Despite the emphasis we placed on designing and executing a rigorous school-randomized controlled trial, several limitations of this study should be mentioned. Unlike Frey et al. (2005), we did not use observational measures of playground behaviors despite the fact that their study found intervention effects in several bullying-related behaviors with this method. The use of students' self-report of bullying and bullying-related behaviors has been brought into question (Frey et al., 2009; Jenson & Dieterich, 2007). Although other sources of data (e.g., firsthand playground observations and direct observation of teachers in classrooms) could provide a unique perspective to program implementation and effectiveness, these are expensive and difficult to implement in large-scale school-randomized trials. …show more content…
As the STR program is designed to be administered to elementary students across three consecutive grades, longterm follow-up of students would allow for a better assessment of program sustainability. In addition, our assessment of program fidelity, as implemented in the classroom, was measured using online checklist data reported by teachers. Although analyses of these data indicated fidelity to the classroom curricula, observation of teachers administering the program in classrooms by trained observers could have provided an alternative assessment of classroom program fidelity, independent from teachers' self-report. We note that effect sizes presented in this study are relatively small and that the generalization of study findings also may be limited by selection of schools with at least 80% agreement from teachers to participate in the trial and the unique geographical and demographic features of the school sample recruitment area (i.e., northern-central