The study determined that 43.3 percent of the offenders were in the "perpetrator-victim" category as defined by the researchers. This means that while incarcerated they have been both perpetrators and victims of aggression; 36.7 percent were perpetrators of aggression against other inmates but not victims of aggression; 6.7 percent were victims of aggression from other inmates without being perpetrators of aggression; and 13.3 percent of the men had not been either victims or perpetrators of aggression while incarcerated ("uninvolved"). The "victim" and "uninvolved" groups had attitudes that were significantly more supportive of victims or victimization than were the attitudes of "perpetrator-victims" and "perpetrators." There were no significant differences among the groups regarding moral reasoning, and there was no link between provictim attitudes and moral reasoning. Possible explanations for these findings are discussed. Of the 60 participants, 33.4 percent were serving a sentence for offenses against a person, 50 percent were convicted for property offenses, 8.3 percent were sentenced for drug offenses, and 8.3 percent were imprisoned for other offenses. All of the men were administered the following instruments: The Direct and Indirect Prisoner Behavior Checklist, which measures interpersonal aggression or victimization; the provictim scale (Rigby and Slee, 1991), which assesses attitudes toward victims; and the Sociomoral Reflection Measure-Short Form, which measures sociomoral reasoning by requiring respondents to justify the importance of the moral values of contract and truth, affiliation, life, property and law, and legal justice. 2 tables and 45 references