Adverse Impact in Performance Management: What HR Should Know in 2026

A performance rating looks like a single number, yet it carries real weight because that score can shape a promotion, a bonus, a succession slot, or a layoff list. So here is the key question: can a performance management process produce unequal outcomes even when its rules look neutral? It can, and HR professionals call that pattern adverse impact.

Adverse impact does not need a biased manager or a bad policy, because a fuzzy rating scale, uneven calibration, or a rigid ranking rule can tilt results quietly.

A well-run performance management system helps you spot those patterns early, and performance management software adds the data layer. You can compare ratings, promotions, and pay across groups and review cycles in one place.

The legal picture also shifted in June 2026. The DOJ’s Office of Legal Counsel concluded on June 9 that the EEOC’s long-standing disparate-impact approach under Title VII is unconstitutional, and the EEOC’s new National Enforcement Plan puts intentional discrimination first while de-emphasizing disparate-impact claims. That story is still unfolding and is not settled law, so this guide explains what changed, what did not, and what HR should do next.

What Is Adverse Impact in Performance Management?

Adverse impact happens when a neutral-looking practice produces clearly different results for a protected group. In performance management, that practice could be a rating scale, a calibration meeting, or a bonus formula, and the result shows up in who earns top ratings, promotions, and raises.

Adverse Impact vs. Disparate Treatment

Disparate treatment is intentional: a manager rates someone lower because of race, sex, age, or another protected trait. Adverse impact looks at effects instead of motives, because a rule applies to everyone yet one group ends up with worse outcomes.

This distinction matters because hunting only for biased managers misses flawed processes, since a rating form can disadvantage a group while every manager acts in good faith.

Where Adverse Impact Can Enter the Performance Cycle

Picture the cycle as a chain: Goals → Performance Review → Rating → Calibration → Promotion → Pay → Career Progression. Unequal outcomes can enter at any link.

• Goals: some employees get visible, high-credit projects, while others get routine work that rarely earns recognition.
• Review: vague criteria leave room for personal judgment.
• Rating: managers apply the same scale in different ways.
• Calibration: persuasive managers win better scores for their teams.
• Promotion: ratings feed straight into advancement lists.
• Pay: merit increases and bonuses follow the rating.
• Career progression: small early gaps compound over the years.

An uneven outcome is a signal, not a verdict, and a gap in ratings does not prove unlawful discrimination. Tenure, role mix, and documented results may explain it, so your job is to investigate before you conclude.

How Performance Reviews Can Create Adverse Impact

Subjective Performance Ratings

Vague criteria hand managers wide discretion, because phrases like “strong attitude” or “leadership presence” mean different things to different raters. Stanford researchers who coded reviews at a Fortune 500 technology company found that gender stereotypes shaped what managers noticed and rewarded, and they concluded that ill-defined protocols let bias persist.

Measurable indicators work better, because closed deals, defect rates, and project milestones give raters something concrete, and behavioral judgments still work when each rating level carries a written example.

Inconsistent Manager Ratings

Two managers can watch identical performance and assign different scores. One treats “meets expectations” as the realistic ceiling, while another hands out “exceeds” to anyone who finishes on time.

Lenient raters push their people ahead in the promotion queue while strict raters hold equally strong employees back, and if certain groups cluster under certain managers, those habits become group-level gaps. Performance reviews built on shared templates and defined rating levels narrow that spread.

Self-Assessments and Manager Evaluations

Differences between self-ratings and manager ratings deserve a close look because they show how people present their work and how managers respond to it.

A 2023 study in the Journal of Economic Behavior & Organization examined more than 100,000 performance reviews across 170 companies, and RAND’s summary reports that women rated their own performance lower than men did, even after accounting for manager and peer reviews.

A 2025 study in the same journal examined a multinational firm and found that managers rated people of color lower than White employees, while women of color gave themselves the lowest self-ratings. When a one-year disruption hid self-evaluations from managers, ratings tracked self-scores less closely, yet the gaps stayed because managers anchored on prior-year ratings.

Self-assessments influence outcomes, and so does rating history, so track both.

How to Measure Adverse Impact in Performance Ratings

Calculate Selection Rates

Start with one simple formula:

Selection Rate = Number Selected ÷ Number Considered

Apply it to any decision point, where “selected” might mean earning a top rating or getting promoted. Suppose 40 of 80 men earn an “exceeds expectations” rating, a selection rate of 50%, while 24 of 80 women earn the same rating, a rate of 30%.

Use the Four-Fifths Rule Carefully

The four-fifths rule gives you a quick screen. Under the Uniform Guidelines on Employee Selection Procedures (29 CFR §1607.4(D)), a selection rate for any group below four-fifths (80%) of the highest group’s rate is generally regarded as evidence of adverse impact.

Run the numbers from above: divide 30% by 50%, and you get 60%, which falls below 80% and deserves a closer look.

Treat the rule as a rule of thumb, though, because falling below 80% does not prove unlawful discrimination. Sample size changes everything: on a team of ten with five people per group, one extra top rating can move a ratio from 67% to 100%, since 2 of 5 compared with 3 of 5 becomes 3 of 5 compared with 3 of 5.

