In this analysis of extant data from 3,011 schools implementing school-wide positive behavioral interventions and supports (SWPBIS) across multiple years, we assessed the predictive power of various school characteristics and speed of initial implementation on sustained fidelity of implementation of SWPBIS at 1, 3, and 5 years. In addition, we partitioned variance in whether schools sustained SWPBIS at the school, district, and state levels. Results showed that the largest differences in fidelity were at the state level. In addition, school characteristics (particularly grade levels served) and speed of initial implementation were significant but small predictors of sustained implementation. Results are discussed in terms of effective supports at the state and district levels, as well as addressing types of schools at greater risk for abandonment, such as high schools. [This paper was published in "Journal of Positive Behavior Interventions" (EJ1113868).]