Optimal Design of Multisite Experiments Investigating Moderation and Main Effects

• In Event: Statistical Power Considerations in Multilevel Designs

Abstract

Researchers often need to investigate whether interventions work or not (through main effect analysis) and under what conditions they work (through moderation analysis). When designing multisite randomized trials to investigate these effects, researchers must determine the sample size needed to detect both effects while efficiently using resources. This study develops a flexible optimal design framework to identify the optimal sample allocation that achieves the maximum statistical to detect both moderation and main effects under a fixed budget. We present illustrative examples to visualize the optimization and implement the method in the R package XXX.

Authors

• Zuchao Shen, University of Georgia
  Presenting Author

• Benjamin Kelcey, University of Cincinnati
  Non-Presenting Author