Researchers want to redefine the threshold for scientific discovery from 0.05 to 0.005 — Quartz In 1925, Fisher published Statistical Methods for Research Workers, whete he explained the concept of statistical significance. Somewhat arbitrarily, Fisher chose to define statistical significance as a difference that had less than a .05 probability of occurring by random chance (in technical terms, this is called a p-value).  Fast forward more than a century later,