the tendency of journals to publish positive, significant results while rejecting null results — meant that the published literature was a selected sample biased toward false positives. If ten research teams independently study a hypothesis, and by chance one finds p < 0.05 while nine find null resu