The bystanders selected as study subjects were all college students who were unaware that the thefts were simulations. The general procedure was similar for all the studies. The subjects were relaxing, usually reading, and seated by themselves in a corridor of a classroom building. One experimenter, who served as the potential victim, approached the bystander subject and secured the bystander's verbal commitment to watch some personal belongings. The experimenter then set several objects near the bystander. Objects ranged from items such as folders which had low value to items such as an electronic calculator which had high value. The first experimenter then moved to a distant location to record the subject's responses as a second experimenter, who served as the thief, approached, paused to examine the articles, and then conspicuously took them. The study's dependent variable was the percentage of subjects who intervened to stop the theft. Subjects were also administered an open-ended questionnaire regarding their awareness of the theft, their reaction, and reasons for their action or inaction. Results showed that a high percentage of females intervened regardless of the amount of potential loss, whereas males generally helped only when the harm to the victim was high. Female victims received significantly more aid while the thief's sex had no effect. Results supported the usefulness of an interactionist approach to the question of how individual variables and situational variables affect prosocial action. Tables, footnotes, and 22 references are provided. (Author abstract modified)