How many people have to die over a type II error?

摘要

Although much concern over type I errors has permeated psychology for decades,nthere is less concern over type II errors. In fact, type II errors constitute a seriousnproblem in safety research that can result in accidents and fatalities becausenresearchers fail to reject the null hypothesis due to arbitrary probabilitynthresholds. The purpose of this paper is to reveal how often type II errorsnoccur and the effect they have on applied ergonomics research. Computernsimulations using population parameters were generated, revealing that type IInerrors happen quite often, particularly with effect sizes between 0.2 and 1.2.nA utility analysis also reveals that the cost of type II errors on society is muchngreater than it needs to be. Solutions for avoiding type II errors are discussed.