While the empirical association of drinking and problem drinking to violence is well established, the etiological nature of the relationship is poorly understood. Using data collected from 1,149 convicted male felons, the acute (drinking just before the violent event) and chronic (a psychiatric diagnosis of alcohol abuse or dependence) effects of alcohol use on violence were analyzed. Logistic regression models were used to examine the relationship of acute and chronic alcohol effects to incarceration for a violent offense and arrest for a violent offense, with demographic and criminal history factors controlled. The acute effects of alcohol were found to be significantly associated with incarceration for a violent offense, but the net explanatory capacity of acute alcohol effects was not large. Chronic alcohol effects were not significantly associated with incarceration for a violent offense or arrest for a violent offense in the previous year. The findings were interpreted as being consistent with the hypothesis that alcohol effects violence directly, acting through the acute effects of use, rather than indirectly through the effects of underlying or mediating factors.