While self-organized groups are the subject of contemporary social science research, measuring how discrete entities contribute to the formation of groups remains a challenge. Recent advances in computer-based simulation, specifically Agent Based Models (ABM), provide a tool to address this problem.  ABMs allow the investigator to ascribe relevant attributes to each individual (i.e., agent), and moreover, implement algorithms that modify the environmental field that permits the expression of these attributes. 

To illustrate the usefulness of ABM as a computational tool, we show how PlayMate, a multi-threaded ABM, simulates the formation of playgroups in preschool children.  The current version attempts to identify attributes generated by, and reflective of, peer dynamics; in addition, homophily, sex, and memory are used as other critical features of group emergence.  While the model remains specific to play partner preferences and peer processes, we attempt to establish the validity of such an approach for answering broader social questions.