5.4 Four-Fifths (80%) Rule, Disparate Impact Analysis, Standard Deviation Rule & Adverse Impact Remediation
Key Takeaways
- Adverse impact is a substantially different rate of selection in hiring, promotion, or other employment decisions which works to the disadvantage of members of a race, sex, or ethnic group.
- The Four-Fifths (80%) Rule establishes that a selection rate for any protected group which is less than four-fifths (80%) of the rate for the group with the highest selection rate is generally regarded as prima facie evidence of adverse impact.
- In large sample populations, federal courts enforce statistical significance tests—specifically the Two-Standard-Deviation Rule (Castaneda/Hazelwood) and Chi-Square analysis (Z >= 1.96, p < .05)—to determine if disparities are legally significant.
- Under Connecticut v. Teal (1982), a favorable 'bottom-line' hiring result does not insulate an employer from Title VII liability if an intermediate, non-validated hurdle (such as a written exam) produced adverse impact.
- Defensible adverse impact remediation strategies include holistic assessment batteries (pairing cognitive tests with structured interviews/work samples), Pareto-optimal weighting, robust pre-test recruitment, and score banding.
5.4 Four-Fifths (80%) Rule, Disparate Impact Analysis, Standard Deviation Rule & Adverse Impact Remediation
In public sector personnel administration, monitoring and analyzing applicant flow data for Adverse Impact (Disparate Impact) is a fundamental compliance responsibility. When an employment test, minimum qualification screening, or promotion process produces disproportionately lower success rates for protected demographic groups, the public agency faces substantial legal liability under Title VII unless the process is validated and compliant with federal standards.
1. Disparate Treatment vs. Disparate Impact
Title VII jurisprudence recognizes two distinct theories of employment discrimination:
+-----------------------------------------------------------------------------------------+
| DISPARATE TREATMENT VS. DISPARATE IMPACT |
| |
| EVALUATION AXIS DISPARATE TREATMENT DISPARATE IMPACT |
| ------------------- ----------------------------- ----------------------------- |
| Core Focus Discriminatory INTENT & MOTIVE Discriminatory CONSEQUENCES |
| |
| Legal Standard McDonnell Douglas Corp. v. Griggs v. Duke Power Co. |
| Green (1973) (1971); CRA 1991 (Sec. 105) |
| |
| Nature of Practice Explicit differential rules or Facially neutral policy applied |
| deliberate biased decisions identically to all candidates |
| |
| Primary Employer Bona Fide Occupational Job-Relatedness & Business |
| Defense Qualification (BFOQ) Necessity (Validation Study) |
| |
| Remedy Available Compensatory & Punitive Injunctive Relief, Back Pay, |
| Damages (CRA 1991) Front Pay, Policy Invalidation |
+-----------------------------------------------------------------------------------------+
2. The Four-Fifths (80%) Rule: Four-Step Calculation Protocol
Under Section 4D of the Uniform Guidelines on Employee Selection Procedures (UGESP), the Four-Fifths (80%) Rule serves as the primary federal administrative screening rule of thumb for determining adverse impact.
+-----------------------------------------------------------------------------------------+
| THE FOUR-STEP FOUR-FIFTHS (80%) RULE PROTOCOL |
| |
| STEP 1: Calculate Selection Rate (SR) for EACH Demographic Group |
| SR = (Number of Selected Candidates) / (Number of Total Applicants) |
| |
| STEP 2: Identify the BENCHMARK GROUP |
| Benchmark Group = Group with the Highest Selection Rate (SR_max) |
| |
| STEP 3: Calculate the IMPACT RATIO (IR) for Each Protected Group |
| Impact Ratio (IR) = (SR of Protected Group) / (SR of Benchmark Group) |
| |
| STEP 4: Compare IR to 0.80 (80% Threshold) |
| - If IR >= 0.80 (80%): NO Adverse Impact indicated. |
| - If IR < 0.80 (80%): ADVERSE IMPACT IS INDICATED (Prima Facie Evidence). |
+-----------------------------------------------------------------------------------------+
Comprehensive Case Study: Municipal Police Officer Examination
A city civil service commission administers an entry-level written examination and physical agility screening for Police Officer. The applicant flow and selection data are recorded as follows:
| Demographic Group | Number of Applicants ($N$) | Number Passing / Hired ($n$) | Selection Rate ($SR = n/N$) | Impact Ratio Calculation ($IR = SR / SR_{max}$) | Adverse Impact Status |
|---|---|---|---|---|---|
| Caucasian (White) | 300 | 180 | 60.0% ($180 / 300 = 0.60$) | Benchmark Group ($SR_{max} = 0.60$) | N/A (Benchmark) |
| African American | 120 | 48 | 40.0% ($48 / 120 = 0.40$) | $0.40 / 0.60 = \mathbf{0.667\text{ (66.7%)}}$ | ADVERSE IMPACT ($66.7% < 80%$) |
| Hispanic / Latino | 100 | 52 | 52.0% ($52 / 100 = 0.52$) | $0.52 / 0.60 = \mathbf{0.867\text{ (86.7%)}}$ | NO IMPACT ($86.7% \ge 80%$) |
| Asian American | 50 | 22 | 44.0% ($22 / 50 = 0.44$) | $0.44 / 0.60 = \mathbf{0.733\text{ (73.3%)}}$ | ADVERSE IMPACT ($73.3% < 80%$) |
Diagnostic Findings:
- The Caucasian applicant group represents the benchmark group with the highest selection rate ($60.0%$).
