To attain academic goals, school personnel must effectively manage 20% of students who engage in the disruptive behaviors that interrupt instruction, create teacher stress, and contribute to poor student outcomes. Without effective strategies, school personnel often respond to disruptive students with ineffective authoritarian tactics, exclusionary policies, and special education referrals. However, federal policies aim to improve student outcomes and reduce special education referrals. To achieve these goals, schools are integrating universal, selective, and indicated practices and programs in tiered response models. Though many effective universal programs exist, only a few selective programs are available. The few available and widely-used selective strategies are rooted in behaviorism, managed by school personnel, and do not integrate scientifically-based efforts that improve self-management outcomes for students. The purpose of the dissertation was to study the effectiveness of STARS, a manualized self-monitoring program designed to be a selective strategy within a tiered response model. The study relied upon a randomized trial with 108 disruptive students across 42 classrooms and 7 schools. Results indicated STARS was feasible, acceptable, and related to improvements in behavior, social competence, authority acceptance, and student-teacher relations. Mediation models confirm direct instruction through STARS in social competencies caused improvements in student behaviors.