mean per capita family income was $242 per month ($2,892 per year, and $11,616 for a family of four), and the mean Hollingshead socioeconomic status (SES) score was 24.5, which is indicative of a working class sample. Initially, 421 families were approached at WIC sites. Fourteen (3.3%) declined to participate at the time of recruitment and an additional 97 declined before the first assessment. Thus, of the 421 families asked, 310 (71%) participated in the first assessment at 18 months. Of the 310 families seen at the 18-month assessment, data were available on 302 at the 24-month assessment, 282 at the 42-month assessment, and 291 at the age 5 assessment, an attrition rate of 6.1%. It is our belief that attrition was lower at the age 5 assessment because these visits were carried out in family's homes as opposed to the laboratory.

Procedures

When children were 18 and 42 months, they were seen for a 2-hour visit at the laboratory, during which time mothers and sons participated in a number of parent-child interaction tasks and mothers completed a series of questionnaires. At 24 months, a similar combined home-lab visit was conducted, lasting approximately 31/2 hours. At 60 months, two separate home visits were conducted, both with the target child and his nearest-age sibling -- one assessment with the mother and one with the father. During these visits, observations of sibling and parent-child interaction occurred, at which time mothers, fathers, and alternative caregivers completed the Achenbach Child Behavior Checklist (CBCL). Mothers completed CBCLs at all assessment points, and alternative caregivers and fathers completed CBCLs at 60 months (available for 62% and 65% of cases, respectively). Alternative caregivers included family members who provided care for the target children on a regular basis, and included children's grandmothers (35%), aunts (19%), day care workers, babysitters, and teachers (25%), and family friends and other relatives (21%). Mothers were reimbursed for their participation at the end of each visit ($25 at all visits except $35 for the combined home-lab visit), as were fathers and alternative caregivers for completing questionnaires ($10).

Measures

Demographic Interview. At the beginning of each assessment, mothers were asked about SES variables, income, ethnicity, and marital status.

Child Behavior Checklist for Ages 2-3 and 4-16 (CBCL, Achenbach, 1991, 1992). The CBCL is a questionnaire designed to assess behavioral and emotional problems in children ages 2-3 and 4-16. When children were 2 and 3� years old, mothers completed the age 2-3 version, and at age 5, mothers, fathers, and alternative caregivers completed the age 4-16 CBCL. Achenbach recommends continuing to use the age 2-3 versus the age 4-16 version of the CBCL when children are between ages 3 and 4 and previous assessments have been conducted (Achenbach, 1992). The questionnaire generates two broad-band factors, Externalizing and Internalizing, both of which were used for analyses in this study. The mean test-retest reliability has been reported by the authors to be .87 (Achenbach, 1992).

Larzelere Toddler Behavior Checklist (TBC, Larzelere, Martin, & Amberson, 1989). Similar to the CBCL, but appropriate for infants as young as 9 months old and children as old as school-age, the TBC was administered to mothers when infants were 18 months. For the present study, the sum of two factors, oppositional and physical aggression, was used to form a composite factor of early disruptive behavior comparable to the broad-band CBCL Externalizing factor r = .72, p < .001 for correlation between oppositional and aggressive). Raw scores from the two factors were converted to z-scores before being summed. The oppositional factor consists of 22 items dealing with anger expression and noncompliance, and the physical aggression factor includes 14 items that tap willful physical aggression such as fighting and biting. The Chronbach alphas for these scales are .91 and .83, respectively.

Marital Adjustment Test. Maternal satisfaction with her marital or significant other relationship was assessed using the short form of the Marital Adjustment Test at the 18-month laboratory visit (Locke & Wallace, 1959). This measure has proven successful in discriminating harmonious and disturbed marriages (Locke & Wallace, 1959), and predicting children's behavior problems (Emery & O'Leary, 1982).

Child-Rearing Disagreements Scale (CRD). The CRD is a 21-item measure of common topics concerning child-rearing disagreements based on interviews with parents over the course of several years (Jouriles et al., 1991). The CRD was administered to mothers at the 24-month assessment. For the purposes of the present study, the frequency of childrearing disagreements was used as the outcome measure. In comparison with general marital satisfaction scales, the CRD has been found to be correlated with a greater variety of child behavior problems and account for significant variance in behavior problems after considering non-child disagreements and child exposure to marital conflict.

Conflict Tactics Scale (CTS). The CTS is a reliable instrument of family resolution tactics (Strauss, 1979). It was administered to mothers at the 42-month lab assessment. It includes factors for verbal and physical strategies used by couples to resolve disputes. In the present study, the Verbal Aggression scale was used instead of Physical Aggression because of the former's greater range of scores, its greater comparability of item content with the CRD, and established relations with children's behavior problems (Christopoulos et al., 1987).

