What Is Adverse Impact in Hiring, and Why Does It Matter?

Adverse impact occurs when a selection procedure produces substantially different rates of selection across groups defined by race, sex, national origin, or other protected characteristics. It is not about intent. An organization can have fully non-discriminatory intentions and still produce adverse impact through a selection process that has differential effects.

The legal standard comes from the EEOC’s Uniform Guidelines on Employee Selection Procedures (UGESP), which established the 4/5ths rule as the primary test.

The 4/5ths rule

The 4/5ths rule (also called the 80% rule) compares the selection rate of a protected group to the selection rate of the highest-selected group. If the ratio falls below 0.80, adverse impact is indicated.

For example: if the selection rate for white applicants is 50% and the selection rate for Black applicants is 35%, the ratio is 0.70. That is below 0.80, and adverse impact is indicated for that selection procedure.

What adverse impact is not

Adverse impact is not proof of discrimination. It is a flag. When adverse impact is detected, the employer must either demonstrate the validity of the selection procedure, modify the procedure to reduce the disparity, or discontinue it.

Adverse impact analysis also does not mean you need equal outcomes. It means the ratios need to stay within the 4/5ths threshold, and deviations require justification grounded in validity evidence.

Intersectional analysis

Standard adverse impact analysis looks at single-axis comparisons: race, sex, disability status independently. Intersectional analysis combines dimensions and looks at compound groups: Black women, Hispanic men over 40, and so on.

Intersectional analysis often reveals disparities invisible in single-axis analysis. It is not currently required by UGESP, but it is increasingly expected as a best practice and is required under some emerging regulatory frameworks.

Why continuous monitoring matters

Adverse impact can emerge over time even in procedures that showed no disparity at launch. Applicant pool composition shifts. Role requirements change. The model sees new data distributions. A selection procedure that was valid and fair at deployment may develop adverse impact after twelve months.

Continuous monitoring, not point-in-time auditing, is the defensible standard.