Adolescents encounter a myriad of risk behaviors, ranging from substance use to violence to sexual activity, and the public often views peers as the culprit for a child's participation in these harmful activities. However, recent reviews suggest that a child's peers play an important role in the uptake of both positive and harmful activities. While parents, teachers, and the media provide children with information about risk behaviors, peers also play a crucial role in behavior development by shaping normative beliefs and the perceived validity of information regarding risk behaviors. In essence, peer norms determine whether a behavior is 'hip,' safe, or desirable. Early involvement with risk activities can lead adolescents to develop lifelong behavior patterns, which may adversely affect their mental and physical health, as well as their family's well being. Yet, we do not have a clear understanding of the peer influence process. Specifically, we do not know the most effective way to measure group influence; we do not know whether a child's own behavior stage effects his susceptibility to influence; we do not clearly know whether a best friend is more influential than other friends; and we do not know how a child's self-esteem or family may effect the influence process. This dissertation seeks to answer these questions by examining peer influence across a variety of risk behaviors: smoking cigarettes, drinking alcohol, using marijuana, chewing tobacco, sexual debut and condom use. The project will use the Add Health longitudinal network data, which contains information on both the subjects and their nominated friends. The adolescent's time two risk behavior will be predicted with logistic regressions which control for the subject's pre-existing behavior, and include their friends' behaviors and moderating variables, such as family communication patterns, self-esteem, and friendship reciprocity.
