Publish or Perish: The Bias of Positive Findings

Less than 40% of scientific conference presentations lead to publications in peer-reviewed journals, with negative or null results being significantly underrepresented. The trend of publication bias has intensified, increasing by 22% from 1990 to 2007, and varies by discipline and region. This bias has tangible consequences, as demonstrated in clinical trials of antidepressants where nearly one-third remain unpublished, predominantly those with negative outcomes. Consequently, published trials misleadingly suggest a 94% positive result rate, while official evaluations reveal only 51% are genuinely positive. Such selective reporting distorts perceptions of drug effectiveness and misguides meta-analyses, which often incorporate predominantly positive studies.