Glossary · Assessment Science
Adverse Impact
Adverse impact occurs when a selection procedure — a test, interview, or other hurdle — produces a substantially different selection rate for members of a protected group compared with the highest-passing group. In US employment law, the threshold is the four-fifths rule: a selection rate below 80 % of the top group's rate signals potential adverse impact.
Why it matters in hiring and assessment.
Adverse impact is both a legal concept and an ethical one. Under the US Uniform Guidelines on Employee Selection Procedures (1978), employers must monitor their selection tools for adverse impact and be able to demonstrate business necessity and job-relatedness if disparate rates appear. In India, while there is no single equivalent regulation, equal-opportunity commitments and regulatory scrutiny of large employers make the concept directly applicable to structured hiring.
For assessment practitioners, adverse impact analysis is a validity check as much as a compliance check. If a coding test, aptitude battery, or situational judgement test eliminates candidates from a particular group at a disproportionate rate, that disparity may signal that the test is measuring something other than job-relevant ability — for example, test-taking familiarity, language exposure, or socioeconomic background. Identifying this early allows you to revise the test, adjust the cut score, or provide practice materials before the disparity becomes a pattern.
Organisations that track adverse impact proactively are better positioned to defend their selection tools if challenged, and tend to build more diverse, higher-performing teams over time — because they catch irrelevant test elements before they filter out qualified candidates.
Example.
An IT company uses an English-language verbal reasoning test to screen all software developer applicants. Of 100 candidates from Group A, 60 pass (60 %). Of 80 candidates from Group B, 30 pass (37.5 %). The selection rate for Group B is 37.5 ÷ 60 = 62.5 % of Group A's rate — well below the four-fifths (80 %) threshold. This triggers an adverse impact flag. The company should examine whether verbal reasoning is genuinely job-relevant for the developer role, and consider whether the test can be redesigned or whether a job-relevant alternative (a coding test) would produce fairer outcomes.
Related terms.
- Criterion Validity
Demonstrating that a test predicts actual job performance — the business-necessity evidence required when adverse impact is found.
- Cut Score
The passing threshold that determines who advances — adjusting the cut score is one lever for managing adverse impact rates.
- Construct Validity
Evidence that the test measures what it claims to measure — necessary for establishing that a disparate-impact tool is genuinely job-relevant.
- Norm-Referenced Scoring
Scoring relative to a reference population — the composition of the norm group affects whether adverse impact analysis reflects the actual applicant pool.