Abstract

The tools of statistical inference that are usually adopted assume that all relevant decisions be made before data begin being collected. That is, all statistical choices need to be made before results become known. Therefore, pre-registered reports simply render it explicit what before was taken for granted implicitly. Instead of pre-registered reports one can make recourse to reports that relate exploratory studies. It is doubtful, however, that we have at our disposal statistical procedures that can be safely applied post-hoc. Finally, the inordinate increase in the number of scientific papers renders the entirety of their results impossible to grasp. That perhaps signals the end of journal papers as the principal tool for making results available to the scientific community.