Bayesian tests of extra-binomial variability. Statistics in Medicine

uTaipei > College of Science > Department of Mathematic and Computer Science Education > Periodical Articles > Item 987654321/2697

Please use this identifier to cite or link to this item: http://utaipeir.lib.utaipei.edu.tw/dspace/handle/987654321/2697

Title:	Bayesian tests of extra-binomial variability. Statistics in Medicine
Authors:	Hsiao Chuhsing Kate Lee Mei-hsien Kass Robert E 李美賢蕭朱杏
Contributors:	臺北市立教育大學數學資訊教育學系
Date:	2005
Issue Date:	2009-08-04 10:58:40 (UTC+8)
Abstract:	A simple model for extra-Binomial variability is the Beta-Binomial. A complication in testing the Binomial against the Beta-Binomial alternative is that the Binomial lies on the boundary of the Beta-Binomial, which forces modifications to the usual asymptotic arguments. In this paper, we propose a Bayesian test using a pair of approximate Bayes factors, one for the case in which the maximum likelihood estimator (MLE) of the extra-Binomial variability is zero and one for the case in which it is positive. These approximate Bayes factors are easy to compute. We evaluate the operating characteristics of the Bayes factors and find them to be more powerful than the likelihood ratio test. We then apply the method to three data sets, including one in which the issue is whether a logistic regression intercept should be considered a random effect. In each case, our approximate Bayes factors are close to the exact Bayes factors, which may also be computed with additional effort.
Relation:	BIOMED EXPERTS, 24(1), P.49-64.
Appears in Collections:	[Department of Mathematic and Computer Science Education] Periodical Articles

Files in This Item:

There are no files associated with this item.