P-Values in a Post-Truth World

TL;DR

This paper critically evaluates the American Statistical Association's recent statements on p-values, highlighting their shortcomings and emphasizing the importance of statisticians' role in fostering trustworthiness amid misinformation.

Contribution

It analyzes the ASA's guiding principles on p-values and demonstrates their failure to adhere to these principles, advocating for greater responsibility among statisticians.

Findings

ASA's statements on p-values do not fully adhere to their own principles

Statisticians have a crucial role in promoting trustworthiness in data interpretation

The paper calls for increased responsibility and role modeling by statisticians

Abstract

The role of statisticians in society is to provide tools, techniques, and guidance with regards to how much to trust data. This role is increasingly more important with more data and more misinformation than ever before. The American Statistical Association recently released two statements on p-values, and provided four guiding principles. We evaluate their claims using these principles and find that they failed to adhere to them. In this age of distrust, we have an opportunity to be role models of trustworthiness, and responsibility to take it.