A hiring process can look consistent on paper and still produce materially different outcomes across demographic groups. This adverse impact analysis guide gives HR leaders, talent acquisition teams, and assessment consultants a practical way to identify those differences early, investigate their cause, and improve decision quality without lowering job-related standards.

Adverse impact analysis is not a one-time compliance exercise completed after a complaint. It is a recurring control for any organization that uses applications, resumes, assessments, interviews, background screens, or automated decision tools to narrow a candidate pool. When it is built into selection operations, it helps protect both the organization and the quality of its workforce decisions.

What adverse impact analysis measures

Adverse impact occurs when a selection procedure appears neutral but disproportionately excludes members of a protected group. The analysis compares selection rates between groups at a specific stage of the hiring process. The issue is not whether every group has identical results. The question is whether the difference is large enough to require closer review of the procedure and its job-related basis.

The most widely recognized initial screen is the four-fifths rule, also called the 80% rule. Calculate the selection rate for each group by dividing the number selected by the number of applicants in that group. Then divide the selection rate of the group being evaluated by the selection rate of the group with the highest selection rate.

For example, assume 100 applicants in Group A apply and 40 advance, creating a 40% selection rate. In Group B, 100 applicants apply and 28 advance, creating a 28% rate. The impact ratio is 28% divided by 40%, or 70%. Because 70% is below 80%, the result warrants investigation.

The four-fifths rule is a practical screening benchmark, not a legal conclusion. Small applicant pools can create dramatic ratios from a few decisions, while larger pools may reveal meaningful differences that deserve statistical analysis even when the ratio is above 80%. Organizations should evaluate practical impact, sample size, statistical significance, trends over time, and the requirements that apply to their workforce and locations.

Start with the right selection data

An adverse impact analysis is only as credible as its underlying data. Before running calculations, define the exact hiring population and preserve data for every meaningful decision point. That means tracking who applied, who met minimum qualifications, who completed an assessment, who passed a screen, who was interviewed, who received an offer, and who was hired.

Each stage should be analyzed separately. Looking only at final hires can hide the source of a problem. A résumé screen may have no concerning pattern, while a cognitive assessment cutoff, interview score, physical requirement, background criterion, or recruiter disposition code may be causing a substantial drop in selection rates.

Use consistent definitions. If one requisition treats withdrawn applicants as applicants and another does not, comparisons become unreliable. Confirm that demographic data is collected, stored, and reported appropriately under applicable law and company policy. Access should be limited to personnel with a legitimate business need, and individual-level results should not be used to make employment decisions.

Define the comparison group carefully

The appropriate comparison is usually between candidates competing for substantially similar jobs under the same process. Combining entry-level customer service applicants with experienced sales candidates can create misleading results because the job requirements, labor markets, and selection standards are different.

Segment analyses by job family, location, requisition type, and selection procedure when the volume supports it. At the same time, avoid slicing the data so narrowly that every group becomes too small to interpret. This is a judgment call. A practical reporting approach often includes a high-level enterprise view, job-family results, and deeper review of high-volume or higher-risk roles.

How to conduct an adverse impact analysis

A disciplined process is more valuable than a spreadsheet run once a year. The following sequence keeps the work connected to hiring operations.

First, map the selection process from application to hire. Identify every hurdle, the decision maker or system responsible, the scoring method, and the pass or fail rule. Include vendor tools, automated ranking logic, recruiter reviews, structured interviews, reference checks, and screening criteria.

Next, calculate selection rates and impact ratios at each hurdle and for the overall process. Use the group with the highest selection rate as the reference group for the four-fifths calculation. Review results across a meaningful period rather than reacting to a single week or requisition unless a serious concern is clear.

Then, examine flagged results in context. Ask whether the affected procedure measures a requirement that is genuinely necessary for successful performance. Determine whether the same standard was applied consistently, whether scores were recorded accurately, and whether there are alternative procedures that could achieve the business objective with less adverse impact.

Finally, document the review and the resulting decision. Documentation should show what was analyzed, the population and time period used, the findings, the business rationale for retaining or changing a process, and the follow-up plan. If a result may create material legal risk, involve qualified employment counsel before making conclusions or implementing significant changes.

Validate the procedure, not just the outcome

A favorable impact ratio does not prove that a selection tool is effective. Likewise, an unfavorable ratio does not automatically mean the tool must be removed. The central operational question is whether the procedure is job-related, appropriately administered, and supported by evidence that it predicts or relates to successful performance.

Validated assessments can strengthen selection decisions when they are aligned to the role, used according to their technical documentation, and combined with other relevant job information. For example, a sales assessment should relate to the competencies that distinguish successful sales performance in that environment. A behavioral profile may support structured interview questions and development planning, but it should not be treated as a shortcut for job analysis or used beyond its intended purpose.

Validation is especially important when organizations set cutoff scores. A cutoff should reflect a defensible performance need, not an arbitrary preference for fewer candidates to review. Raising a threshold may improve administrative efficiency, but it can also reduce the available talent pool and increase adverse impact risk. The best threshold is the one supported by job requirements and performance evidence.

Review human judgment and automated tools

Human decisions can introduce inconsistency even when the formal process is sound. Unstructured interviews, vague definitions of culture fit, and undocumented recruiter rejections are common weak points. Structured questions, anchored rating scales, trained interviewers, and documented disposition reasons make decisions more consistent and easier to evaluate.

Technology requires the same discipline. Applicant tracking systems, résumé ranking tools, AI-enabled screening, and vendor assessments should be reviewed for their inputs, scoring rules, intended use, and results. A vendor’s assurance is not a substitute for an employer’s oversight. The organization remains responsible for understanding how a tool affects its own applicant population and job decisions.

What to do when you find a concerning result

Do not assume the answer is to eliminate every procedure with an impact ratio below 80%. Start by confirming the data, the group definitions, and the calculation. Then identify the precise stage where the gap emerges. A change to a poorly defined minimum qualification may matter more than replacing an assessment that is performing as intended.

Potential corrective actions depend on the findings. An organization may clarify job requirements, remove nonessential screens, revise a cutoff score, standardize interviewer training, improve accommodations processes, use a more job-relevant assessment, or add a complementary evaluation method. Any change should be tested and monitored. A well-meaning adjustment that reduces one disparity but weakens prediction of performance or creates inconsistency elsewhere is not a durable solution.

Maintain a regular cadence for review. High-volume hiring programs may need quarterly or monthly monitoring, while lower-volume roles may be reviewed semiannually or annually. Trigger additional analysis when the organization introduces a new assessment, changes a screening standard, expands into a new labor market, or sees a sharp change in candidate flow.

Maximum Potential helps organizations apply validated assessment tools within a broader decision-support process, connecting pre-hire evaluation with the competencies and development needs that matter after hire. That connection supports a more useful goal than simply passing a ratio test: selecting people with evidence-based standards that are relevant to performance.

Better hiring decisions are not produced by ignoring differences in selection outcomes or by abandoning standards at the first warning sign. They come from measuring outcomes carefully, validating what the process is designed to assess, and making targeted improvements before a small process flaw becomes a costly workforce problem.