The study found that family disorganization measures were the most significant predictors of violence between siblings, overriding the characteristics of children or particular family demands. Family-disorganization factors that were strongly related to sibling violence were parent-to-child violence, mother's disciplinary practices, and father's intimidation of the children. These factors apparently contributed to a child's aggressive impulse to physically attack a sibling. Younger children and, to a lesser extent, male children were more likely to be violent toward their siblings. As the age of the child increased, rates of sibling physical violence declined. There was a link between marital problems and sibling physical violence. This study used a subsample of the married couples who were interviewed in the 1976 National Survey of Physical Violence in American Families. The subsample consisted of 994 respondents who had at least 2 children living at home. One person from each couple was interviewed about a range of witnessed and experienced family violence, including the amount of directed or initiated aggression by a child randomly selected in the interview protocol. The primary dependent variable was the frequency of physical violence among siblings in the past year. The measure of sibling violence included only the items on the Conflict Tactics Scale that represented more serious violence. Independent variables pertained to sociodemographic and family composition, including percentage of male children, parents' marital status, family resources, family stability, involvement with kin and community, parental models of punishment, and the family's power structure. 4 tables, 52 references, and appended relevant items of the Conflict Tactics Scale