The findings of this study confirm that exposure to deviant peers is an important factor in the promotion of delinquent behavior. It also shows that a mechanism of "imbalance" in delinquency between close friends predicted within-individual change in delinquency. Since it is rare for two individuals to bring the same level and content of delinquent attitudes/behaviors to a friendship, this imbalance has an interactive effect on the individuals that can either increase or decrease delinquent behavior, depending on the dynamics and personalities of the individuals involved. The context for peer interactions also influences exposure to the varying delinquent behaviors of interacting peers. In unstructured and unsupervised interactions, delinquent behavior is likely to increase, regardless of whether the peers have an established pattern of delinquency. These findings suggest multiple pathways by which particular individuals in interactions with one another in various contexts will commit different types of delinquent acts that may or may not become stable behaviors. Further research is needed in order to identify the precise factors in the pathways of delinquency that evolve from peer interactions. This study used information from the AddHealth data, which are longitudinal and have consistent deviance measures across both waves, which affords the opportunity to assess within-individual change in delinquency over time. Measures of peer deviance allow for the construction of a relative peer deviance that measures deviance imbalance. 5 tables, 1 figure, and 99 references