Rejecting Parenting. Maternal rejecting parenting was measured using the Early Parenting Coding System (EPCS). Observer ratings of parenting were made from videotaped mother-child interactions during a structured clean-up task at the 18- and 24-month lab assessments. The EPCS consists of 9 categories of parenting strategies coded molecularly as well as six global ratings (Winslow, Shaw, Bruns, & Kiebler, 1995). For the purposes of the present study, only molecular and global ratings relevant to rejecting parenting were employed. These included two molecular ratings--verbal/physical approval and critical statement --as well as three global ratings--hostility, warmth, and punitiveness. Hostility was defined as the emotional expression of anger by the mother toward the child as indicated by tone of voice and mannerisms. The warmth rating was an evaluation of the amount of positive affect expressed toward the child. Punitiveness was defined as the extent to which the mother was too strict, demanding, or harsh, considering the child's behavior. For molecular codes, Cohen's kappa coefficients were .85 for approval and .75 for critical statement. For global ratings, weighted kappa coefficients were .84, .81, and .89 for hostility, warmth, and punitiveness, respectively. Coders were blind to scores on all other measures used in the study.

Principal components analysis was conducted with the five parenting variables at each age to create composites corresponding to rejecting parenting. At 24 component loadings exceeded .40 in the expected directions for all five variables. However, at 18 months molecular approval did not load on a component with the other four variables; therefore, approval was not included in the rejecting composite at 18 months. Alpha coefficients for the rejecting parenting scales were .68 at 18 months and .70 at 24 months.

Results

Overview of Group Formation

Family structure was identified prospectively based on marital status at the 18-, 24-, 42-, and 60-month assessments. Always two-parent families were those in which biologic parents were married or living together continuously from the 18- to 60-month assessments, with no known history of prior separations during the target child's life (n = 136 at 18 months). To-be-divorced families were those who would eventually experience a divorce or separation at the 24-, 42-, or 60-month assessment but who were married or living together at 18 months (n = 55); thus, membership for the to-be-divorced group decreased in size with time. At 18 months, families who had experienced a separation or divorce were placed in the already divorced group, which increased in size from 18 to 60 months (n = 24 at 18 months, n = 41 at 60 months). Always-single families included those continuously headed by single mothers from the 18- to 60-month assessments, with no known history of prior marriages or live-in partners during the target child's life (n = 51 at 18 months). Families who did not fall into one of the four aforementioned categories were excluded from analyses. Families excluded were those in which the mother was remarried or living with a nonbiologic partner, reconciled with the biologic father after a separation, widowed, single at 18 months but married by 60 months, or divorced or separated for the second time or more.

Comparisons between marital groups were made when sample sizes permitted (e.g., between to-be-divorced and always two-parent families at 18, 24, and 42 months; for always-single and always two-parent families at all time points), allowing us to test whether earlier child behavior, family relations, or sociodemographic factors accounted for group differences. For comparisons involving to-be-divorced or already-divorced families, child behavior, marital conflict, and parenting were assessed when partners were still living together.

Descriptive Statistics

In Table 1, descriptive statistics are presented for the ethnic composition of family structure groups. At each timepoint, a much greater percentage of Caucasian versus African American families were always two-parent (64-65% vs. 15-17%), and a significantly higher percentage of African American families were always single (32% vs. 4-5%). African American families also were significantly more likely to be in the already-divorced group starting at age 2. In addition, African American families were more likely to be in the "other" group at all ages (22-34% versus 10-21%), primarily due to establishing and ending relationships from previously single-parent families. Fifty-three percent of African American families who were married or living together with the biologic father when the child was 18 months became divorced or separated by the time the child was age 5, in comparison to the 19% of Caucasian families who divorced or separated between 18 and 60 months.

Marital Group Differences in Child Adjustment

The next set of analyses involved testing for marital group differences in child adjustment. To test the hypothesis that children from to-be-divorced and already-divorced families would show higher rates of behavior problems than those from always two-parent families, a series of planned t-tests were computed using mean TBC and CBCL scores as dependent variables. Results of these one-tailed t-tests are presented in Table 2. At 24 months, children in to-be-divorced families showed significantly higher CBCL Externalizing (t = 1.80, p < .05) and Internalizing (t = 1.80, p < .05) problems than children from always two-parent families. At 42 months, despite the small group size (n = 9), children from to-be-divorced families continued to show significantly higher Internalizing problems (t = 2.09, p < .05). Only one main effect was found between already-divorced and always two-parent families -- already-divorced boys demonstrated significantly higher Externalizing scores at 42 months (t = 1.82, p < .05). No significant differences in child behavior were found between to-be-divorced and already-divorced families. When always single-parent families were compared to other marital groups, boys from always-single families showed significantly higher Externalizing and Internalizing scores than boys from always two-parent homes at 24 (t = 1.82, p, < .05 for Externalizing; t = 3.21, p < .01 for Internalizing), 42 (t = 2.66, p, < .01 for Externalizing; t = 2.86, p < .01 for Internalizing), and 60 (t = 2.46, p, < .01 for Externalizing; t = 2.03, p < .05 for Internalizing) months according to both mothers and secondary caregivers (at 60 months for alternative caregivers, t = 2.17, p <.05 for Externalizing; t = 2.02, p < .05 for Internalizing). At 24 months, boys from always single-parent families also demonstrated higher rates of both Externalizing and Internalizing problems than those from recently divorced families (Externalizing, t = 1.69, p < .05; Internalizing, t = 2.18, p < .05). However, always single and already-divorced families did not differ significantly on child adjustment at older ages. Moreover, there were no significant differences between always single and to-be-divorced families on behavior problems at any age (18-42 months).

