Las pruebas de significaciónuna polémica abierta

Abstract

Statistical significance tests are the most frequently statistical tests used for analyzing research data, both in Natural Sciences and Social Sciences, but their application has caused continuous controversy which is prevailing nowadays. The different opinions about these tests concern to three aspects: (a) the wrong meaning about the significance level; (b) the incorporation of the size effect as a complementary value for the explanation of significant hypothesis, and (c) the important role of this kind of tests in order to accept research studies for publication. In this work some opinions about the present controversy are exposed, both those of them who are really opposed and those of them who defend still their importance and utility. Also some recent alternative analysis are introduced, although none of them is seen like a valid substitute for the significance tests. For these reasons, we conclude that the generalized application of the significance tests must be avoided, and when they are used they have to be enclosed by statistical procedures with information about the real size, direction and importance of the results reported.