Principal Investigators: Sharlene Wolchik and Irwin Sandler

Level of Intervention: Selected

Target Population: Parents divorced within 2 years, with a child between 8 and 15 years of age

References: Wolchik, West, Westover, Sandler, Martin, Lustig, Tein, and Fisher (1993)

Theory (Risk and Protective Factors Targeted):
Parental divorce is a stressful life event and has the potential to affect children’s adjustment in significant ways. Divorce has been linked with childhood aggression, internalizing symptoms, poor academic performance, and poor peer relations. The authors stressed the importance of using an empirical approach to the design and evaluation of prevention programs. Based on "small theory", the Children of Divorce Parenting Program targeted a number of modifiable processes ("putative mediators") that have been identified in past research as being associated with the quality of children’s adjustment after divorce: 1) quality of the child’s relationship with the custodial parent, 2) contact with the noncustodial parent, 3) negative divorce-related events including interparental conflict, 4) support from nonparental adults, and 5) discipline strategies.

Description of Intervention:
The Children of Divorce Parenting Program is a parent-based intervention designed to improve the quality of the parent child relationships by encouraging parents to spend quality time with their children, listen to their children, and reinforce positive behavior. Parents are taught how to use clear and consistent discipline practices, and to use anger management skills to reduce interparental conflict. In addition, parents are made aware of the importance of the father-child relationship and non-parental adults as a source of social support for the child, and given the opportunity to problem solve ways to establish these resources. The program consisted of 10 group and 2 individual sessions that were presented to small groups of 6 to 8 participants. Each group was co-led by a male-female team that were primarily graduate students in clinical psychology.

Research Subjects:
Mothers who were divorced within the past 2 years were recruited for participation in the intervention through a random sampling of court records, media articles, and school presentations. A two-step screening process was used to identify subjects. In an initial phone call, parents were interviewed to determine if the family met 7 criteria: 1) the divorce was granted within the last 2 years, 2) the mother had a child between 8 and 15 years of age, 3) the custodial parent was female, 4) no family member was in treatment, 5) the custodial parent had no plans to remarry, 6) the custody arrangement was stable, and 7) English was the primary language in the home. In a pre-test interview, subjects were eliminated (46 out of 177 excluded) if their scores on the putative mediators indicated low-risk (scores at or above 30%) or if there was evidence that the child or mother reported clinical levels of depression (21 children and 1 mother excluded). There were 94 families assigned to groups and 70 who remained in the study through post-test. Ninety percent of the participants were Caucasian, and 61% of the children interviewed were male. The average yearly income was between 20,001 to 25,000 (range = less than 5,000 to 50,000) and 74% of he mothers had attended college.

Research Design:
Participants were randomly assigned to an intervention or wait-list control group.

There were no significant group differences in attrition rates between those individuals that dropped out of the program and those that remained through post-test. When differential attrition rates were compared by intervention condition, only two significant group differences were found. According to child reports, control subjects who completed the program were significantly less aggressive (p<.01) and slightly less depressed (p<.10) than those who dropped out of the program. Within the intervention group, control subjects who completed the program tended to report higher levels of depression than those who left the program. The authors noted that these group differences would most likely bias effects against showing favorable outcomes for treatment.

A series of ANCOVA were conducted in which pre-test scores were used as covariates and post-test scores were used as dependent variables.

Parents who participated in the intervention reported significantly lower (p<.05) levels of total problem behaviors on the Child Behavior Checklist (CBCL; Achenbach & Edelbrock, 1983) compared to control parents.

Children of intervention participants reported significantly lower levels (p<.01) of aggressive behavior on the Youth Report of Hostility Scale (Cook, 1985) compared to controls, the intervention children reported significantly higher levels of depressive symptoms (p<.05) on the Child Assessment Schedule (CAS; Hodges, Kline, Stern, Cytryn, & McKnew, 1982) compared to controls. The authors qualified this finding by pointing out that there were no group differences on Child Depression Inventory (CDI; Kovacs, 1981), and that the difference on the CAS may have been due to an extreme outlier in the sample. There was a possibility of differential attrition on child ratings but similar results were found when this was controlled in subsequent analyses.

On the mediating variables significant group differences were also found. A number of these involved interactions with pre-test level. The children of parents in the program who perceived better quality relationships (as reflected in higher Acceptance/Rejection ratings) prior to the intervention, reported significantly higher levels of warmth and acceptance (p<.05) in their relationship with their parent compared to controls. One counter-intuitive finding with child reports was that program subjects reported receiving less support from non-parental adults than children in the control group (p<.05).

On parent ratings of mediating variables, group differences were found on two measures of the quality of the custodial parent-child relationship. Parents in the program rated the quality of communication (p<.01) and the positive routines in their family (p<.01) more positively than controls. In addition, at post-test mothers who participated in the program were significantly more likely than controls to be willing to change visitation by the ex-spouse if it was requested (p<.05). Interactions were found between intervention condition and pre-test levels on two of the parent measures. Mothers who reported less consistent discipline at pre-test reported significantly more improvement at post-test compared to controls (p<.01). When mothers reported high numbers of negative events at pre-test, program participants reported significantly lower rates of negative events at post-test.

Mediational analyses were conducted and confirmed that the quality of the custodial mother-child relationship mediated the effect of the intervention on changes in child behavior. Specifically, 43% of the change in behavior was attributed to mediation by these parent-child relationship factors.

Strengths and Limitations:
The Children of Divorce Parenting Program targets a number of modifiable factors under parental control that have been identified in past research as being associated with the quality of children’s adjustment after divorce. It is a parent-based intervention program for mothers with a child between 8 and 15 years of age, who divorced within the last two years. According to parent reports, children of participants evidenced fewer problem behaviors at the end of the program compared to controls. Child reports of behavior change were mixed and somewhat difficult to interpret. There was support from the parent perspective that the program affected mediating factors, related to child behavior. Indeed, forty-three percent of the main effect of the intervention on post-test changes in maternal ratings of child behavior were mediated by improvements in the parent-child relationship. While these results of this randomized trial are promising, they should be interpreted cautiously. The sample size was very small, unbiased reporters did not make the ratings, and there have been no independent replications of the program.

The authors made significant efforts towards program fidelity, which should be noted. Leaders received intensive training regarding the theoretical basis of the program. They were also trained in the program with videotapes of sessions and role-plays. Group leaders used detailed program outlines to guide the sessions and received weekly supervision (1.5 hours). A 5-part process evaluation was conducted that included monitoring how much time was spent on each component of the sessions, attendance of participants, ratings of the leader’s knowledge of the program, participant evaluations of the leaders, and parent ratings of their usage of the program. Results of this evaluation indicated that participants attended a mean of 9.6 of the group sessions, all major components were covered, and evaluations of the group leaders were very positive.

back to text     back to table of contents     external links to program     contact investigator