Consistent with previous research, the results support the hypothesis that offenders are much more likely to be drinking when they physically assault strangers than when they physically assault people they know. Offenders who committed sexual assaults were no more likely to be drinking than offenders who committed physical assault. In summation, alcohol can contribute to any type of violent outcome, but it is much more important for some types of violence than others. The results suggest that it is particularly likely to increase the likelihood of a physical assault on a stranger, and it is least important for assaults on partners. This pattern is attributed to the nature of conflict in different relationships, however, other explanations are acknowledged. This pattern is strong enough to justify consideration in discussions of the role of alcohol in violent crime. Tables, figures, references