Related Experiment Videos
Publication bias: the problem that won't go away
1Department of Epidemiology and Preventive Medicine, University of Maryland School of Medicine, Baltimore, MD 21201.
Annals of the New York Academy of Sciences
|December 31, 1993
Summary
Publication bias favors significant study findings, threatening medical conclusions. Investigator decisions, not editorial ones, cause nonpublication, highlighting the need for trial registration systems.
Area of Science:
- Medical research methodology
- Scientific publishing ethics
Background:
- The validity of medical conclusions relies on published scientific literature.
- Publication bias, where study results influence publication decisions, compromises this validity.
Purpose of the Study:
- To investigate the association between study result significance and publication status.
- To identify the reasons for nonpublication of initiated studies.
Main Methods:
- Meta-analysis of data from four prospective investigations.
- Tracking 997 initiated studies for results, publication status, and reasons for nonpublication.
- Analysis restricted to controlled trials and randomized trials.
Main Results:
- A positive association exists between "significant" study results and publication (OR = 2.88).
- This association is stronger in controlled trials (OR = 6.15) and randomized trials (OR = 8.72).
- Nonpublication was consistently investigator-based, not editor-based.
Conclusions:
- Study results significantly influence publication, indicating publication bias.
- Investigator decisions to suppress results contribute to this bias.
- Registration systems for clinical trials are crucial to mitigate future publication bias.