Contextual Influences on the Implementation of a School-Wide Intervention

Purpose:  School-wide interventions are among the most effective approaches for improving students’ behavioral and academic outcomes. However, scholars have documented consistent challenges with the implementation fidelity of these approaches (Payne et al., 2006). This mixed methods study examines contextual influences on the implementation of a whole school intervention called Responsive Classroom (RC).  In a random control trial and longitudinal studies, RC has shown efficacy in improving social, emotional, literacy and math outcomes for disadvantaged students with behavior problems by building on the assets of teachers to assess and intervene with students in the classroom setting (Rimm-Kaufman et al., 2007).   

Methods:  We conducted ten focus groups with teachers, school social work interns, and school administrators in the fall of 2013 (n=15) and spring of 2014 (n=19), collected observational data from teachers using the Classroom Practices Observation Measure (n=24) at two time points, and administered the Classroom Practices Frequency Survey at the end of the school year (n=19). Both quantitative measures were created specifically for assessing RC implementation and have been found to be reliable and valid (Abry, 2010). Template analysis (King, 2012) was used to analyze the qualitative data and independent samples t-tests were used on the quantitative data. We integrated qualitative themes with quantitative results for a convergent mixed methods analysis.

Results:  Nine codes represent the qualitative findings, indicating contextual influences aligned with three categories described by Durlak and DuPre (2008) in their framework for conceptualizing implementation effects: 1) provider characteristics such as beliefs about behavior change and management; 2) organizational capacity such as principal buy-in, and 3) the delivery support system such as training and technical assistance.  Quantitative findings supported key themes from the focus groups.  For example, the influence of organizational capacity was evident in both types of data.  Focus group results indicated that principal buy-in was stronger among the elementary grades than the middle school grades (each grade level had its own principal). Survey results mirrored this qualitative findings and demonstrated that teachers in the elementary grades (M = 4.9, SD = 1.18) used the RC strategies more often than those in the middle schools grades (M = 3.0, SD = .99) [t (15) = 2.91, p = .01]. These patterns were also evident in the classroom observation data, with elementary teachers mean use of the intervention (M = 2.09, SD = .60) greater than middle school teachers (M = 1.2, SD = .43) [t(22) = 4.01, p = .001].

Conclusions:  There is growing consensus among school social work scholars that the most promising service models reflect an ecological orientation that targets the school environment (Kelly, Raines, Stone, & Frey, 2010). Our findings indicate a need for additional attention and resources to support the translation of school-wide interventions into public schools. In particular, our results suggest that when implementing school-wide interventions there is need for a differentiated model of professional development by grade level that also addresses teacher beliefs about behavior management. Results also indicate the importance of cultivating strong principal buy-in prior to program implementation.