To examine whether or not marital group differences in child adjustment varied by ethnicity, a series of two-way ANOVAs were computed using ethnicity and family structure as independent variables and TBC and CBCL factors as dependent variables. When interactions with ethnicity were significant, t-tests were computed separately for African American and Caucasian subjects. In the first comparisons involving to-be-divorced and always two-parent families, a significant ethnicity by family structure interaction occurred for 24-month Externalizing, F(1, 147) = 4.82, p < .05 (2-tailed). Results are displayed in Figure 1. Caucasian boys from to-be-divorced families had significantly higher Externalizing scores than Caucasian boys from always two-parent families (t = 2.77, p < .01). For African Americans, there were no significant differences between to-be-divorced and always two-parent families with respect to 24-month Externalizing.

Marital Group Differences in Family Process

A similar set of t-tests was computed between marital groups for sociodemographic, parenting, and parental conflict/satisfaction measures. As illustrated in Table 3, to-be-divorced parents were significantly lower in SES at 42 months (t = -1.73, p < .05) and had more acrimonious marriages at 18 (t = -2.66, p < .01), 24 (t = 1.85, p < .05), and 42 (t = 1.67, p < .05) months than always two-parent families. Already-divorced families also differed from always two-parent families in expected ways, demonstrating lower per capita income at 24 (t = -1.79, p < .05) and 42 (t = -3.09, p < .01) months and higher scores on rejecting parenting at 24 months (t = 1.70, p < .05). As expected, always single-parents consistently demonstrated significantly lower family income and SES than always two-parent and to-be-divorced families (For comparisons with always two-parent families, all t-values for income > 5.0 and all p-values < .001; all t-values for SES > 3.0 and all p-values < .01. For to-be-divorced families, all t-values for income and SES > 2.0 and all p-values < .05). Always single-parent families also showed significantly lower family income and SES than divorced families at 24 (for income, t = -1.88, p < .05; for SES, t = -1.84, p < .05) and 42 (for income, t = -1.69, p < .05; for SES, t -2.10, p < .05) months. In addition, always-single mothers demonstrated higher levels of rejecting parenting compared to always two-parent mothers at 18 (t = 2.58, p < .01) and 24 (t = 2.53, p < .01) months and to-be-divorced parents at 18 months (t = 2.29, p < .05).

When interactions involving ethnicity and family structure were computed, an interaction was found with respect to rejecting parenting at 24 months, F (1, 143) = 11.16, p <.01 (two-tailed). In t-tests performed separately by ethnicity, to-be-divorced African American mothers were significantly more rejecting than always two-parent African American mothers (t = 3.09, p < .01); however, there were no significant differences between to-be-divorced and always two-parent Caucasian mothers. Scheffe post-hoc tests indicated that to-be-divorced African American mothers were significantly more rejecting than to-be-divorced Caucasian mothers, and more rejecting than both Caucasian and African American always two-parent mothers. A significant interaction was also found between per capita income at 18 months and ethnicity, F(1, 173) = 3.99, p < .05 (two-tailed). In t-tests performed separately by ethnicity, to-be-divorced African Americans had significantly lower income than always two-parent African Americans (t = -1.73, p < .05), but there were no significant differences in income between to-be-divorced and always two-parent Caucasians.

Factors Accounting for Marital Group Differences in Child Adjustment

The next set of analyses involved using family process variables as covariates to test whether their influence would account for marital group differences in behavior problems. For all covariate analyses, we considered the following family functioning variables as potential factors that might account for marital group differences in child adjustment: SES, per capita income, ethnicity, rejecting parenting, and marital conflict. However, covariate analyses were only conducted when the following criteria were met: (1) there was a significant difference between marital groups on child adjustment, and (2) there was a significant difference between marital groups on the family process variable. To limit the number of analyses but restrict case loss to a minimum, covariates were focused on the most proximal measure of each construct (e.g., for those divorced at 42 months, 24-month CRD was used rather than the 18-month Locke & Wallace). Analyses involving alternative caregiver reports of 60-month child behavior were not computed because cell sizes were typically less than 5 when ethnicity was included as a covariate. Because parenting measures were not available at 42 and 60 months, for covariate analyses at these ages a composite of 18- and 24-month rejecting parenting was created by averaging standardized scores to improve generalizability. Results are presented in Table 4.