Statistical significance tests help here, since analysts often use Fisher’s exact test to check whether a gap could be random chance. Because the June 2026 DOJ opinion criticizes the EEOC’s disparate-impact guidelines, ask employment counsel how the Uniform Guidelines apply to you today, although the calculation still works as an internal screening tool.

Look Beyond a Single Percentage

One ratio tells a thin story, so build a fuller picture by analyzing:

• Rating distributions across groups
• Promotion rates after each cycle
• Compensation outcomes, including merit increases
• Bonus eligibility and payout levels
• Manager-level patterns
• Historical trends across several review cycles

A gap that repeats for three cycles under the same managers needs attention.

Adverse Impact in Promotions and Compensation

When Performance Ratings Influence Promotion Decisions

Many companies require a top rating before they consider someone for promotion, so a small rating gap becomes a large promotion gap.

Compare rating outcomes with actual promotion rates. If two groups earn top ratings at similar rates but one group gets promoted less often, that points to a problem after the rating stage, so look for the differences first and assign causes later.

Performance Ratings and Pay Decisions

Pay-for-performance systems turn ratings into dollars through merit matrices, bonus pools, and equity grants. Check whether the same rating produces the same pay result for everyone. An “exceeds” rating should lead to a similar raise across teams, and if one department pays out more than another, you have found a consistency issue worth fixing.

Audit the Full Decision Chain

Use a simple framework: Rating → Calibration → Promotion → Pay.

Reviewing only the final rating misses too much, because ratings may look balanced while calibration quietly shifts scores down for one group, so check each stage separately to learn where the unevenness began.

Manager Bias and Adverse Impact in Performance Reviews

Common Sources of Rating Bias

Most managers want to rate fairly, but common thinking errors creep in:

• Halo and horns effects: one strong or weak trait colors the whole rating.
• Leniency and severity: some managers score high across the board, and others score low.
• Similarity bias: managers favor people who resemble them.
• Recency effects: the last six weeks outweigh the other ten months.
• Personal relationships: friendship softens the score.
• Vague criteria: unclear standards let all of the above grow.

The CIPD’s evidence review on performance management also examines interpersonal bias as a factor that affects the fairness of appraisal.

Why Manager Consistency Matters

Fair systems apply the same standards to every employee. Train managers with real rating examples for each performance level, hold calibration sessions where managers defend scores with evidence, and require written documentation behind every rating.

Short notes on specific results and dates make a score easier to defend, and frequent check-ins and one-on-ones help too, because they create a year-round record so managers rely less on memory.

Can Forced Ranking Create Adverse Impact?

How Forced Distribution Works

Forced distribution sorts employees into preset categories, such as 20% in the top tier and 10% in the bottom, and managers rank people against each other rather than against defined goals. Goal-based evaluation asks whether each person met the objective, while forced ranking asks who ranks above whom.

Risks of Ranking Employees Against Each Other

A fixed bottom percentage creates trouble when team strength varies, because a high-performing team still has to label someone a low performer while a weaker team may reward average work with a top score.

Research backs the concern. A 2015 Journal of Applied Psychology simulation study found that using a forced distribution rating system for layoffs, rather than retention, may raise the likelihood of adverse impact violations, with violations likely to peak in the first year of use.

If you use forced ranking, run adverse impact checks every cycle.

How Performance Management Software Helps Identify Unequal Outcomes

Spreadsheets can flag a problem once, but they struggle to track patterns across cycles, managers, and decisions, so purpose-built performance management software handles that work better.

Compare Performance Ratings Across Groups

Good analytics let you slice rating distributions by department, manager, location, job level, or other legally appropriate categories, so you see unusual patterns fast. Treat each pattern as a question rather than an answer.

Connect Ratings With Promotion and Compensation Data

Ratings tell only half the story, so connect them to promotion and pay records to see whether rating differences carry forward into bigger decisions. Compare outcomes across review cycles to separate blips from lasting trends.

Support Calibration and Consistent Reviews

Software also supports consistency before bias appears. Look for tools that:

• Centralize performance criteria in one place
• Capture feedback and supporting evidence during the cycle
• Give HR visibility into rating patterns by manager
• Keep an auditable record of changes and decisions

eLeaP’s employee performance management software covers customizable review templates, analytics, and audit trails in one platform, which gives HR teams a clear view of how ratings flow into later decisions.

Can AI Performance Reviews Create Adverse Impact?

How AI Can Influence Performance Decisions

AI now touches performance management in several ways:

• Automated scoring of work output
• Employee monitoring data
• Predictive performance analytics
• Automated recommendations for ratings or promotions
• AI-generated performance summaries

Each tool can speed up reviews and also carry bias into the final decision.

Where Algorithmic Bias Can Enter

Bias can sneak in through several doors:

• Historical training data: the model learns from past ratings that may already contain bias.
• Biased performance measures: if the metric favors certain work styles, the model repeats that preference.
• Proxy variables: factors like commute distance or schedule patterns can stand in for protected traits.
• Poorly designed scoring models: weak design amplifies small errors.
• Overreliance on recommendations: managers accept the output without questioning it.

