Adverse impact is a group-level difference in selection rates within a defined employment process; one applicant’s personality report cannot establish it. Under the U.S. federal Uniform Guidelines, the four-fifths ratio is an initial signal, not a complete finding about cause, job-related support, or legality. Interpreting a signal means defining the job and decision path, examining how scores affected advancement, and considering job evidence, accessibility, and suitable alternatives.
What does adverse impact describe in a personality assessment process?
In the U.S. federal employment-selection context, adverse impact describes a substantially different selection rate that disadvantages a race, sex, or ethnic group. It is a pattern in outcomes across groups within an employment decision process, not a label for one candidate’s score, a disappointing result, or an employer’s intent. The EEOC’s *Questions and Answers to Clarify and Provide a Common Interpretation of the Uniform Guidelines on Employee Selection Procedures* frames the inquiry around group selection rates. Its role here is technical assistance interpreting the Guidelines; it says it does not have the force and effect of law, and it cannot decide what happened in a particular company’s process.
A personality assessment can still be relevant to that review when an employer uses it as one step in selecting applicants. The key distinction is what the evidence describes: an individual report records one person’s responses or result, while adverse impact concerns the outcomes of a defined group of people moving through an employment decision. A single report cannot establish whether group rates differed. Nor does a rejection, by itself, show that the assessment caused a group-level pattern. Those conclusions require process information beyond the candidate’s personal result, including group-specific counts and a clear account of how the employer made the relevant decision.
Four questions therefore need to stay separate. First, did selection rates differ between the groups being compared? Second, which step or combination of steps contributed to that outcome? Third, is the assessment’s particular use supported for the job in question? Fourth, what response is appropriate if a concerning pattern is found? The Uniform Guidelines address these as related but distinct parts of an employment-selection inquiry. This article can explain the distinctions and the evidence each question calls for; it cannot adjudicate a company’s case from an individual personality report.
How is a selection-rate comparison calculated?
A selection rate is the number of people selected from a defined group divided by the number of people from that group considered at the decision point being examined. An impact ratio compares one group’s rate with the highest group rate in that same comparison: divide the lower rate by the highest rate. The EEOC’s *Questions and Answers to Clarify and Provide a Common Interpretation of the Uniform Guidelines on Employee Selection Procedures* lays out this rate-and-comparison method. The Office of Personnel Management’s *Designing an Assessment Strategy* also describes comparing a lower passing or selection rate with the highest rate. Both calculations depend on first defining who was considered and what counted as selection. The comparison also needs a stated group basis and a common point in the process, so the two percentages have a meaningful relationship rather than merely sharing a label such as “applicants.”
For a hypothetical arithmetic example, suppose 60 of 100 people in Group A are selected, while 40 of 80 people in Group B are selected at the same decision point for the same defined job. Group A’s selection rate is 60/100, or 60%. Group B’s rate is 40/80, or 50%. The impact ratio for Group B relative to the highest rate is 50% divided by 60%, which is about 83%. These numbers are invented solely to show the operation; they are not company data or findings from a study.
The numerator and denominator must refer to the same kind of decision and a clearly bounded pool. If one rate counts people who passed an assessment and another counts people who received a job offer, the comparison mixes decision points. If one denominator includes everyone who applied while another includes only those who reached an interview, the rates answer different questions about different pools. A reader should therefore ask what group entered each denominator, what outcome was counted in each numerator, and whether both rates concern the same job and stage. Changing those definitions can change the comparison even when the arithmetic is flawless.
A legible ratio is useful for describing the comparison that was defined. It does not explain why outcomes differ, identify which part of a hiring process contributed, or establish whether an assessment is supported for a job. Those require separate evidence and questions; the calculation itself only states how the two defined rates relate.
Sources: Questions and Answers to Clarify and Provide a Common Interpretation of the Uniform Guidelines on Employee Selection Procedures; Designing an Assessment Strategy
What can the four-fifths ratio tell you—and what can it miss?