- The Hispanic group achieves an impact ratio of 86.7%, which exceeds 80%; therefore, no adverse impact is indicated.
- The African American group ($66.7%$) and Asian American group ($73.3%$) both fall below the 80% threshold, establishing prima facie evidence of adverse impact under UGESP.
3. Statistical Significance and the Standard Deviation Rule
While the Four-Fifths Rule is an administrative guideline, federal courts rely primarily on statistical significance tests—specifically when dealing with large applicant pools where minor numerical differences can trigger the 80% rule, or small samples where random variation distorts percentages.
The Standard Deviation Rule (Castaneda / Hazelwood Doctrine)
In Castaneda v. Partida, 430 U.S. 482 (1977) and Hazelwood School District v. United States, 433 U.S. 299 (1977), the Supreme Court established the Two-Standard-Deviation Rule: if the difference between the expected selection rate and the observed selection rate exceeds two to three standard deviations, the hypothesis that the selection process is neutral is rejected, establishing statistically significant disparity ($p < .05$).
+-----------------------------------------------------------------------------------------+
| STANDARD DEVIATION FORMULA FOR ADVERSE IMPACT (Z-SCORE) |
| |
| | A - E | |
| Z = ------------- |
| sqrt(V) |
| |
| Where: |
| A = Actual number of protected group candidates selected |
| E = Expected number of protected group candidates selected: |
| E = n_protected * (Total Selected / Total Applicants) |
| V = Variance of the hypergeometric/binomial distribution: |
| V = n_protected * p * (1 - p) * [ (N_total - n_selected) / (N_total - 1) ] |
| (In large applicant pools, the finite population correction factor is omitted: |
| sqrt(V) = sqrt( n_protected * p * (1 - p) ) ) |
| |
| *Decision Threshold: If Z >= 1.96 (p <= .05, two-tailed), the disparity is |
| statistically significant and constitutes legal disparate impact.|
+-----------------------------------------------------------------------------------------+
Small Sample Exception (Fisher's Exact Test)
UGESP Section 4D explicitly notes that the 80% rule is not applicable to sample sizes too small to be statistically meaningful. In small municipal departments (e.g., hiring 3 officers from 12 applicants), a single candidate hiring decision swings selection percentages by 25–33%. In such cases, Fisher's Exact Test is utilized to evaluate exact probability.
4. The "Bottom-Line" Defense Fallacy: Connecticut v. Teal (1982)
A critical legal doctrine tested on the PSHRA-CP exam is the Bottom-Line Defense.
+-----------------------------------------------------------------------------------------+
| THE LESSON OF CONNECTICUT V. TEAL (1982) |
| |
| [STAGE 1: PASS/FAIL WRITTEN TEST] =====================> ADVERSE IMPACT DETECTED |
| - White Pass Rate: 79.5% Impact Ratio: 68.2% (< 80%) |
| - Black Pass Rate: 54.2% |
| |
| [STAGE 2: FINAL PROMOTIONS MADE FROM PASSERS] ==========> NO BOTTOM-LINE IMPACT |
| - Agency promoted 22.9% of Black applicants Bottom-Line Ratio: 169% |
| - Agency promoted 13.5% of White applicants |
| |
| ----------------------------------------------------------------------------------- |
| SUPREME COURT RULING: |
| Title VII protects the INDIVIDUAL CANDIDATE, not just demographic group bottom lines. |
| An employer CANNOT use a favorable bottom line to excuse an unvalidated pass/fail |
| hurdle that unfairly excluded individuals at an earlier stage. |
+-----------------------------------------------------------------------------------------+
In Connecticut v. Teal, 457 U.S. 440 (1982), the State of Connecticut administered a written exam for permanent Supervisor positions. The test had adverse impact against Black candidates. To compensate, the State promoted a significantly higher percentage of Black candidates who passed the test, resulting in a favorable "bottom line." Disqualified Black candidates sued.
