A Bias-Corrected Method for Estimating the Proportion of True Null Hypotheses

• Activity Number: 360
• Type: Contributed
• Date/Time: Tuesday, August 3, 2010 : 10:30 AM to 12:20 PM
• Sponsor: Biometrics Section
• Author(s): Dexiang Gao*+ and Tiejun Tong
• Companies: The Children's Hospital - Denver and University of Colorado, Boulder
• Keywords: Effect size ; False-null p-value ; Microarray data ; Multiple testing ; Probability density function ; Upper tail probability

Abstract

We consider the problem of estimating the proportion of the true null hypotheses, when testing on a large number of hypotheses. Besides its own interest to know how many hypotheses are true nulls, an accurate estimate of this parameter is key for many multiple testing procedures. It is known that most existing methods for estimating the proportion of true null in the literature are conservative. In this paper, we propose a bias-corrected method by incorporating the distribution of false nulls. Through both theoretical justification and simulation studies, we show that the proposed estimator may substantially decrease the bias than the existing methods and thus improve the overall performance.