The four-fifths rule is a screening convention: under the Uniform Guidelines, a group’s selection rate below four-fifths (80%) of the rate for the group with the highest rate is generally regarded as evidence of adverse impact. The Office of Personnel Management’s *Designing an Assessment Strategy* calls this a widely used rule of thumb. Its value is practical: it gives an employer or reviewer a simple, consistent first comparison that can flag a pattern for closer examination. It is a signal about the rates as defined, not a finding about why they differ, the merit of any individual, or whether the assessment predicts relevant work. A result at or above 80% can therefore help triage a comparison without certifying that the procedure is fair or properly supported.
The EEOC’s *Questions and Answers to Clarify and Provide a Common Interpretation of the Uniform Guidelines on Employee Selection Procedures* explains why the threshold cannot be used as a verdict in either direction. A difference smaller than 20% may still warrant concern when statistical and practical significance indicate that the disparity is meaningful. Conversely, a raw rate comparison alone does not automatically determine that unlawful discrimination occurred. The Guidelines’ approach considers the particular employment process and other relevant evidence; the ratio does not establish cause, intent, job relationship, or the appropriate legal conclusion. Thus, a below-threshold comparison calls for context rather than a declaration, while a ratio above the threshold cannot close the inquiry where other evidence points to a material disparity.
Sparse counts illustrate a mathematical limit without establishing an empirical cutoff. If a defined group has only a few people in its denominator, one additional selection changes its rate by a large fraction of that denominator; with a much larger denominator, one person changes the rate less. The resulting ratio can move sharply when counts are small, even though the arithmetic is correct. This observation does not mean that small samples always produce impact, or that a particular minimum count makes a comparison reliable. It shows why the underlying counts and the way the comparison was defined matter alongside the percentage. Readers should not infer stability, statistical significance, or practical importance from the ratio alone; those are separate questions requiring appropriate evidence.
The useful judgment is therefore to treat four-fifths as a common first screen, then ask what else bears on the apparent pattern: how many people were counted, whether the comparison is meaningful for this decision, and whether other evidence changes its interpretation. Dismissing the convention because it is incomplete would discard an accessible warning signal. Treating it as a safe harbor or automatic legal test would give it more force than the EEOC guidance assigns. It organizes the next question; it does not answer every question about fairness or establish a company’s legal responsibility.
Sources: Questions and Answers to Clarify and Provide a Common Interpretation of the Uniform Guidelines on Employee Selection Procedures; Designing an Assessment Strategy
Why does the whole job-specific decision path matter?
The Uniform Guidelines start by asking whether the total selection process for a particular job shows adverse impact. The EEOC’s *Questions and Answers to Clarify and Provide a Common Interpretation of the Uniform Guidelines on Employee Selection Procedures* describes the total selection process as the combined procedures leading to the final decision. This process-level view answers an outcome question: across the route applicants actually take for that job, do the combined steps produce different selection rates? It prevents a reviewer from treating the personality assessment’s isolated pass rate as though it were automatically the final employment outcome, especially when later stages also affect who is selected.
Consider a hypothetical funnel, offered only as an illustration: applicants first pass an application review, then complete a personality assessment, interview, and receive a final decision. Each stage can change who remains eligible or advances. The rate at the end reflects the combined route, not the assessment alone. Looking at the whole path first keeps the analysis connected to the employment decision people ultimately face. It also makes the question job-specific: a route for one role may not be the same decision process as a route for another, even if both use a questionnaire with the same name. The illustration supplies no company facts or empirical result; it shows why the sequence of decisions matters to interpretation.
The Guidelines then make the sequence conditional: when the total selection process for a job has adverse impact, the individual procedures within it are analyzed to determine which procedure or procedures have adverse impact. That second, component-level inquiry serves a different purpose. It helps locate where a disparity appears or which step may contribute; it does not replace the process-level question about the combined outcome. A component can show a concerning pattern without its rate being identical to the final selection rate, because the people and decision point under examination differ. Keeping these questions distinct avoids treating one stage’s comparison as a substitute for the result of the complete route.
