In many scientific fields, a p value less than 0.05 is considered significant, pointing to real difference in reality, whereas anything above it cannot be accepted (i.e. null hypothesis cannot be rejected, which means your results cannot be accepted for publication), From a perspective of pure statistics, this threshold of p value is completely arbitrary. Why have the scientific community settled on this value, and is willing to reject the null hypothesis only for for anything with a probability less than 5%?

  • 1
    $\begingroup$ Should be migrated to statistics $\endgroup$ – Peter Jansson Sep 22 '13 at 15:08
  • 4
    $\begingroup$ In particle physics they use 0.0000003. $\endgroup$ – Noah Snyder Sep 22 '13 at 16:11