Standard Article

Publication Bias

  1. Hannah R. Rothstein

Published Online: 14 DEC 2007

DOI: 10.1002/9780471462422.eoct323

Wiley Encyclopedia of Clinical Trials

Wiley Encyclopedia of Clinical Trials

How to Cite

Rothstein, H. R. 2007. Publication Bias. Wiley Encyclopedia of Clinical Trials. 1–11.

Author Information

  1. City University of New York, Baruch College, New York City, New York

Publication History

  1. Published Online: 14 DEC 2007

Abstract

Publication bias is defined as the problem that results from systematic differences between the results of all the completed studies on a topic and the results of the subset of those studies that are published. The consequences of this problem include the danger that readers and reviewers will reach the wrong conclusion about what the evidence shows, leading at times to the use of unsafe or ineffective treatments. The role of meta-analysis in highlighting the existence of publication bias is explained, after which a brief summary of the evidence that publication bias exists in the biomedical trials literature is provided. In addition to “classic” publication bias (bias due to selection on the basis of statistical significance), other potential information suppression mechanisms that represent a threat to the validity of research reviews are explained, including language, citation, availability, cost, outcome, and duplication biases. After this, ways of preventing publication bias are introduced, including universal prospective registration of trials, and design of prospective meta-analyses. The role of thorough information search in reducing the impact of publication bias on meta-analytic results is explained, and suggestions of what to search are made. The major procedures for assessing publication bias are summarized, including file drawer (failsafe N) analysis, the funnel plot, the Begg and Mazumdar test, the Egger test, and the trim and fill method. Differences in the various procedures' specific hypotheses about the distribution of study results as well as their limitations are noted. Finally, results of surveys of the impact of publication bias on health-care meta-analysis are described and their implications discussed.

Keywords:

  • publication bias;
  • information suppression;
  • meta-analysis;
  • trials registry;
  • information retrieval