Improving Power in One-Sided Multiple Testing
Published:Dec 31, 2025 03:26
•1 min read
•ArXiv
Analysis
This paper addresses the problem of conservative p-values in one-sided multiple testing, which leads to a loss of power. The authors propose a method to refine p-values by estimating the null distribution, allowing for improved power without modifying existing multiple testing procedures. This is a practical improvement for researchers using standard multiple testing methods.
Key Takeaways
Reference
“The proposed method substantially improves power when p-values are conservative, while achieving comparable performance to existing methods when p-values are exact.”