Classroom- and School-Level Predictors of Implementation in a Multisite RCT to Promote Early Academic Engagement

Introduction: Understanding the sources of variation underlying implementation fidelity contributes to improvements in internal validity (linking program model to outcomes), external validity (i.e., replication), program theory (i.e., how does a program work), program impact, and sustainability. It is becoming increasingly accepted that implementation quality strongly influences whether programs have an impact or are maximally effective. Optimal implementation may be particularly challenging for social–ecological or multicomponent approaches in public school contexts, which often require coordination among various groups within an organization (e.g., teachers and administrators), as well as in terms of acquisition and integration of skills into daily practice. In this paper, we explore predictors of implementation of a widely disseminated SEL program (Second Step©) and Proactive Classroom Management (PCM), in the context of a randomized, controlled trial among a highly diverse sample of early elementary school students (K–3).

Methods and Results: This presentation draws on Year 1 of a randomized, controlled efficacy trial involving 61 elementary schools in Washington and Arizona. Participants were 7,277 students and 324 teachers randomly selected to be part of data collection. Teachers received training in Second Step and a half-day training in PCM. To support the integrity of implementation of the Second Step curriculum and PCM strategies, monthly tips and reminders were developed and disseminated to teachers via email and paper flyers. Data about adherence to and engagement with the Second Step Curriculum were gathered through teacher implementation logs, and data about adherence to PCM strategies were gathered through three observations throughout the year. Teachers were expected to complete 25 lessons, though dosage varied greatly in the study (average delivered = 17, range = 5–25 lessons). Adherence to and engagement among students also varied greatly across classrooms. Data suggest little uptake of the PCM strategies. This paper draws on survey and observational data to (1) describe variability in implementation, and (2) examine the school-level (SES, % minority, amount of school-level programming, administrative attitudes) and classroom-level (% ELL, grade level, observed disruptive behavior) predictors of adherence and engagement.

Discussion: Findings are discussed in regard to classroom and to broad school-level threats to optimal implementation and the need for enhanced supports for PCM. Implementation is discussed in the broader context of cultural diversity.