Using multilevel modeling to examine the effects of multitiered interventions

Abstract

AbstractData collected in school settings are inherently hierarchical. At the same time, it is becoming increasingly common for interventions to be implemented within the context of a similar multitiered intervention framework where different interventions are provided at different levels of the hierarchy, often simultaneously. This prevalence of hierarchical structures in educational intervention research makes the use of multilevel models (MMs) an appealing methodology for the study of multitiered interventions. We provide an overview of (1) the hierarchical nature of educational intervention research, (2) relevant practical issues in the design of multitiered intervention studies, and (3) the use of MMs for providing a descriptive characterization of multitiered intervention effects. Findings from a hypothetical data example demonstrate that MMs provide noticeably more accurate estimates of effects than traditional single‐level analyses and clarify the advantages of MMs for examining data analysis issues relevant to multitiered intervention studies. Finally, the article provides recommendations for the use of MMs in future intervention studies, especially within the context of multitiered intervention models and response to intervention frameworks. © 2007 Wiley Periodicals, Inc. Psychol Schs 44: 503–513, 2007.