The first set of ANCOVAs involved testing group differences between to-be-divorced and always two-parent families on 24- and 42-month CBCL factors. Group differences on Internalizing scores at 42 months were accounted for by the marital conflict measure concurrently. However, for 24-month Externalizing, differences between to-be-divorced and always two-parent Caucasian families remained once marital conflict at 24 months was taken into account. In the ANCOVA conducted between always two-parent and already-divorced groups, family structure differences on Externalizing at 42 months were accounted for by 24-month Externalizing.

Rejecting parenting accounted for marital group effects when used as a covariate in comparisons involving always-single and always two-parent groups for maternal report of Externalizing and Internalizing at 60 months. At 24 months, income, SES, and ethnicity were included as covariates in comparisons between already-divorced and always-single families; however, none of the covariates was significantly related to Externalizing or Internalizing in these ANCOVAs (Analyses were not presented due to space limitations).

Discussion

The present study sought to extend our understanding of the effects of marital structure and transitions on young boys' behavioral adjustment using a predominantly low-income, ethnically diverse sample. When to-be-divorced and always two-parent families were compared, modest but significantly higher rates of CBCL problems were evident for boys in to-be-divorced families at 24 and 42 months. To-be-divorced parents also were found to have less satisfactory marital relationships at 18 months, higher rates of childrearing disagreements at 24 months, higher rates of verbally aggressive conflict, and lower socioeconomic status at 42 months than always two-parent families. These pre-divorce differences appeared to hold true for both Caucasian and African American families, while only to-be-divorced African American mothers demonstrated significantly higher levels of rejecting parenting and lower levels of per capita income. When family relations were used as covariates to account for group differences between to-be-divorced and always two-parent families, measures of marital adjustment/conflict accounted for differences on children's Internalizing but not Externalizing problems. The results suggest that for young children in families about to separate, parental conflict may play an important role in the development of early internalizing symptoms. However, for boys' early externalizing problems, there appear to be other components of the pre-divorce environment in addition to parental conflict that contribute to disruptive behavior. One possible explanation involves changes in the stability of disruptive behavior, which has been found to increase from 24 months onward as the child is more able to express his/her feelings verbally and parents are likely to show less tolerance of children's aggressive behavior (Shaw & Bell, 1993). These developmental changes might help explain why group differences on externalizing-type symptoms were not present among any marital structure groups at 18 months, but were present at 24 months and thereafter in several instances.

Group differences between to-be-divorced and always two-parent families were more consistently found for internalizing symptoms, suggesting that young children may be more likely to respond to parental acrimony in the pre-divorce environment by demonstrating symptoms of distress and anxiety rather than acting in a disruptive manner. Unfortunately, there are few prospective studies of divorce that have been able to examine internalizing symptoms in preschool-age children, although a number of studies have documented higher rates of internalizing symptoms among school-age children and adolescents from divorced families (Furstenberg & Allison, 1989; Forehand, McCombs, Wierson, Brody, & Fauber, 1990; Hoyt, Cowan, Pedro-Carroll, & Alpert-Gillis, 1990). The differences on internalizing symptoms may reflect young children's immature cognitive understanding of parental conflict (Brody & Neubaum, 1996; Hetherington, 1991; Wallerstein & Kelly, 1980). Parental conflict may be more anxiety-inducing for young children given their greater dependence on the primary caregivers.

Results regarding differences in the pre-divorce adjustment of children are consistent with the studies of Block et al. (1986) and Cherlin (1992) showing that differences in adjustment patterns of to-be-divorced children are evident as early as two years of age. It should be noted that these effects were relatively modest, averaging .3-.4 standard deviations on CBCL factors, and more consistent for Caucasians than African Americans. The results also are in accord with the Block et al. (1986) and Shaw et al. (1993) studies, which found that pre-divorce family environments were characterized by more dissatisfied and conflictual parental relations.

In contrast to research with predominantly middle-class, Caucasian school-age children (e.g., Fauber et al., 1990; Hetherington et al., 1985), differences in problem behavior between already-divorced families and always two-parent families were found inconsistently. When differences did emerge, they were in accord with previous research -- already-divorced families showed lower income at all but 60 months, higher rates of rejecting parenting at 24 months, and higher rates of Externalizing problems at 42 months. However, pre-divorce child Externalizing scores accounted for these group differences in Externalizing problems. This further bolsters the finding that behavior problems associated with divorce often begin prior to parental separation, rather than as a consequence of divorce.

