This study informs community science, and seeks to narrow the research-to-practice gap, by examining how the interpersonal networks within a setting influence individuals' use of interventions. More specifically, it explores the role of two network mechanisms-cohesion and structural similarity-in urban elementary school teachers' use of interventions designed to improve academic and behavioral outcomes for students. Lagged regression models examine how position in advice giving networks influenced weekly use of the daily report card and peer assisted learning by kindergarten through fourth grade teachers in three schools. Results indicate that intervention use spreads among teachers with similar patterns of advice-giving relationships (i.e., via structural similarity), rather than from teachers who are sources of advice (i.e., via cohesion). These results are consistent with findings in other settings, and suggest that researchers wishing to increase the use of an intervention should select change agents based on their patterns of their relationships, rather than on their direct connections.