There is a reason to inspect components: a total-process rate can hide where a divergence begins. Under the Guidelines’ sequence, component analysis helps examine that possibility after the overall process signals impact, subject to applicable exceptions. The practical reading is not to choose one number and ignore the other, but to match each number to its question: the total process describes the combined job-specific outcome; component rates can help locate a contributing procedure. Neither rate, by itself, explains why it arose or establishes that a personality score caused it.
Which applicant pool and decision rule are actually being compared?
A meaningful rate comparison needs a defined job, a defined route into that job, and a specific point at which people were considered and advanced. Applicants should not be silently pooled across unlike jobs or decisions merely because the same personality assessment appeared in each process. A rate for one position may describe a different selection process from a rate for another position, even when titles sound similar or both use the same instrument. The question is whether the people in the denominator faced the same relevant decision under a comparable rule.
There is an important qualification: the EEOC’s Questions and Answers to Clarify and Provide a Common Interpretation of the Uniform Guidelines on Employee Selection Procedures says that, where more than one route leads to the same kind of employment decision, the total selection process can encompass the combined results of those routes. Its example addresses external applicants and internal candidates pursuing the same kind of decision. This means neither automatic pooling nor automatic separation is sound. The process definition controls: routes that culminate in the same kind of job decision may need to be considered together, while distinct jobs or decisions should not be collapsed without a reason. The point is to preserve the actual process boundary, not to choose whichever grouping gives a more favorable ratio.
For interpretation, ask what job was under consideration; whether the figures combine internal and external routes or other entry paths; what stage the reported rate describes; who counted as considered at that stage; and what rule determined advancement. Also ask whether the assessment was one screen among several or itself the decision point. These questions make the comparison intelligible and help distinguish an assessment-stage rate from the combined outcome across routes. They do not require an applicant to reconstruct the employer’s complete dataset or independently inspect protected-group records. An individual can ask what role the assessment played and what kind of comparison is being described, while recognizing that group-level analysis depends on process records beyond one person’s report.
This boundary matters even when the arithmetic is correct: changing the job, route, stage, or advancement rule changes what population the rate represents. The Uniform Guidelines Q&A supplies a way to define the combined process when routes serve the same kind of decision; it does not authorize combining dissimilar decisions or establish what information a particular employer must disclose. The useful standard is a comparison that follows the real job-related process, with its pooling choices made explicit.
How can the score-use rule change who advances?
The assessment’s name does not fully describe its employment effect. A score can be combined with other information, used to rank applicants, grouped into bands, or treated as a threshold that an applicant must cross. These are different decision rules: the instrument produces a result, while the employer’s rule determines how that result affects who continues. To understand a personality assessment’s role, a reader needs both pieces—the score and the rule that turns it into an employment decision.
The abstract for Personality Testing in Personnel Selection: Adverse impact and differential hiring rates reports a study of 398 bus-operator candidates. It says that potential adverse-impact and differential hiring-rate issues could depend on the selection decision method and the designated group assessed. That is a qualified result from a particular candidate sample and a study focused on decision methods. The accessible record is an abstract; the full text was unavailable, so it does not support claims here about exact group outcomes, effect sizes, specific score rules tested, or which method produced a particular pattern. It also does not establish how personality tests generally affect applicant groups. Its bounded contribution is to show why the method of using assessment results belongs in the question alongside the instrument itself.
Conceptually, a compensatory rule allows a stronger result on one relevant measure to offset a weaker result on another when forming an overall score. A noncompensatory threshold, by contrast, requires a person to meet a minimum on a specified measure regardless of stronger results elsewhere. Ranking orders people relative to one another, so advancement may depend on how many positions are available and where a person falls in that ordering. A banding rule groups scores into ranges treated as sufficiently similar for some subsequent decision. These definitions describe possible approaches; they do not reconstruct the bus-operator study’s methods. Each can produce a different set of people advancing from the same initial scores because each applies a different mapping from results to decisions.
For someone reading a report after an application, the practical question is therefore not only “What did this score mean?” but “How was it used?” Ask whether the score was combined with interview or experience ratings, whether a minimum score excluded candidates, whether applicants were ranked, or whether scores were grouped into bands. Ask at what stage that rule operated and whether it determined advancement or only informed later review. The answers help explain the path from a reported result to a decision; the instrument label alone cannot supply that account.