In the course of comparing to-be-divorced, already-divorced, and always two-parent families, we discovered several ethnic differences. The incomes of to-be-divorced and always two-parent Caucasian families did not differ. However, African American mothers in our sample were less likely to be married or living with a partner at each age, and more likely to be always single, compared to Caucasian mothers. Moreover, among those married or living with a partner when the target child was 18 months old, African American mothers were more likely to divorce or separate by age 5 than Caucasians. Interestingly, African American parents who were married or living together initially, but separated by the time the target child was age 5, had significantly less income than stable, two-parent African American families. However, the incomes of to-be-divorced and always two-parent Caucasian families did not differ significantly.

Similar findings have been reported in the sociological literature. For example, Wilson and Neckerman (1986) noted negative associations between the amount of money earned by husbands and the probability of divorce among African American couples. In fact, Wilson (1996) has attributed the high rate of marital instability among low-income, urban African American families to the high rate of unemployment found among lower-SES African American men, due to steady declines in the demand for low-skilled workers. Moreover, Bane (1986) pointed out that, in contrast to lower-income Caucasians, most poor, African American single-mothers could not have avoided poverty by maintaining the traditional two-parent family structure. Thus, it appears that poverty-stricken, African American couples more often separate, whereas African American couples who have more financial resources may tend to stay together.

We also found some ethnic differences in parenting among mothers who would eventually separate from their partners versus those already divorced. Specifically, for Caucasian mothers in the sample, levels of rejecting parenting at age 2 were higher among already-divorced mothers than among to-be-separated mothers; whereas for African Americans, to-be-separated mothers were more rejecting than already-separated mothers. On the other hand, to-be-separated Caucasian and African American couples did not differ significantly with respect to levels of reported marital satisfaction or conflict prior to separation. Thus, although both subgroups of mothers experienced conflictual relations with their spouses prior to separation, African Americans seemed to exhibit higher levels of disrupted parenting before divorce than afterwards, whereas Caucasian mothers tended to show more disrupted parenting following the marital separation than before. These data point to the possibility that effects of family structure on the family system may vary for low-income Caucasian and African American families. In order to understand these differences more clearly, it will be necessary to examine characteristics of the family context in more minute detail, such as investigating the specific reasons for parental separation as well as paternal psychological functioning prior to separation.

We found that rejecting parenting accounted for the higher levels of behavior problems found among boys from always single homes using maternal report of externalizing and internalizing problems. Many researchers have pointed to differences between the life circumstances of poor African American and Caucasian families, which might explain why the always single-parent family structure seems to be associated with worse outcomes for low-income, African American boys than low-income, Caucasian boys. Specifically, low-income African Americans tend to live in highly segregated neighborhoods characterized by a high proportion of unemployed and poverty-stricken residents, many female-headed households, and high rates of crime (Massey, Gross, & Shibuya, 1994; Sampson, 1987; South & Crowder, 1997, Wilson, 1996). (All of the aforementioned characteristics tend to co-occur in ghetto or underclass neighborhoods, although researchers often use the criterion of a poverty rate among residents greater than 40%). In contrast, poverty-stricken Caucasians rarely experience neighborhood conditions comparable to low-income African Americans. For example, in a high-risk sample of male adolescents recruited from the same city as our sample, Pittsburgh, Peeples and Loeber (1994) discovered that they did not have enough Caucasian participants living in underclass neighborhoods to permit statistical analyses on them. Only 2% of Caucasians lived in underclass neighborhoods, compared to 41% of African Americans in their sample.

Living in underclass neighborhoods presents parents with stressors not experienced to the same degree by residents of working-class neighborhoods, where low-income Caucasians typically live. For example, ghetto residents often believe they cannot trust their neighbors due to high rates of crime and substance abuse among residents in the community. Many fear for the safety of

themselves and their children to the extent that they may curtail activities outside the home, sometimes even preventing their children from participating in structured activities at school, church, or neighborhood organizations (Furstenberg, 1993). Supportive characteristics of more organized communities, such as neighbors who watch over each others' property and assume

responsibility for supervising the activities of children in the neighborhood, might be nonexistent in underclass neighborhoods (Sampson, 1987). Thus, single-mothers in underclass neighborhoods may receive less caregiving support than other single-mothers, hampering their ability to provide nurturant caregiving.

