The following five hypotheses were made: (1) physiological arousal and aggressive thoughts were predicted to be higher in the virtual reality immersion condition than in the virtual reality observation and control conditions; (2) subjects in the virtual reality observation condition were predicted to report more aggressive thoughts than were those in the control condition; (3) hostile feelings were not expected to decrease from baseline to treatment in the virtual reality or observation conditions, as would be predicted via a drivereduction hypothesis; (4) because one aggresses directly in the virtual reality immersion condition, gender differences in aggressive thoughts were not predicted within this condition; and (5) as aggression is more in keeping with male rather than female roles, men were expected to report more aggressive thoughts than were women in the observation condition.