Participants in this study are part of a larger longitudinal study entitled LONGSCAN, which is a longitudinal investigation of the antecedents and consequences of child abuse/neglect. This investigation included all of the LONGSCAN children who had completed interviews when they were 4 and 6 years old and were living with their biological mothers at the time of the interviews. A total of 682 subjects from 4 different regions of the country met these criteria. The mothers or other primary caregivers were asked to participate in a 2-hour face-to-face interview that was composed of both standardized and project-developed measures. In addition, each child was administered a number of age-appropriate measures during brief sessions. The variable of family violence was defined as "violent or aggressive behavior that involved family members or intimate partners." A distinction was made between violence that was witnessed by children and violence that was directed at them. In addition, this violence was further differentiated in terms of whether it involved physical acts of aggression or psychological or verbal acts of aggression. These measures were administered at the age-6 interview. After controlling for mothers' reports of child problem behaviors on the Child Behavior Checklist at age 4, the study found that subsequent exposure to family violence predicted reported problem behaviors at age 6. Although mothers' reports of child victimization predicted subsequent problem behaviors, witnessed violence was related to these problems only when both mothers and children reported its occurrence. These study results suggest that even though there was a relationship between witnessed and directly experienced family violence, both had independent, noninteractive effects on subsequent behavior problems. The findings suggest that attempts to understand the impact of violence on children and youth will not be complete without considering exposure to family violence as a set of complex, independent predictors. Further research should continue to differentiate and better assess types of violence exposure within prospective designs, while also assessing different domains of functioning using a variety of informants. 5 tables and 36 references