Free Trial

Safari Books Online is a digital library providing on-demand subscription access to thousands of learning resources.

Share this Page URL

Can You Reject the Null Hypothesis When p > 0.05? 251 CAN YOU REJECT THE NULL HYPOTHESIS WHEN p > 0.05? On One Hand Setting = 0.05 provides significant control over the likelihood of a Type I error (rejecting the null hypothesis when there is actually no difference). With this test criterion, any result with p < 0.05 is, by definition, statistically significant; all others are not. Over the long run, you should only make a Type I error once out of every 20 tests. In the late 19th century, Francis Edgeworth, one of the first statisticians to routinely conduct tests of significance, used a very conservative = 0.005 (Stigler, 1999). The first formal statement of judging significance with p < 0.05 dates back to Fisher in the early 20th century, although there is evidence that it had been conventional for some time (Cowles, 1989). On the Other Hand "Surely, God loves the 0.06 nearly as much as the 0.05" (Rosnow and Rosenthal, 1989, p. 1277). "Use common sense to extract the meaning from your data. Let the science of human factors and psychology drive the statistics; do not let statistics drive the science" (Wickens,