Evidence-based interventions are being disseminated broadly in schools across the U.S., but the implementation levels achieved in community settings vary considerably. The current study examined the extent to which teacher and school factors were associated with implementation dosage and quality of the PAX Good Behavior Game (PAX GBG), a universal classroom-based preventive intervention designed to improve student social-emotional competence and behavior. Specifically, dosage (i.e., number of games and duration of games) across the school year and quality (i.e., how well the game is delivered) of PAX GBG implementation across four time points in a school year were examined. Hierarchical linear modeling was used to examine the association between teacher-level factors (e.g., demographics, self-reports of personal resources, attitudes toward the intervention, and workplace perceptions) and longitudinal implementation data. We also accounted for school-level factors, including demographic characteristics of the students and ratings of the schools' organizational health. Findings indicated that only a few teacher-level factors were significantly related to variation in implementation, whereby perceptions (e.g., fit with teaching style, emotional exhaustion) were generally related to dosage, whereas demographic factors (e.g., teachers' age) were related to quality. These findings highlight the importance of school contextual and proximal teacher factors on the implementation of classroom-based programs. [This paper was published in "Prevention Science" v16 n8 p1064-1074 2015.]