Power Analysis for Multisite Randomized Cost Effectiveness Trials With Site Fixed Effects

• In Event: Advancing Complex Multilevel Modeling Techniques Through Simulation

Abstract

Multisite randomized controlled trials have been widely used in education to evaluate the causal effects of educational interventions. Recently, education researchers started to incorporate a cost study into their multisite studies to evaluate the cost-effectiveness of an intervention. Prior studies have introduced statistical power analysis for multisite randomized cost-effectiveness trials (MRCETs) using random effect models (e.g., multilevel models); however, fixed-effect models (e.g., OLS regression with school dummy variables) are widely used in education when analyzing multisite studies. This study is designed to address the limitations of power analysis for MRCETs by extending prior work to accommodate models with site fixed effects for two- and three-level MRCETs. We also demonstrate power computations using a free and user-friendly tool we developed.

Authors

• Wei Li, University of Florida

• Nianbo Dong, University of North Carolina - Chapel Hill

• Rebecca A. Maynard, University of Pennsylvania

• Benjamin Kelcey, University of Cincinnati

• Jessaca K. Spybrook, Western Michigan University

• Yue Xu, University of Florida