A Web-Based Application for Constrained Optimal Sample Allocation in Design of Multilevel Randomized Experiments

• In Event: Marriott Roundtable Session 41
  In Roundtable Session: Methodological Design and Illustrations

Abstract

There are several tools to design multilevel randomized experiments. Majority of these tools center their framework on statistical power and minimum detectable effect size calculation, whereas relatively few has focused on sample size calculation. Sample size calculation in multilevel randomized experiments can be complicated due to multilevel structure, having a fixed budget, and other limitations with sample sizes. There exists a few user-friendly tools to help researcher in this endeavor (CRT-Power and PowerUp!), however, a web-application that can potentially reach many researcher has been a missing component. In this study, we propose a front-end web application to PowerUpR, an R package developed from PowerUp! and PowerUp!-Moderator, and allows constrained optimal sample allocation for average treatment and moderator effects.

Authors

• Metin Bulus, University of Missouri - Columbia

• Nianbo Dong, University of Missouri - Columbia

• Benjamin Kelcey, University of Cincinnati

• Jessaca K. Spybrook, Western Michigan University