PhilSci Archive

The severity score is fundamentally flawed: A reply to Spanos' "Severity and Trustworthy Evidence"

Rochefort-Maranda, Guillaume (2022) The severity score is fundamentally flawed: A reply to Spanos' "Severity and Trustworthy Evidence". [Preprint]


Download (165kB) | Preview


Aris Spanos' recent critical appraisal of my paper on the severity measure of evidence and the Winner's Curse has the merit of allowing me to clarify my arguments. I show that Spanos' paper contains mistakes, contradictions, and an example that ultimately supports my conclusions. The bottom line is that the severity score is fundamentally defective because it favors the rejection of the null hypothesis due to sampling error as opposed to the discovery of true discrepancies from the null. Since the severity score is a function of the test statistics, it is especially corrupted by the large sampling error of underpowered test.

Export/Citation: EndNote | BibTeX | Dublin Core | ASCII/Text Citation (Chicago) | HTML Citation | OpenURL
Social Networking:
Share |

Item Type: Preprint
Additional Information: The Appendix can be found on my researchgate page
Keywords: severity; evidence; statistics; error probability
Subjects: Specific Sciences > Probability/Statistics
Depositing User: Dr. Guillaume Rochefort-Maranda
Date Deposited: 02 Nov 2022 21:22
Last Modified: 02 Nov 2022 21:22
Item ID: 21348
Subjects: Specific Sciences > Probability/Statistics
Date: 4 May 2022

Monthly Views for the past 3 years

Monthly Downloads for the past 3 years

Plum Analytics

Actions (login required)

View Item View Item