Subjects taking part in the study were 153 college students enrolled in a behavioral science course. Students ranged in age from 17 to 50, with a mean age of 23. Students were approached after lectures and asked to volunteer for research in jury decisionmaking; the majority of those present volunteered. The subjects read evidence presented at two political trials, one a terrorist bombing and the other a politically motivated burglary. In both crimes, the defendant was sometimes identified as a radical leftist and sometimes identified as a member of the radical right. Subjects were asked to render verdicts, to recommend sentences, and to answer questions about their own beliefs. Analysis of resultant data indicates that as objective similarity increased, the probability of finding the defendant guilty decreased in the political burglary situation (a crime involving no loss of life). This effect was not observed in the terrorist bombing situation, which did involve loss of life. In both situations there was a tendency for similarity to lead to a more punitive sentence when the subject was convinced of the defendant's guilt. conclusions drawn from study findings are limited by three factors, including the use of university students as subjects, the lack of group deliberation, and the problem of artificiality, i.e., reading about a trial rather than experiencing it. Tables and 21 references are included in the article. (Author abstract modified)