Statistics and Its Interface

Volume 5 (2012)

Number 4

A new non-inferiority test based on Bayesian estimation in matched-pairs design

Pages: 443 – 449

DOI: https://dx.doi.org/10.4310/SII.2012.v5.n4.a7

Authors

Weiquan Chen (School of Mathematical Sciences, South China Normal University, Guangzhou, China)

Hua Jin (School of Mathematical Sciences, South China Normal University, Guangzhou, China)

Abstract

Non-inferiority of one treatment to another is a common issue in medical research. In this paper, a new test using an approximate $p$-value based on Bayesian estimation is proposed. Our test is based on only one point of the twodimension nuisance parameter space for accuracy improvement and computational purposes. The sizes and powers of our test are considered. Simulation results suggest that our test can control the type I error rates with reasonable powers while the asymptotic normal test cannot for most cases. In comparison to Sidik’s exact test, our test is much easier to implement.

Keywords

matched-pairs sample, test of non-inferiority, Bayesian estimation, approximate p-value, Type I error rates of tests, powers of tests

2010 Mathematics Subject Classification

60K35

Published 16 November 2012