There are methodological limitations of the study that need to be mentioned. The participants represent a restricted low-income community sample. By recruiting families from WIC, the range of dysfunction was most likely curtailed, as all participating families were involved in a program designed to improve the quality of their children's nutritional needs. Thus, we believe our results may be limited to relatively high-functioning low-income families. Similarly, because we restricted recruitment to families with more than one child, the study's results are only applicable to low-income families with multiple children. Similar prospective studies of girls are needed particularly in light of the findings regarding risk of internalizing problems in children from to-be-divorced families. Sample sizes for specific comparisons were variable. Although we eliminated analyses that included cell sizes less than 5, there was still great variability particularly in analyses in which ethnicity was considered. Replication of these findings needs to be carried out in larger epidemiological studies now that salient issues have been identified, including ethnicity, pre-divorce adjustment and family environment, and the nature of single-parent status. Finally, we used conservative criteria to identify families with different marital structures. In doing so, a large minority of families were not used in the present analysis because they did not fit our criteria (e.g., single parents who eventually marry/live together, divorced families who remarry). For better or worse, instability in marital structure is becoming more common among both Caucasian and African American families (Wilson, 1996); thus, as researchers we need to develop more sophisticated methods of categorizing families that truly reflect the family's parental structure. In coding marital structure among our subjects, we were struck by its diversity, and frustrated because of insufficient numbers to study important sub-types (e.g., African American families that began as always-single but became two-parent families for some or part of the child's first 5 years).

Despite these caveats, this study provides important prospective data on the effects of marital structure and family processes in relation to the adjustment of low-income, young boys from urban backgrounds. It suggests that ethnicity needs to be considered in assessing the consequences of the pre- and post-divorce environment for children, and that preschool boys in pre-divorce environments may be likely to show internalizing symptoms, and to some extent externalizing problems, in response to the parents' impending separation. Many of the effects of both the pre- and post-divorce context appear to be accounted for by parental conflict or earlier child behavior. In the case of single-parent families, rejecting parenting appeared to play a similar role in accounting for the effects of family structure on child adjustment.

Author Note

The research reported in this paper was supported by grants to Daniel Shaw from the National Institute of Mental Health, grant MH 46925 and MH 50907. We greatly appreciate the assistance of Robert Emery in reviewing an earlier version of the manuscript.

Correspondence concerning this article should be addressed to the first author at the following address: Department of Psychology, University of Pittsburgh, 4015 O'Hara Street, 604 OEH, Pittsburgh, PA, 15260-0001.

References

Achenbach, T.M. (1991). Manual for the Child Behavior Checklist/4-18 and 1991 profiles. Burlington, VT, University of Vermont Department of Psychiatry.

Achenbach, T.M. (1992). Manual for the Child Behavior Checklist 2/3 and 1992 profile. Burlington, VT, University of Vermont Department of Psychiatry.

Amato, P. & Keith, B. (1991). Parental divorce and the well-being of children: A meta-

analysis. Psychological Bulletin, 110, 26-46.

Amato, P.R. & Rezac, S.J. (1994). Contact with nonresident parents, interparental conflict, and children's behavior. Journal of Family Issues, 15, 191-207.

Bane, M. J. (1986). Household composition and poverty: Which comes first? In S. Danzinger & D. Weinberg (Eds.), Fighting poverty: What works and what doesn't. Cambridge, MA: Harvard University Press.

Baruch, D. W., & Wilcox, J. A. (1944). A study of sex differences in preschool children's coexistent with interpersonal tensions. The Journal of Genetic Psychology, 64, 281-303.

Baumrind, D. (1973). The development of instrumental competence through socialization. In A. Pick (Ed.), Minnesota Symposia on Child Psychology (Vol. 7, pp. 3-46). Minneapolis: University of Minnesota Press.

Block, J.H., Block, J., & Gjerde, P.F. (1986). The personality of children prior to divorce: A prospective study. Child Development, 57, 827-840.

Block, J.H., Block, J., & Gjerde, P.F. (1988). parental functioning and the home environment in families of divorce: Prospective and concurrent analyses. Journal of the American Academy of Child and Adolescent Psychiatry, 27, 207-213.

Brody, G.H. & Neubaum, E. (1996). Family transitions as stressors in children and

adolescents. In C.R. Pfeffer (Ed.), Severe Stress and Mental Disturbance in Children.

Washington, D.C.: American Psychiatric Press, 559-590.

Camara, K.A. & Resnick, G. (1988). Interparental conflict and cooperation: Factors moderating children's post-divorce adjustment. In E.M. Hetherington & J. Aratesh (Eds.), Impact of divorce, single parenting, and step-parenting on children (pp. 169-195). Hillsdale, NJ: Erlbaum.

Cherlin, A.J. (1992). Marriage, divorce, & remarriage. Cambridge, MA: Harvard University Press.

Cherlin, A.J., Furstenberg, F.F., Chase-Landsdale, P.L., Kiernan, K.E., Robins, P.K., Morrison, D.R., & Teitler, J.O. (1991). Longitudinal studies of effects of divorce on children in Great Britain and the United States. Science, 252, 1386-1389.

Chess, S., & Thomas, A. (1984). Origins and evolution of behavior disorders from infancy to early adult life. New York: Brunner/Mazel.

