Testing the hypothesis that group-centered leadership leads to less problem behavior among emotionally disturbed adolescents, 20 male and female adolescents aged 14-18 (from Karitas Community in Crystal Springs, Mississippi) were divided into leader-centered and group-centered groups and exposed to 3 weekend camping experiences. Assessments were made via Bovard Scale of Interpersonal Affect (level of interpersonal liking for other group members); Wagman Group as a Whole Scale (individual's degree of like or dislike for the group as a whole); and Devereux Adolescent Behavior Rating Scale (amount and type of problem behavior displayed by participants). Statistical analysis involved the: Kruskal-Wallis One Way Analysis of Variance; Mann-Whitney U Test; and Spearmann Rank Order Correlation Coefficient. Results indicated: participation in a camping program with either group- or leader-centered leadership did not improve emotionally disturbed adolescents' level of interpersonal affect for other group members; group-centered leadership led to greater increase in liking for the group as a whole than did leader-centered leadership; the hypothesis was not unconditionally supported, although the only positive behavior changes that occurred and the one positive behavior change that continued past the first camping trip occurred in the group-centered leadership group. (JC)