The systematic tendency for studies with positive, statistically significant results to be published more readily than studies with negative, null, or inconclusive findings. This creates a distorted evidence base where the published literature overrepresents positive outcomes.

Publication bias is particularly acute for compounds with small evidence bases. If ten groups study a peptide and three find positive effects while seven find nothing, the three positive studies get published and the seven null studies sit in file drawers. The result: the published evidence makes the compound look effective when the full evidence does not. This is why systematic reviews search for unpublished data, why trial registries like ClinicalTrials.gov exist, and why the number of published studies on a compound is not the same as the strength of its evidence.