Study data were obtained from the 1998 Minnesota Student Survey (MSS), an anonymous questionnaire that has been administered to public school students in the 6th, 9th, and 12th grades every 3 years since 1989. The 1998 MSS included questions about family, peers, school, and activities. Of the 133,794 survey respondents, 131,281 (98 percent) answered at least 1 of the questions that pertained to physical violence in the family. For the purpose of this study, physical violence in the family was defined as hitting hard enough to leave a mark or to make the victim afraid of the perpetrator. Cross-tabulation analyses showed statistically significant relationships between the variables of grade, race/ethnicity, and family structure and the variables of attempted suicide, fighting, and gun-carrying at school. The analysis controlled for the demographic variables in the logistic regression equation. Logistic regression was used to assess the effect of witnessing physical violence in the family, being the object of physical violence, and both witnessing and being the object of violence. Any physical violence in the family correlated significantly with having attempted suicide, fighting, and gun-carrying. In bivariate analyses, all of the family violence variables assessed were significantly associated with attempting suicide, physical fighting, and gun-carrying by both girls and boys. Direct victims of family violence were more likely to engage in self-directed and other-directed violence than were youth who only witnessed violence. Thus, programs aimed at decreasing juvenile violence should address the role of violence in the family. 3 tables and 33 references