Optimal Design of Experiments Evaluating Main and Mediation Effects

• In Event: Hilton Roundtable Session 1
  In Roundtable Session: Methodological Advances in Experimental Design

Abstract

The investigation of main and mediation effects is key to understand whether interventions work or not and the mechanisms of interventions. Literature has rarely provided guidance on how to design efficient and well-powered experiments aim to assess both main and mediation effects. In this study, we developed an optimal design framework that identifies the optimal sample allocation, achieving the maximum statistical power to detect mediation and main effects under a fixed budget. The results show that the framework often can identify more efficient designs than those frequently used (e.g., balanced designs). To facilitate end-user calculations, we have also implemented these formulas in the R package XXX.

Authors

• Zuchao Shen, University of Georgia

• Benjamin Kelcey, University of Cincinnati

• Walter L. Leite, University of Florida

• Wujun Sun, Capital Normal University