How Individual and School Aggregate Baseline Behavior Levels Moderate Response to a Primary Level Behavior Intervention

In our previous research (Benner, Nelson, Sanders, & Ralston, 2012), elementary schools were randomly assigned to either a primary-level behavior intervention directed at externalizing behavior (treatment, n = 7 schools) or to business-as-usual condition (control, n = 6 schools). A screening procedure was used to identify K through 3-grade students who exhibited externalizing behavior. The results showed that treatment students (n = 44) exhibited lower levels of problem behavior (ES = -0.99) and higher rates of on-task behavior (ES = 0.61) compared to controls (n = 26). The results also suggested that individual baseline behavior and school context as measured by free and reduced-priced meal rates moderated the findings. Given this latter finding, our interest in the current study was to test whether individual and school aggregate baseline behavior levels moderate students' response to two years of the behavior intervention. Schools originally assigned to treatment continued to implement the intervention during the subsequent academic year; schools in the control condition continued with their business-as-usual behavior management approaches. During the winter of the second year, students who remained in their schools were re-assessed on their behavior (n = 35 and n = 19 in treatment and control conditions, respectively). Analyses involving multilevel context modeling showed that students with relatively high initial baseline behavior levels (pretreatment) were more responsive to the intervention if they were in schools with higher school aggregate baseline behavior levels. Conversely, students tended to be less responsive to intervention if they were in schools with lower school aggregate baseline behavior levels. Limitations and practical implications are discussed. (Contains 2 tables and 1 figure.)