Aggression is inherently social. Evolutionary theories, for instance, suggest that the peer group within which an aggressor is embedded is of central importance to the use of aggression. However, there is disagreement in the field with regard to understanding precisely how aggression and peer relationships should relate. As such, in a series of three empirical studies, my dissertation takes a relational approach and addresses some of the inconsistencies present in the extant literature. In Study 1, I examined how qualities of youth's close friendships contributed to the use of aggression, both concurrently and over time. I found that youth with large friendship networks were more aggressive, whereas those with highly interconnected friendship network decreased in aggression over time. Using a dyadic mediation model, the second study considered the precursors to aggressors' friendships with peers. Specifically, I explored aggressive youth's interactions with unfamiliar peers and assessed how the interactions that unfold affected the quality of the relationship. I found that dyads who were highly discrepant in their tendencies toward aggression failed to collaborate well with one another, and this led to less positive perceptions of one another. Whereas the first two studies concerned aggressors' relationships with their friends (Study 1) and acquaintances (Study 2), Study 3 focused on a different type of relationship – the relationship between an aggressor and his or her victim(s). In the third study, I explored how power dynamics operate within an aggressor-victim dyad and assessed whether differences in the balance of power between the aggressor and victim affected the strength of their relationship. I found that more aggressor-victim dyads were characterized by a relative balance than imbalance in power, and that power balanced dyads had stronger and more sustained aggressor-victim relationships. By taking a relational approach to the study of aggression, this dissertation has advanced extant work in the field. That is, these findings move away from the simplification and aggregation of relational constructs (e.g., relationships, friendships), and instead consider the nuances of specific types of relationships or interactions with specific peers, allowing for a better understanding of the relational nature of aggression.