The proposed project is a 6 year follow-up of 249 families who participated in an experimental evaluation of a prevention program for children of divorce. The evaluation included three randomly assigned conditions; a mother-only program, a combined program which involved separate mother and child components and a self-study, guided reading condition. The interventions targeted empirically-supported risk and protective factors (putative mediators) and were well-implemented with a high degree of fidelity. Participant retention was excellent, and multiagent, multimethod assessments were conducted. Analyses of immediate posttest and short-term follow-up data revealed positive effects on the targeted putative mediators, as well as on children's mental health outcomes. Although several studies have demonstrated similar immediate and short-term mental health gains for preventive interventions targeting children of divorce, researchers have not yet addressed whether positive program effects persist into adolescence. Long-term follow-up is particularly important for this population, given that children of divorce are at an elevated risk for developing problems that normatively increase during adolescence. The proposed project has four specific aims: to examine whether there are persistent program effects on behavioral and mental health problems during adolescence; to examine whether program effects obtained at the 6 year follow-up are mediated by improvement on the putative mediators targeted in the intervention; to examine moderators of long term intervention effects; and to test alternative theoretical models in which stressful contextual factors, mother-child relationship variables, and child characteristics assessed in childhood are prospectively linked to mental health outcomes in adolescence. Multimethod, multiagent assessments of internalizing and externalizing problems and behavioral and mental problems that are rare in childhood but more prevalent during adolescence will be conducted. Adolescents, mothers, (step)fathers, and teachers will complete questionnaire data; mothers and adolescents will engage in problem-solving videotaped interactions; archival data will be collected from school records. Analyses of covariance, latent growth curve modeling, structural equation modeling, and survival analyses will be employed to test the specific aims.
