Date of this Version
An adolescent’s peer group has been theorized to influence the development of psychopathology. However, little research has examined the adolescent peer group using information obtained directly from peers in a longitudinal framework. Research has also been limited on peer group influence on the development of internalizing disorders. The study used Social Network Analysis to examine self-reported anxiety, depression, aggression, and delinquency in the fall and spring of one school year for students in a rural high school. In addition to examining the effect of the peer group on individual reports of psychopathology, the strength of this relation was compared to that of the adolescent’s closest friend. Potential moderators (peer group density, grade, and gender) of the relation between the peer group and individual psychopathology were examined. Results suggested that how the peer group variable is constructed affected the findings. When the peer group variable was constructed from reciprocated peer nominations, the peer group level of anxiety or depression predicted later individual changes in these measures. However, when the peer group variable was constructed from all incoming and outgoing nominations, regardless of reciprocation, the level of delinquency reported in the peer group predicted later change in individual delinquency. The peer group’s level of aggression was not related to concurrent or later individual aggression. The adolescent’s closest friend’s level of psychopathology was not related to concurrent or later psychopathology. Peer group density was supported as a moderator of the relation between reciprocated peer group and individual anxiety, such that individuals from less dense peer groups were more influenced by the peer group. Grade was supported as a moderator of the relation between reciprocated peer group and individual level of anxiety, depression, and delinquency, with anxiety and depression showing the expected negative quadratic moderation effect, and delinquency showing an unexpected, positive moderation effect for grade. Gender was not supported as a moderator. Conceptual and methodological implications are discussed with recommendations for clinical practice and policy.