Statistical Inference as Severe Testing: How to Get Beyond the Statistics Wars

This eye-opener illuminates controversies surrounding widely used statistical methods across the physical, social, and biological sciences. New solutions to philosophical problems of induction, falsification, science vs. pseudoscience are put to work to let statisticians and reproducibility researchers get beyond hardened conceptual disagreements.

Format: Hardback
Length: 500 pages
Publication date: 20 September 2018
Publisher: Cambridge University Press

The mounting failures of replication in social and biological sciences have underscored the critical need for a rigorous evaluation of proposed reforms. This book delves into the deep-rooted disagreements between experts tasked with restoring the integrity of science. It challenges two prevailing perspectives on the role of probability in inference: namely, assigning degrees of belief and controlling error rates in the long run. If statistical consumers remain unaware of the assumptions underlying rival evidence reforms, they cannot effectively scrutinize the consequences that impact their fields, such as personalized medicine and psychology.

To address this issue, the book employs a straightforward tool: if little has been done to rule out flaws in inferring a claim, then it has not passed a severe test. Many methods advocated by data experts fail to withstand rigorous scrutiny and are at odds with effective strategies for mitigating cherry picking, selective reporting, and other forms of bias. Through a series of engaging excursions and exhibits, the philosophy and history of inductive inference come alive. Philosophical tools are applied to address problems related to science, pseudoscience, induction, and falsification, providing a comprehensive exploration of these critical concepts.

This book is a valuable resource for anyone interested in understanding the foundations of statistical inference, the role of probability in scientific inquiry, and the challenges faced by researchers in combating bias and promoting integrity in the scientific process. By shedding light on the disagreements between experts and offering practical solutions, it empowers readers to engage more critically with statistical evidence and make informed decisions based on rigorous scientific principles.

Weight: 898g
Dimension: 161 x 235 x 30 (mm)
ISBN-13: 9781107054134