Partially Nested Multisite Cluster-Randomized Trials: Estimation, Inference, and Design Randomized Trials

• In Event: Estimation Methods for Multilevel Models

Abstract

Our study is to advance experimental design in partially nested studies by developing principles and strategies for multisite partially nested randomized trials. We begin by describing the nature and utility of multisite partially nested randomized trials and detailing the corresponding statistical models. Then, we develop principles of estimation and inference for the design, and expressions to track the statistical power with which the design can detect main effects and effect heterogeneity across sites. Subsequently, we discuss design implications and strategies that emerge from these results. To conclude, we compare our results with those from a commonplace model that ignores the partially nested structure. The results provided a set of principles and formulas to guide researchers in the design and analysis.

Authors

• Yanli Xie, Florida State University

• Benjamin Kelcey, University of Cincinnati