Shortcomings of the 80% Rule - Insensitive to Magnitude of Disparity

Aside from being insensitive to the sample size, the 80% Rule is also insensitive to the magnitude of the disparity. To demonstrate this, consider the following two examples.


Assume that we have a population of 10,000 protected and 10,000 non-protected individuals who are applicants for a given position. Further assume that "Qualifications Test 1" disqualifies 3,000 protected individuals and no non-protected individuals for hire, and that "Qualifications Test 2" disqualifies 9,993 protected individuals and 9,990 non-protected individuals:






Using the 80% Rule, Qualifications Tests 1 and 2 appear to be identical in that they both result in a ratio of 70%. However, the two tests are quite different in the number of protected class members affected. In Qualifications Test 1, there are 3,000 more protected class members than non-protected class members who have been adversely affected. In Qualifications Test 2, there are 3 more protected class members than non-protected class members who have been adversely affected. There is certainly a difference of practical significance between 3,000 protected class members and 3 protected class members.



0 comments: