"The single most important problem with null hypothesis testing is that it provides researchers with no incentive to develop precise hypotheses. To perform a significance test, one need not specify the predictions of either one’s own research hypothesis or those of alternative hypotheses"
(Gerd Gigerenzer 1993)