Group density (the extent of anticipated or actual interaction between group members independent of the subject) may be a situational antecedent of paranoid behavior. Members of a high density group can discuss others in their absence, and the potential for criticism and restricted privacy is greater than in a low density group where fewer communication channels exist. To test the prediction that under evaluative circumstances, a high density group would be more likely to promote paranoid behavior than a low density group, male subjects were tested in 6 groups of 10 in high or low density conditions. Subjects were told that evaluations of each person's self-disclosures would be used to exclude two subjects from a group problem. High density subjects expected other group members to discuss and evaluate them in their absence, while low density subjects expected to be present for all discussions between group members. Prior to the expected evaluations, measures associated with paranoid behavior (defensiveness, anxiety, self-disclosure, and projection) were administered. The results indicated a significant overall effect for density. High density subjects scored higher in anxiety, projection, and guardedness concerning self-disclosure than low density subjects. Differences in defensiveness were nonsignificant. (Author/PAS)