Design and Analysis of Complex Multisite Partially Nested Randomized Trials

• In Event: Experiment Designs of Educational Research Studies

Abstract

Our purpose of this study is to develop statistical models that map out the type of complex partially nested multisite designs, outline estimation methods, and develop closed-form expressions to track statistical power and optimal sampling allocation when testing the main treatment effect and its variability across sites. These works allow researchers to probe the capacity of potential designs incorporating partially nested data structure and determine requisite sample sizes, provide insight into constructing their analytical model and conducting statistical analysis.

Authors

• Yanli Xie, University of Cincinnati

• Benjamin Kelcey, University of Cincinnati