Trajectories for Coaching Dosage As a Means for Improving Implementation of the Good Behavior Game

Introduction: With implementation of evidence-based interventions typically being poorer in real-world settings than in efficacy trials (Dusenbury et al., 2003; Ringwalt et al., 2003), there is increasing interest among federal agencies, researchers, and policy makers in translational research (SPR, 2008). Although research shows that on-going, rather than one-time or occasional, training is needed in order to promote implementation and sustainability (Fixsen et al., 2003), it is currently unknown how much or how often “on-going” supports are needed in order to promote teacher implementation and thus student outcomes. This paper addresses this gap in the literature by determining specific patterns of coaching support to enhance teachers’ implementation of the Good Behavior Game (GBG; Barrish et al., 1969; Embry et al., 2003). Specifically, data from a one-year randomized controlled trial of the GBG will be used to determine whether coaches provide varying levels support in terms of time spent with teachers. We also examine the specific coaching activities engaged in and determine whether these activities are associated with teacher characteristics or GBG implementation.

Method: Coaching supports were provided to 220 teachers in 18 schools. After each contact with a teacher, coaches recorded the number of minutes spent on activities (e.g., checking in, modeling skills, providing technical assistance and feedback, delivering GBG, and collecting additional data). Coaches began the school year by providing consistent support for all teachers, but then provided more tailored coaching supports based on implementation observations.

Results: Teachers received an average of 9 hours of coaching supports over the course of the school year. The largest proportion of time was spent doing check-ins with teachers followed closely by needs assessments; however, there were differences in the time spent on activities between the initial and tailored phases of the coaching process. For example, during the universal coaching period, coaches spent the most time conducting needs assessments. It is anticipated that distinct coaching patterns will emerge with more intensive coaching, including a greater amount of modeling and feedback, being provided to teachers with lower initial implementation of the GBG. Latent variable modeling will be used to derive distinct patterns of coaching support and examine whether these patterns vary over time. Teacher implementation and perceptions will be modeled as predictors of coaching trajectories.

Discussion: Implications for the dissemination of prevention interventions will be discussed.