< back

Adverse Impact

Related Additional Information topics:

For a description of associated Data Tables:

To learn about ADVERSE IMPACT reporting, see:


Adverse Impact Ratio Analysis by Favored Group

balanceAAP can perform ADVERSE IMPACT reporting on personnel actions to support your organization's Affirmative Action monitoring program by flagging potential problem areas. The system will flag statistical results on ADVERSE IMPACT reports, when they meet the parameters below.

Four-Fifths Rule

The "Four-Fifths" (or 80%) rule to is used in Adverse Impact testing to compare the selection rates for classes of applicants or employees against the rate at which the most-represented class is selected. If the selection rate for a lesser-represented class is less (positive action) or more (negative action) than four-fifths (or 80%) of the selection rate for the most-represented group, the results are flagged for potential Adverse Impact.

Positive Employment Action (Applicants, Promotions)

The favored class is the one with the highest selection rate. Statistical evidence exists for Adverse Impact if the selection rate of the non-favored class is less than 80 percent of the favored class.

Example:
Applicant Pool for Job Group = 100 Males and 68 Females
Number of Females Selected: 10 / 68 = 0.147 (15%)
Number of Males Selected: 6 / 100 = 0.60 (6%)
Favored Group = Females
0.06 / 0.15 = 0.40 (40%)

There is statistical evidence that Adverse Impact exists.

Negative Employment Action (Terminations, Involuntary Promotions)

The favored class is the one with the lowest selection rate. Statistical evidence exists for Adverse Impact if the selection rate of the favored class is greater than 80 percent of the non-favored class.

Example:
Termination Pool for Job Group = 168 Whites and 39 Minorities
Number of Minorities Selected: 2 / 39 = 0.05 (5%)
Number of Whites Selected: 10 / 168 = 0.06 (6%)
Favored Group = Minorities
0.05 / 0.06 = 0.83 (83%)

The statistical evidence does not show Adverse Impact.

Note: balanceAAP relies on Standard Deviation or Fisher’s Exact test to determine if the above results are statistically significant, depending on sample size. Results will be flagged if they are statistically significant.