Academic journal article Public Personnel Management

Disparate Results in Adverse Impact Tests: The 4/5ths Rule and the Chi Square Test

Academic journal article Public Personnel Management

Disparate Results in Adverse Impact Tests: The 4/5ths Rule and the Chi Square Test

Article excerpt

In a Title VII sex discrimination case, a judge may be faced with the following scenario. The plaintiff claims that the organization's selection procedure has adverse impact on women, and shows that the selection ratio of women is significantly less than the selection ratio of men, based on the Chi Square test. The organization rebuts the claim of adverse impact by using the exact same numbers and shows that the selection ratio of women is more than 4/5ths of the selection ratio of men, therefore there is no adverse impact according to the 4/5ths rule. Is there adverse impact or not? The answer is ... yes and no. The two tests for adverse impact sometimes yield conflicting results. The Chi Square is the appropriate statistical test (testing for significant differences in selection ratios between minority and majority applicants), and if the sample size is large, the Chi Square will indicate that there is adverse impact even when the 4/5ths rule indicates there is not. But if the sample size is small, the Chi Square will sometimes indicate that there is no adverse impact when the 4/5ths rule says there is. The reason for the lack of agreement is that the 4/5ths rule is only a rough approximation of the Chi Square test: when the sample size is small, the Chi Square test has low power; and when the sample size is large, the Chi Square test will identify even small differences in selection ratios as adverse impact.

The Uniform Guidelines on Employee Selection Procedures (1) defines adverse impact as "a substantially different rate of selection in hiring, promotion, or other employment decision which works to the disadvantage of members of a race, sex, or ethnic group." Section 4D of the Guidelines describes how employers and the EEOC will determine whether adverse impact exists: "A selection rate for any race, sex, or ethnic group which is less than four-fifths (4/5) (or eighty percent) of the rate for the group with the highest rate will generally be regarded by the federal enforcement agencies as evidence of adverse impact."

With small sample sizes, the outcome for one or a few individuals may dramatically affect the calculation of adverse impact by the 4/5ths rule. For example, if there are 40 applicants, 20 women and 20 men, and 5 women and 8 men are hired, there is adverse impact according to the 4/5ths rule because SRminority = 5/20 = .25, SRmajority = 8/20 = .40, Adverse Impact Ratio (AIR) = .25/.40 = .63. If 6 women and 7 men had been hired, there would have been no Adverse Impact according to the 4/5ths rule, because SRminority = 6/20 = .30, SRmajority = 7/20 = .35, Adverse Impact Ratio = .86.

Although the Guidelines explicitly define the 4/5ths rule as a test for adverse impact, it has been criticized by a number of researchers. (2,3,4,5,6,7) Judges rarely include the raw data in their decisions, but cases of disagreement between the 4/5ths rule and the Chi Square test can still be found. For example, in Jackson v. Nassau County Civil Service Commission, (8) the court ruled that there was no adverse impact for a test with a passing rate for blacks of 40/55 = 72.7% and a passing rate for whites of 99/113 = 87.6%, and AIR = .83. However, the test had adverse impact according to the Chi Square test, with [2.sub.(1)] = 5.74, p = .017. Other examples can be found covering the scope of Title VII. (7,9,10,11,12,13,14,15,16,17)

It is important for personnel managers to be aware that there is an area of disagreement between the 4/5ths rule and the Chi Square test, so that they can effectively monitor their organization's EEO compliance and are not surprised by contradictory tests of adverse impact. How often will the two tests lead to conflicting conclusions about the presence of adverse impact? When the two tests for adverse impact disagree, which test should be relied upon under which circumstances? And if a suit is filed against the organization, how can the strongest and most persuasive defense be presented? …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.