The purpose of the study was to provide empirical data to inform the debate and lead to meaningful answers to important questions in custody mediation and domestic violence including equity of outcomes for families with partner violence versus those without any evidence of violence, and the issue of health and safety outcomes for the children.
In order to examine custody mediation and domestic violence, researchers examined every seventh action filed in the Family Court in San Diego during the calendar year 1996. Researchers eliminated cases in which the parties resolved the dispute either before or during mediation. All remaining cases were those in which there was a mediation that failed to resolve the dispute, either in full or in part. Each of these 512 mediations resulted in a formal report with custody recommendations by the mediator. Each nonagreement mediation report and its associated Court Screening Form were scanned in its entirety and subsequently printed. Mediation reports ranged from 2 to 11 pages in length, and in general, contained 2 sections. In one section, the mediator made detailed recommendations about custody and visitation. In the other section, the mediator wrote a narrative describing the mediation process and the rationale for any recommendations given. Each report also contained an appended page containing a detailed parenting plan in which the mediator summarized the percentage of physical custody recommended for each parent. A content analysis system was developed to systematically capture the information in the mediator's report. Each report was studied and content analyzed by three trained raters who had no access to the Court Screening Form and so were blind to any a prior domestic violence allegations. The majority of coding items called for yes/no responses. Other coding items allowed the raters to choose from among more than two options. A variable was coded in a particular direction if all three raters agreed or if two of the three raters agreed. In only rare instances was there no consistency among the three raters. In those instances, the investigators independently evaluated the reports, so that each final data point was based on full agreement by at least two people. The domestic violence sample (n = 200) contained mediations in which there were indications of domestic violence in the case file available to the mediator or the mediator identified domestic violence. The nondomestic violence sample (n = 200) contained cases with no indicators of domestic violence in the file and no mention of domestic violence in the mediation report.
The study had two randomly selected samples. The starting point for the study sample was every single action filed in the Family Court in San Diego during the calendar year 1996. Every seventh filing was examined. If there was a custody dispute in that filing, it was selected for further study. If there was no custody dispute, the researchers examined each subsequent filing sequentially and selected the first custody dispute that could be identified. From the resultant 948 files, researchers eliminated cases in which the parties resolved the dispute either before or during mediation. All remaining cases were those in which there was a mediation that failed to resolve the dispute, either in full or in part. Each of these 512 mediations resulted in a formal report with custody recommendations by the mediator. The domestic violence sample (n = 200) included all of the selected cases with domestic violence indicators on the current screening form, prior allegation of domestic violence on a court screening form and/or a permanent domestic violence restraining order in the case file that was available to the mediator, and clear cases of domestic violence as documented by the mediator in which there was absolutely no indication of domestic violence on any screening form or in the case file. The nondomestic violence sample (n = 200) included 200 of the remaining 312 cases that had no indicators of domestic violence in the file and no mention of domestic violence in the mediation report. A random number generator was used to eliminate nondomestic violence cases until researchers were left with a sample size of 200 cases.
Every action filed in the Family Court in San Diego during the calendar year 1996.
Family Court filing
The data are record abstracts collected from actions filed in the Family Court in San Diego during the calendar year 1996.
administrative records data
The data include variables from five categories, including general case information, parental factors, child factors, family interaction factors, and current parental relationship. Case information variables include previous mediation, legainclue
l and physical custody recommendations, and types of visitation. Parental factors include mother's living arrangements, occupation, and work schedule, father's living arrangements, occupation, and work schedule, parental concern for child's well-being, current involvement with child, parental physical appearance, body language, emotional factors, hostility, and willingness to compromise, and drug and alcohol use. Child factor variables include child's living arrangements, day care plan, separation from primary caregiver, child's regressive behaviors, age, gender, and past occurances of child abuse. Family interaction factors include domestic violence mentioned, child exchange difficulties, and extended family involvement. Current parental relationship variables include parent's current relationships, parental hostility, parent/child interaction, allegations regarding safety concerns, legal concerns or threats, and challenges to current visitations.
Likert-type scales were used.