Sequential Inference on a Measure of Reduction for Two Binomial Proportions

• Activity Number: 644
• Type: Contributed
• Date/Time: Thursday, August 7, 2014 : 10:30 AM to 12:20 PM
• Sponsor: Biometrics Section
• Author(s): Hokwon Cho*+ and Zhou Wang
• Companies: University of Nevada, Las Vegas and University of Nevada, Las Vegas
• Keywords: Measure of reduction ; Modified MLE ; Sequential confidence limts ; Optimal sample sizes ; Asymptotics

Abstract

We propose sequential methods to establish approximate confidence limits of the measure of reduction (MOR) for two independent binomial proportions. The measure is considered based on a modified maximum likelihood estimator. We investigate the desirable properties of the estimator focusing on the unbiasedness, variance and its asymptotic normality. We also find the optimal sample sizes of the interval for the true values of the measure. For numerical study, the Monte Carlo experiment is carried out for the various scenarios with two samples to examine its finite sample behavior.