|
Activity Number:
|
548
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Thursday, August 6, 2009 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Section on Bayesian Statistical Science
|
| Abstract - #303903 |
|
Title:
|
Semiparametric Bayesian Approach for Testing Noninferiority Using Relative Risk and Odds Ratio for Binary Data
|
|
Author(s):
|
Muhtarjan Osman*+ and Sujit Ghosh
|
|
Companies:
|
North Carolina State University and North Carolina State University
|
|
Address:
|
, , ,
|
|
Keywords:
|
Bayes factor ; Bernstein polynomial ; mixture prior ; non-inferiority ; odds ratio ; relative risk
|
|
Abstract:
|
A semi-parametric testing approach using Bayes factor is developed for non-inferiority in the setting of two-sample binary data. Mixture prior is constructed based on Berstein polynomial for two binomial proportions without putting any assumption the form of the density function. The proposed prior also assigns approximately equal probability to the two competing hypotheses to avoid bias. The critical value of the Bayes factor is determined to approximately control the frequentist type I error rate to a desired level. Through numerous simulation studies it is demonstrated that the power of the Bayesian procedures are substantially higher than that of the popularly used frequentist procedure. This in turn indicates that the proposed Bayesian procedures could lead to substantial decrease in sample size for desired level of the power.
|
