search for




 

Bayesian hierarchical model for publication bias
Journal of the Korean Data & Information Science Society 2019;30:1-10
Published online January 31, 2019;  https://doi.org/10.7465/jkdi.2019.30.1.1
© 2019 Korean Data and Information Science Society.

Eun Jin Jang1 · Dal Ho Kim2

1Department of Information Statistics, Andong National University
2Department of Statistics, Kyungpook National University
Correspondence to: Professor, Department of Statistics, Kyungpook National University, Daegu 41566, Korea. E-mail: dalkim@knu.ac.kr
Received December 26, 2018; Revised January 9, 2019; Accepted January 11, 2019.
This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/3.0) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.
Abstract
Meta-analysis is a statistical methodology aimed at quantitatively integrating the results of homogeneous studies on specific research topics. For meta-analysis, statistically significant or clinically positive results are more likely to be published, and the results of meta-analysis using published studies tend to overestimate the effect size, which may result in publication bias. Although regression-based methods are widely used to assess the publication bias in meta-analysis, fewer studies included in the meta-analysis have reduced accuracy. Therefore, this study considers the Bayesian hierarchical model for more accurate estimation when the number of studies is small, and compares based on simulation and real data application. As a result, the greater the degree of publication bias and statistical heterogeneity, the more likely the bias of the estimated effect size to be reduced using the Bayesian hierarchical model.
Keywords : Bayesian hierarchical model, Egger’s test, meta analysis, publication bias.