Evaluating Population-Level Implementations of Evidence-Based Programming: PAX Good Behavior Game and Youth Crime
Abstract
This study analyzes whether the implementation of the PAX Good Behavior Game in elementary schools was associated with lower levels of school-based law enforcement referrals and arrests. It utilizes publicly available administrative data in a quasi-experimental design. School-based law enforcement referrals and arrests in zip codes in which the PAX Good Behavior Game was implemented in schools as part of a statewide population-level public health initiative were compared to those of otherwise similar comparison schools in neighboring zip codes in which the intervention was not yet implemented. For the two geographic regions examined in this study, one region recorded a significant decrease in law enforcement referrals in intervention zip codes compared to comparison zip codes. Neither region recorded significant changes in school arrests among zip codes. The modeling used to detect these changes has meaningful implications for evaluating population-level implementation.