Generalizability of Multiple Measures of Treatment Integrity: An Empirical Replication

Abstract

Treatment integrity is essential for the implementation of interventions in schools as it determines the accuracy or consistency with which different components of a treatment are implemented. There are no current standards regarding the best practices in treatment integrity measurement; however, higher integrity is associated with enhanced student outcomes. At present, there is no database providing information on treatment integrity for practitioners, researchers, and policy-makers to reference for choosing an appropriate level of treatment integrity needed for certain interventions for certain problems. Consequently, there is a need to establish convergent validity among different methods of treatment integrity measurement using multiple evidence-based interventions in order to guide best practices. The current study attempted to replicate and expand the finding that the direct observation method yields the most reliable treatment integrity data the most quickly, followed by self-report, when using an evidence-based intervention (Gresham, Dart, & Collins, 2017). For this study, researchers empirically replicated the methods of Gresham and colleagues’ work to examine two of the three measures of treatment integrity, direct observation and self-report, for six teachers’ implementation of the response cards intervention.