In this study, involving 52 families consisting of 134 family members, the researcher examined the difference in family relations, as measured by Beaver's Self-Report Family Inventory, between families participating in specified group family play and those not participating. The researcher used random sampling to assign the families to four groups according to the Solomon Four Group Design. A one-way analysis of variance found a difference between the groups that would not have occurred by chance. The pair-wise t-tests and Bartlett's Chi Square found that the experimental groups showed significant improvement in family relations and that these differences were not due to testing effects or maturation.A significant relationship between group family play and improved family relations was found at the oc = .001 level of significance. The findings support the use of family play in group settings in both prevention and therapeutic milieus.Families in the 1990s consist of a mosaic of structures, with their composition often changing. Stresses are increasing on the family from a variety of directions, including financial, substance abuse, and alternative lifestyles. Families need to acquire rapid adaptation skills and to develop problem solving skills to keep pace with the demands of an increasingly complex world (Chasin & White, 1989;Harvey, 1990;Keller & Bromley, 1989). As a result, families are now turning to the institutions of society for the support and care once sought from extended families (Josephson, 1993; Sweeney & Landreth, 1993).Churches, schools and non-profit organizations are functioning on ever-tightening budgets, which limits the programs they can Shari E. Duff, Southwest Baptist Theological Seminary This document is copyrighted by the American Psychological Association or one of its allied publishers.This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.