Although child neglect is the most prevalent form of child maltreatment, there has been a comparative lack of empirical research into the characteristics of neglecting families and the psychological consequences of neglect. In studying neglectful families, existing research has failed to distinguish between two forms of neglect -- denial of critical care and supervision -- and to use micro-social coding of direct observations of parent-child interactions to develop an understanding of the parent-child relations in neglecting families. The work will then be able to test the hypothesis that it is the relational aspects of neglect that determine the impact of neglect on the psychological outcome for the child. Because neglect is often associated with physical abuse and punitive discipline, the proposed research is also designed to understand the relative importance of neglect and punitive discipline in the development of children's aggression. By distinguishing between instrumental or proactive, and irritable or reactive aggression, the research will be able to determine whether neglect and punitive discipline differentially influence the development of two different kinds of aggression in young school-age children. To conduct the research and to assure a sample that is ethnically diverse and drawn from both rural and urban areas, 270 maltreating families and 270 economically disadvantaged, or high risk, families will be recruited in Iowa, Nebraska, and Wisconsin. Structural equation modeling will be used to determine the role of neglect, punitive discipline, and parent attributes, as well as the mediating influences of peer interaction and social competence, on young children's aggression. Because all of the constructs studied will be based on a multimethod/multisource approach, including direct observations of family interaction and social competence in children, and peer nomination indices of aggression, the research is designed to provide an understanding of family function and children's aggression that is not compromised by shared method and shared source variance.
