The study expected more aggression in those cases in which the aggressor ascribed dissimilar attitudes to the victims than in those in which the victims showed similarity to the aggressor. Also, more aggression was expected in a situation in which aggression yielded external rewards than in a situation that did not. Subjects included 37 Amsterdam policemen who averaged 21 years of age and who had recently completed their initial training. Subject profiles were constructed based on the results of a questionnaire covering authoritarian views, self-evaluation, issues of current importance to policemen, and political and ideological backgrounds. In the experiments, a policeman acted as the aggressor, and a disguised staff member played the victim. In each case, the victim's character profile was designed to be either similar or dissimilar to that of the aggressor. The pair were placed on either side of a partition, and the policeman was asked to evaluate the victim's judgment regarding distance in different traffic situations. The aggressor was then requested to generate beeps to the victim's headphone so as to disturb the victim's judgment; aggressors were told that certain loud beeps could damage the victim's hearing. The resulting data showed a strong effect of the similarity variable, with significantly more aggression being shown to a dissimilar partner. However, the offer of an external reward (money) did not lead to a clear increase of aggression. The relevance of these data to the problem of controlling violence in society is considered. Related studies are reviewed. Data tables and a 44-item reference list are included.