Adolescence is a unique time during the human lifespan, where children develop into reproductively competent adults, and many behavioral and psychological changes develop. The goal of this dissertation is to examine the underlying functions related to the emergence of risky adolescent "behaviors," broadly defined to include both physical risks (e.g., drinking or physical fighting) and psychological risks (e.g., anxious or depressive symptoms). In this paper I (1) present a manuscript reviewing and synthesizing the relevant evolutionary theories, which bears testable hypotheses regarding sex differences in risky adolescent behavior, and (2) empirically assess these predictions in two separate but theoretically and conceptually related studies that test for theoretically-based hypothesized relations: one specific to males, one to females. Specifically, I test the proposition that males should be more likely to engage in physically risky behavior than females, and that females should be more likely to engage in affective risks (assessed as sensitivity to social evaluation). These hypotheses are generally supported, but shed light on important environmental factors that are associated with the prevalence of risk-taking, including the degree of closeness within the family, timing of pubertal development, and social status. In general, an early pubertal development and low social status are associated with risk-taking, while closeness to family buffers against the likelihood of engaging in later risky behaviors.