Commentary on Livingston (2011): Alcohol outlets and domestic violence -acute effects and the social ecology of neighborhoods may both contribute to the relationshipa dd_3396 926..927 'Location, location, location' are the three most important considerations in finding a place to live. This reflects the knowledge that the physical and social environment matter, that they both constrain and facilitate behavior. [6]. The study by Livingston [7] in the current issue adds to this literature, replicates the findings of linking alcohol outlets to violence and extends this literature in some important ways. Most importantly, it provides longitudinal evidence that changes in the number of alcohol outlets are associated with changes in police reports of 'family incidents'.As with other studies of ecological data, the results may have been produced by several different processes. One explanation suggested by Livingston is that increases in alcohol outlets led to increases in heavy drinking episodes, and through the acute effects of alcohol increased the occurrence or severity of domestic disputes to the extent that they are reported to the police. There is a chain of evidence supporting this view. Experimental studies indicate that intoxication leads to increases in experimentally measured aggressive behavior, mainly in male dyads [8]. Other experiments show that alcohol increases the negativity of marital conflicts [9] and verbal expressions of aggressive intentions among maritally aggressive men [10]. Studies of partner violence episodes indicate that episodes are more severe when the man has been drinking [11,12]. Thus, factors that lead to changes in excessive use, either in individuals or in the community, should, through these acute effects, lead to changes in domestic violence.Notwithstanding my belief in the acute effects model [13], it seems unlikely that acute effects alone can account for the results. Perhaps one-third of domestic violence episodes in the community [14], but closer to 40% of police calls for domestic violence [15] involve alcohol. With a conservative estimate that 50% of the episodes did not involve alcohol, the observed rate of 4.76 events would then consist of 2.38 alcohol-involved events. If the estimated increase of 1.36 episodes consists of entirely alcohol-related episodes, this means that a one-packaged outlet increase per 1000 individuals led to a 56% increase in alcohol-related episodes, leading to the question: 'Can the increase in heavy drinking that occurs with an increase of one packaged outlet account for this large of an increase in alcohol-related domestic violence episodes'?. I think the answer is 'probably not', which should lead us to think about other processes. Alternatively, if the estimated increase of 1.36 episodes per 1000 people consists of many episodes that do not involve alcohol, the conclusion is the same; we must think about other processes in addition to acute alcohol effects.The key problem in specifying other explanations for the results is that the process...