|
|||||
 |
|
|
|
|
|
|
|
Adaptive tests
for mixed paired and two-sample designs
Eli Sun, Thomas Yacovone and Scott Richter |
|
Vol. 19 (2026), No. 3, 393–404
|
Abstract |
|
Adaptive procedures using and Wilcoxon-based tests for testing for treatment difference in mixed paired and two-sample designs are studied. Previous studies have found that -based tests perform better for normal data, while Wilcoxon-based tests perform better for nonnormal data. The proposed adaptive procedures use tail index measures to distinguish between normal and nonnormal data and select the more powerful test. A simulation study is conducted to estimate the power and type I error rate of the proposed adaptive tests. The results suggest that the adaptive tests provide an objective approach to choose the more appropriate test. |
Keywords
paired design, missing data, adaptive test, tail weight
|
Mathematical Subject Classification
Primary: 62-08
|
Milestones
Received: 19 October 2023
Revised: 5 November 2024
Accepted: 22 January 2025
Published: 12 June 2026
Communicated by Javier Rojo
|
Authors |
| Eli Sun | |
| Department of Mathematics University of Rochester Rochester, NY United States |
|
| Thomas Yacovone | |
| Department of Mathematics and
Computer Science College of the Holy Cross Worcester, MA United States |
|
| Scott Richter | |
| Department of Mathematics and
Statistics University of North Carolina at Greensboro Greensboro, NC United States |
|
| © 2026 MSP (Mathematical Sciences Publishers). |
