Systematic naturalistic observation often contradicts experimental laboratory findings and idiosyncratic personal perceptions. In this study, family disputes managed by police officers trained in interpersonal conflict management yielded uniform observational data on 1,388 cases. The view shared by police and by social scientists that family disputes are likely to involve assaultiveness and that such behavior is typically caused by alcohol use was not supported by these data. Instead, the findings suggest that: assaults do not usually precede arrival of police; disputes are not usually influenced by alcohol use; and, indeed, assaults are less common when alcohol has been used.