Research Article
Efficacious End User Measures—Part 1: Relative Class Size and End User Problem Domains
Table 2
This table shows the total and for two tests with the same CPD on equally sized data sets (2250 observations), drawn from the same populations. The only difference is the samples have different rCSs. Because the tests were run on data sets with different sized classes, the equivalence of the CPD’s effectiveness is not obvious. It would be easy for an observer to erroneously conclude the CPDs were significantly different.
| ||||||||||||||||||