Spanos, Aris2024-02-122024-02-12https://hdl.handle.net/10919/117930The replication crisis and the untrustworthiness of empirical evidence is often viewed through the lens of the Medical Diagnostic screening (MDS) perspective, conceived as a surrogate for Neyman-Pearson (N-P) testing. To shed light on this crisis theMDS Positive Predictive Value (PPV) is Metamorphosed into the M-PPV by identifying the false positive/negative probabilities with the type I/II error probabilities. The diagnosis based on M-PPV is that the untrustworthiness of empirical evidence stems from several misuses of N-P testing, including p-hacking, data-dredging, and cherry-picking. The appropriateness of the MDS perspective, as well as the ensuing diagnosis based on the M-PPV, are called into question since they invoke dubious analogies with N-P testing. It is argued that a more pertinent explanation is that the untrustworthiness of evidence stems from a much broader problem relating to the uninformed and recipe-like implementation of frequentist statistics without proper understanding of the invoked assumptions, limitations, and warranted evidential interpretations of the frequentist inference results. This broader perspective, in conjunction with the post-data severity evidential interpretation of the testing results, could potentially address the untrustworthiness of empirical evidence problem.application/pdfenIn CopyrightReplication crisisuntrustworthy evidencefalse positive/negative ratesmedical diagnostic testingPositive Predictive Valuetype I and II error probabilitiestrustworthy evidencesensitivityspecificityprevalencepost-data severitystatistical results vs. evidenceArticle - Refereed