 The conclusion is that using p equals 0.05 to suggest a discovery will result in being wrong at least 30% of the time, and experiments underpowered will increase this error rate. This conclusion is supported by tree diagrams, simulations, and evaluations of effect size overestimation. To keep false discovery rate below 5%, a 3 sigma rule or p less than or equal to 0.001 should be used, and the word significant should be avoided. This article was authored by David Kahoon.