Abstract
It is demonstrated that in significance testing under certain assumptions the lower limit of a confidence ratio (proportion of alternative hypotheses accepted that are true) and the lower limit of a power ratio (proportion of true His tested that are accepted) are both a function of the number of tested hypotheses accepted and the number of tested hypotheses refused acceptance. It is proposed that historical values of the latter quantities are observable and that the lower limits of the ratios may be valuable in justifying the use of significance testing procedures.
Get full access to this article
View all access options for this article.