Keep Human Review in the Decision Process

Treat AI output as decision support rather than unquestioned judgment, test your systems regularly for group-level differences, and document how automated recommendations affect final decisions.

Some state and local rules also require bias testing for automated employment tools. New York City’s Local Law 144, for example, requires bias audits of automated tools used in hiring and promotion decisions, so ask your legal team which rules apply to you.

What Changed Around Adverse Impact in 2026?

DOJ Position on Disparate-Impact Liability

On June 9, 2026, the DOJ’s Office of Legal Counsel released an opinion arguing that disparate-impact liability, as the EEOC applies it, raises equal protection concerns. It also pushed for a stricter reading of Title VII, with a broader business-necessity defense and a requirement that plaintiffs show the challenged practice caused the disparity and identify an equally effective alternative.

Read this as the DOJ’s legal position, not universal law. The opinion does not amend Title VII, override court precedent, or end private litigation and state-law duties, and courts are not bound by it.

EEOC’s 2026 National Enforcement Plan

The EEOC approved its National Enforcement Plan on June 4, 2026, and the plan follows Executive Order 14281, signed on April 23, 2025. It prioritizes disparate-treatment theories over disparate impact, and the agency states that it will not commence, develop, or continue litigation advancing disparate-impact claims.

That shift lowers federal enforcement risk, but it does not remove every risk. Private plaintiffs, state attorneys general, and state civil rights groups can still pursue these claims.

What HR Teams Should Take From the 2026 Changes

Separate legal requirements from internal quality reviews, because your duty to run a fair, job-related process has not gone away. Keep three habits:

1. Continue examining performance data for unexplained patterns.
2. Keep performance criteria job-related and consistently applied.
3. Document the evidence behind ratings and decisions.

Intentional discrimination remains illegal, and the EEOC now makes it a priority, so clean documentation protects you under either theory. This article offers general information, not legal advice, so check with employment counsel about your situation.

How to Reduce Adverse Impact in a Performance Management System

Define Measurable Performance Standards

Start with job-related criteria, replace vague traits with observable results, and link each rating to documented targets. Clear goals and OKRs give managers evidence to rate against, so they lean less on gut feel.

Train Managers on Consistent Rating Practices

Explain what each rating level means, show examples of work at every level, and review managers whose rating patterns look unusual. A learning management system, such as the eLeaP LMS, can deliver rater training and track completion across your whole manager group.

Run Regular Performance Data Reviews

Make data review a routine, and track:

• Rating distributions
• Promotion outcomes
• Compensation outcomes
• Manager-level differences
• Changes across review cycles

Review the Process, Not Just the Final Rating

Treat goals, feedback, ratings, calibration, promotion, and compensation as connected stages, and fix the stage where the gap starts, because adjusting the final score afterward treats the symptom and ignores the cause.

Frequently Asked Questions About Adverse Impact

What is adverse impact in performance management?

It is a pattern where a neutral-looking performance practice produces notably different outcomes for a protected group, and ratings, promotions, or pay may show the gap even without any intent to discriminate.

How is adverse impact calculated?

Divide the number of people selected by the number considered for each group, then compare each group’s rate to the highest group’s rate, and add a statistical significance test for a reliable read.

What is the four-fifths rule?

It is a screening guideline under which a group’s selection rate below 80% of the top group’s rate may signal adverse impact. It starts a review and does not settle one.

Can performance ratings have adverse impact?

Yes, because vague criteria, inconsistent managers, and rigid distributions can all skew ratings, and large review datasets show gender and race gaps in both self-ratings and manager ratings.

What is the difference between adverse impact and disparate treatment?

Disparate treatment involves intentional unequal treatment, while adverse impact involves a neutral practice with uneven effects. Intent matters for the first and does not define the second.

Can performance management software detect adverse impact?

It can surface patterns by comparing ratings, promotions, and pay across groups and cycles, but people still need to interpret them and decide what to fix.

Can AI performance reviews create adverse impact?

Yes, since biased training data, proxy variables, and heavy reliance on automated scores can all create uneven outcomes.

Does an adverse-impact ratio below 80% prove discrimination?

No. The ratio is a warning sign, and small samples, chance, and legitimate job-related factors can all explain a low number, so investigate before you conclude.

How often should companies review performance-rating data?

Review it after every rating cycle at a minimum, and add a mid-year check on promotion and pay decisions.

Conclusion: Use Performance Data to Examine the Full Employee Decision Process

Adverse impact does not live in one spot, because it can enter through goals, ratings, calibration, promotions, or pay, so examine the whole decision chain rather than just the final score. Fair outcomes start with measurable standards, consistent manager practices, and evidence-based decisions.

Performance management software makes this work practical because it connects ratings to promotions and pay and shows patterns across cycles. A platform like eLeaP gives HR teams that visibility without extra spreadsheets, so you can see how ratings shape employee opportunities and fix what the data reveals.