Getting to a post p < 0.05 era
Are you already subscribed?
Login to check
whether this content is already included on your personal or institutional subscription.
Abstract
A special issue of the journal «The American Statistician», published in March 2019, aimed to discuss the problem of reproducibility of scientific research, both from a methodological point of view and from a practical point of view. In this paper we discuss some of the contributions presented in the Supplement, which comprises an Editorial and 43 papers divided in 5 sections. From a methodological point of view, we focus on two statistical procedures that are particularly simple to use and seem to satisfy one of the recent recommendations, namely to avoid a dichotomization of the results in significant vs. non significant. Some Authors also suggested not declaring a result as significant or not significant, but to provide a descriptive interpretation to each statistical indicator. A critical reading of the 43 papers presented in the special issue leads to the conclusion that there is still need of possible proposals that can supplement the p-value and Null Hypothesis Significance Testing.
Keywords
- p-value
- NHST
- significance
- second generation p-value
- credibility prior interval
- false discovery rate
- reproducibilty