Abstract:
Fisher's approach to significance testing has in the past stumbled on how to make sense of the degree of improbability required to eliminate a chance hypothesis. Common significance levels in the social science literature of .05 and .01 seem to lack a rational foundation. Consequently, alternative comparative approaches to testing chance hypotheses have gained popularity. Nevertheless, by factoring in the probabilistic resources relevant to the occurrence of an event (probabilistic resources being the number of opportunities for the event to occur and be specified), it is possible to place Fisher's approach to significance testing on a firm rational foundation. Moreover, by generalizing the rejection regions to which Fisher's approach applies, one can use Fisher's approach to reliably detect intelligent design.