Efficacious End User Measures—Part 1: Relative Class Size and End User Problem Domains
Table 9
The top two JPTs have been “tuned” to a population where the diseased population is one hundredth of the undiseased population. In this environment, the two tests appear statistically indistinguishable.