Christopoulos, C., Cohn, D.A., Shaw, D.S., Joyce, S., Sullivan-Hanson, J., Kraft, S., & Emery, R.E. (1987). Children of abused women: I. Adjustment at time of shelter residence. Journal of Marriage and the Family, 49, 611-619.

Davies, P.T. & Cummings, E.M. (1994). Marital conflict and child adjustment: An emotional security hypothesis. Psychological Bulletin, 116, 387-411.

Dornbusch, S., Carlsmith, J. Bushwall, S., Ritter, P., Leiderman, H., Hastorf, A., & Gross, R. (1985). Single parents, extended households, and the control of adolescents. Child Development, 56, 326-341.

Duncan, G., Brooks-Gunn, J., Klebanov, P. (1994). Economic deprivation and early childhood development. Child Development, 65, 296-318.

Emery, R. E. (1988). Marriage, divorce, and children's adjustment. Beverly Hills: Sage.

Emery, R. E., Hetherington, E. M., & DiLalla, L. F. (1984). Divorce, children, and social policy. In H. W. Stevenson & A. E. Siegel (Eds.), Child development research and social policy. Chicago: University of Chicago Press.

Emery, R.E., & O'Leary, K.D. (1982). Children's perceptions of marital discord and behavior problems of boys and girls. Journal of Abnormal Child Psychology, 10, 11-24.

Erel, O, & Burman, B. (1995). Interrelatedness of marital relations and parent-child relations: A meta-analytic review. Psychological Bulletin, 118, 108-132.

Farnworth, M. (1984). Family structure, family attributes, and delinquency in a sample of low-income, minority males and females. Journal of Youth and Adolescence, 13, 349-364.

Fauber, R., Forehand, R., Thomas, A., & Wierson, M. (1990). A mediational model of the

impact of marital conflict on adolescent adjustment in intact and divorced families: The role of disrupted parenting. Child Development, 61, 1112-1123.

Forehand, R., Armistead, L., & David, C. (1997). Is adolescent adjustment following parental divorce a function of predivorce adjustment? Journal of Abnormal Child Psychology, 25, 157-164.

Forehand, R., McCombs, A.T., Wierson, M., Brody, G., & Fauber, R. (1990). Role of

maternal functioning and parenting skills in adolescent functioning following parental

divorce. Journal of Abnormal Psychology, 99, 278-283.

Furstenberg, F. F., & Allison, P. D. (1989). How marital dissolution affects children: Variations by age and sex. Developmental Psychology, 25, 540-549.

Furstenberg, F. F., Jr., Belzer, A., Davis, C., Levine, J., Morrow, K., & Washington, M. (1993). How families manage risk and opportunity in dangerous neighborhoods. In W. Wilson (Ed.), Sociology and the public agenda. Newbury Park, CA: Sage Publications.

Gardner, F. (1989). Inconsistent parenting: Is there evidence for a link with children's conduct problems? Journal of Abnormal Child Psychology, 17, 223-233.

Glick, P. C., & Lin, S. (1986). Recent changes in divorce and remarriage. Journal of Marriage and the Family, 48, 737-747.

Grych, J.H. & Fincham, F.D. (1990). Marital conflict and children's adjustment: A cognitive-contextual framework. Psychological Bulletin, 108, 267-290.

Haurin, R.J. (1992). Patterns of childhood residence and the relationship to young adult

outcomes. Journal of Marriage and the Family, 54, 846-860.

Hetherington, E. M., Cox, M., & Cox, R. (1979). Family interaction and the social, emotional, and cognitive development of children following divorce. In V. Vaughn & T. Brazelton (Eds.), The family: Setting priorities (pp. 89-128). New York: Science and Medicine.

Hetherington, E. M., Cox, M., & Cox, R. (1985). Long-term effects of divorce and remarriage on the adjustment of children. Journal of the American Academy of Child Psychiatry, 24, 518-530.

Hetherington, E.M. (1991). Families, lies, and videotapes. Presidential address to the meeting of the Society for Research on Adolescence, Atlanta, GA, 1990. Journal of Research on Adolescence, 1, 323-348.

Hoyt, L.A., Cowen, E.L., Pedro-Carroll, J.L., & Alpert-Gillis, L.J. (1990). Anxiety and

depression in young children of divorce. Journal of Clinical Child Psychology, 19, 26-32.

Jouriles, E.N., Murphy, C. M., Farris, A.M., Smith, D. A., Richters, J.E., & Waters, E. (1991). Marital adjustment, parental disagreements about child rearing, and behavior problems in boys: Increasing the specificity of the marital assessment. Child Development, 62, 1424-1433.

Larzelere R.E., Martin, J.A., Amberson, T.G. (1989). The toddler behavior checklist: A parent-completed assessment of social-emotional characteristics of young preschoolers. Family Relations 38, 418-425.

