Systematic tendency to publish positive results while suppressing negative or null findings
Publication bias is the systematic tendency of journals, researchers, and funding bodies to favor the publication of studies with positive or statistically significant results over studies with negative, null, or inconclusive findings. This creates a distorted scientific literature that overstates the effectiveness of interventions and understates their risks.
The phenomenon is well-documented. A 2008 analysis of FDA-registered antidepressant trials found that 94% of published trials reported positive results, while only 51% of all registered trials (published and unpublished) were actually positive. The published literature painted a dramatically more favorable picture than the full evidence base. This pattern has been confirmed across multiple therapeutic areas.
Publication bias operates through multiple mechanisms. Journals preferentially accept papers with novel, positive findings. Researchers are more likely to write up and submit positive results. Pharmaceutical companies publish favorable trials while burying unfavorable ones in regulatory filings that are difficult for clinicians to access. The cumulative effect is a scientific literature that systematically misleads — not through any single act of fraud, but through a structural bias that filters out inconvenient truths.