The Supreme Court ruled in favor of the employees, holding that Title VII Section 703(a)(2) protects individual employees from discriminatory barriers. An employer cannot justify an unvalidated, exclusionary intermediate examination hurdle simply by demonstrating that the ultimate bottom-line hiring figures achieved demographic balance.
5. Strategic Remediation: Reducing Adverse Impact in Public Selection
When a public agency's selection system manifests adverse impact, HR leaders must execute compliant remediation strategies that reduce disparate impact while preserving or enhancing valid predictive utility.
+-----------------------------------------------------------------------------------------+
| STRATEGIES FOR ADVERSE IMPACT REMEDIATION |
| |
| [1. DIVERSIFY ASSESSMENT MODALITIES] |
| Replace standalone cognitive tests with a multi-hurdle composite battery: |
| - Structured Behavioral & Situational Interviews (r = .40 - .50; Low Impact) |
| - Work Sample & Simulation Exercises (r = .45 - .55; Low Impact) |
| - Situational Judgment Tests (SJTs) (r = .35 - .45; Moderate Impact) |
| |
| [2. COMPOSITE BATTERY & PARETO-OPTIMAL WEIGHTING] |
| Avoid 100% weighting on cognitive written exams; weight non-cognitive and behavioral |
| components (e.g., 40% Written Knowledge, 40% Work Simulation, 20% Structured Oral). |
| |
| [3. SCORE BANDING & CATEGORY RATING] |
| Transition from strict top-down ranking to SEM-based score banding or Category |
| Rating (Quality, Well-Qualified, Qualified), expanding selection discretion. |
| |
| [4. PRE-TEST RECRUITMENT & CANDIDATE ORIENTATION WORKSHOPS] |
| Offer free, open test preparation orientations, practice familiarization materials, |
| and job preview guides to eliminate test-taking anxiety and procedural unfamiliarity. |
+-----------------------------------------------------------------------------------------+
Summary Table: Selection Methods Validity vs. Adverse Impact Profile
| Assessment Instrument | Predictive Validity ($r$) | Adverse Impact Risk Profile (Race/Gender) | Recommended Civil Service Deployment |
|---|---|---|---|
| General Cognitive Ability Tests | High ($r \approx .51$) | High against Black and Hispanic candidates. | Combine in composite battery; avoid high standalone pass/fail cutoff. |
| Work Sample Tests | High ($r \approx .54$) | Low across racial and ethnic groups. | Primary testing component for technical and craft positions. |
| Structured Interviews | High ($r \approx .45 - .51$) | Very Low when panel raters are calibrated. | Core component for supervisory, professional, and public-contact jobs. |
| Physical Ability Tests (PAT) | High ($r \approx .40 - .50$) | High against female candidates. | Ensure cutoff reflects minimum essential physical job requirements. |
| Assessment Centers | High ($r \approx .45$) | Low-to-Moderate across all demographic groups. | Gold standard for promotional executive and public safety command ranks. |
| Unstructured Interviews | Low ($r \approx .15 - .20$) | High due to unstandardized rater bias. | Prohibited or strictly discouraged in merit systems. |
A county public works department evaluates 200 male applicants and hires 60. It also evaluates 80 female applicants and hires 16. Using the Four-Fifths Rule, what is the Impact Ratio for female applicants, and does adverse impact exist?
Under the U.S. Supreme Court ruling in Connecticut v. Teal (1982), why was the State's 'bottom-line' defense rejected when an unvalidated promotional examination produced adverse impact?
In Title VII disparate impact litigation, what threshold under the Two-Standard-Deviation Rule (Castaneda v. Partida / Hazelwood School District) establishes statistically significant disparity between expected and observed selection rates?
Which of the following selection redesign strategies is most effective for reducing adverse impact while maintaining or enhancing the overall predictive validity of a civil service selection process?