Locke, H.J., & Wallace, K.M. (1959). Short marital-adjustment and prediction tests: Their reliability and validity. Marriage and Family Living, 21, 251-255.

Loeber, R., & Dishion, T.J. (1983). Early predictors of male delinquency: A review. Psychological Bulletin, 94, 68-99.

Massey, D. S., Gross, A. B., & Shibuya, K. (1994). Migration, segregation, and the geographic concentration of poverty. American Sociological Review, 59, 425-445.

McCord, W., McCord, J., & Zola, I.K. (1959). Origins of crime. New York: Columbia University Press.

McLeod, J., Shanahan, M. (1993). Poverty, parenting, and children's mental health. American Sociological Review, 58, 351-366.

Montare, A., & Boone, S. (1980). Aggression and paternal absence: Racial-ethnic differences among inner-city boys. Journal of Genetic Psychology, 137, 223-232.

Minuchin, S. (1974). Families and family therapy. Cambridge, MA: Harvard University Press.

Offord, D. R., Sullivan, K., Allen, N., & Abrams, N. (1979). Delinquency and hyperactivity. Journal of Nervous and Mental Disease, 167, 734-741.

Patterson, G. (1982). Coercive family processes (Vol. 3). Eugene, OR: Castalia.

Peeples, F., & Loeber, R. (1994). Do individual factors and neighborhood context explain ethnic differences in juvenile delinquency? Journal of Quantitative Criminology, 10, 141-157.

Rothbaum, F., & Weisz, J. (1994). Parental caregiving and child externalizing behavior in nonclinical samples: A meta-analysis. Psychological Bulletin, 116, 55-74.

Sampson, R.J. (1987). Urban black violence: The effect of male joblessnes and family disruption. American Journal of Sociology, 93, 348-382.

Shaw, D.S., & Bell, R.Q. (1993). Developmental theories of parental contributors to antisocial behavior. Journal of Abnormal Child Psychology, 21, 493-518.

Shaw, D.S., Emery, R.E., & Tuer, M.D. (1993). Parental functioning and children's adjustment in families of divorce: A prospective study. Journal of Abnormal Child Psychology, 21, 119-134.

Shaw, D.S., Keenan, K., & Vondra, J.I. (1994). The developmental precursors of antisocial behavior: Ages 1-3. Developmental Psychology, 30, 355-364.

Shaw, D. S., Winslow, E., Owens, E., Vondra, J., Cohn, J., & Bell, R.Q. (1998). The development of early externalizing problems among children from low-income families: A transformational perspective. Journal of Abnormal Child Psychology, 26, 95-107.

South, S. J., & Crowder, K. D. (1997). Escaping distressed neighborhoods: Individual, community, and metropolitan influences. American Journal of Sociology, 102, 1040-1084.

Strauss, M. (1979). Measuring intrafamily conflict and violence: The Conflict Tactics Scales. Journal of Marriage and the Family, 41, 75-88.

Thomas, A., & Chess, S. (1977). Temperament and development. New York: Brunner/Mazel.

U.S. Bureau of the Census. (1992). Current Population Reports, Series P-60, No. 181. Poverty in the United States: 1991. Washington, DC: U.S. Government Printing Office.

United States Department of Health and Human Services (1990). Monthly vital statistics report, 13, No. 13.

Wallerstein, J.S. (1991). The long-term effects of divorce on children: A review. Journal of the Academy of Child and Adolescent Psychiatry, 30, 349-360.

Wallerstein, J. S., & Kelly, J. B. (1980). Surviving the breakup. New York: Basic Books.

Winslow, E.B., Shaw, D.S., Bruns, H., & Kiebler, K. (1995, April). Parenting as a mediator of child behavior problems and maternal stress, support, and adjustment. Paper presented at the biennial meeting of the Society for Research in Child Development, Indianapolis, IN.

Wilson, M., & Tolson, T. (1990). Familial support in the black community. Journal of Clinical Child Psychology, 19, 347-355.

Wilson, W.J. (1996). When work disappears: The world of the urban poor. New York: Alfred Knopf.

Wilson, W. J., & Neckerman, K. (1986). Poverty and family structure: The widening gap between evidence and public policy issues. In S. Danzinger & D. Weinberg (Eds.), Fighting poverty: What works and what doesn't (pp. 232-259). Cambridge, MA: Harvard University Press.

Zill, N, Morrison, D.R., & Coiro, M.J. (1993). Long-term effects of parental divorce on

parent-child relationships, adjustment, and achievement in young adulthood. Journal of

Family Psychology, 7, 91-103.

Zimmerman, M., Salem, D., Maton, K. (1995). Family structure and psychosocial correlates among urban African-American adolescent males. Child Development, 66, 1598-1613.