A method-related difference in one sample does not show that changing a rule will reduce impact in another job, nor that a new rule would preserve the same job performance. Those conclusions need evidence about the local applicant pool, the actual rule, and the work purpose it serves. The study abstract supports attention to decision method and designated group; it does not prescribe a universally fair scoring method or establish a conclusion about any individual personality report.
Sources: Personality Testing in Personnel Selection: Adverse impact and differential hiring rates
What job evidence would support the particular score use?
A job-related claim has to connect the interpretation of a personality score to important elements of the particular job where the score affects selection. In *Section 15: Race and Color Discrimination*, the EEOC illustrates this with a personality test used to identify management potential: if the test has significant disparate impact, the employer must validate its relationship to important elements of manager job performance. The example makes the claim more specific than “this test measures leadership” or “the provider says this trait matters.” It asks what work element the interpretation is intended to inform, and what evidence connects the two in this job. The example states a standard of inquiry; it supplies no validation study for a real instrument or employer.
A provider’s trait description tells a reader how the provider names or explains a scale. It does not, by itself, show that the interpretation is useful for deciding who should advance in a given role. Likewise, evidence that a measure produces consistent scores addresses consistency under specified conditions; it does not alone establish that using those scores for a selection decision bears on important work elements. These are different links in the argument: what the score is said to represent, how dependably it is measured, and whether the employer’s particular interpretation and use are supported for the job. The EEOC’s management-potential example focuses on that last connection. It neither rules out personality measures for occupational criteria nor establishes that any named test meets the requirement.
For an applicant trying to understand a report’s role, a focused question is: which important performance element in this job was this score intended to inform, and how did it affect the decision? A concrete answer should identify the work element and the score interpretation being relied on, rather than simply repeating a broad trait label. That question can clarify whether the employer treated the result as context, a screen, or a decisive criterion, but an individual report cannot provide the employer’s substantiation. That would require job and selection evidence outside the report. If the answer remains only a general claim about what a trait sounds like, the job-specific connection has not yet been explained to the reader; that gap is not, by itself, proof that the assessment is invalid or that the employer’s use was unlawful. The performance element should be stated in terms of work the role requires, not inferred from the label “management potential” alone. That distinction lets a reader separate a plausible description of a scale from evidence supporting its consequential use in this selection context.
What might an employer examine after a signal?
Once an employer has located the procedure associated with a concerning pattern, review can ask what role that component actually played and whether changing it could remove the pattern from the overall selection process. *CM-604 Theories of Discrimination* describes examining components and modifying a procedure or its use to remove impact; the EEOC’s *Section 15: Race and Color Discrimination* also discusses a less discriminatory alternative that meets the employer’s needs. This is a sequence for investigating options, not a promise that one change will solve the problem: first identify the operative component and decision rule, then examine what happens if that use is modified, and assess what remains if the employer continues to rely on it.
If the component remains in use, the next question is what evidence supports that use for the work at issue. The point is to examine the particular procedure and the way its result is applied, rather than presume that the instrument alone explains the pattern. A changed threshold, a different combination with other information, or another method could alter who advances, so the proposed change itself needs to be described precisely. The agency guidance frames employer review; it does not establish which option is appropriate for a particular company, and this reasoning pathway is not a legal checklist or a guaranteed remedy. Its practical value is to keep the review attached to the actual decision mechanism and the job purpose it serves.
A suitable alternative is a comparison against the same work need, not merely a procedure with a lower impact in isolation. The employer would need to consider whether the alternative addresses the relevant job purpose and what evidence bears on its performance for that purpose. A lower observed impact alone does not demonstrate equivalent performance, establish that every substitute is feasible, or settle the other evidence relevant to the decision. Conversely, rejecting an alternative without examining its suitability would leave the comparison incomplete. *CM-604 Theories of Discrimination* and *Section 15: Race and Color Discrimination* make the alternative question part of the employer-side inquiry; neither source identifies a universally applicable replacement. For the report reader, the distinction is useful: “what other procedure or use was considered for the same job need?” is a question about the employer’s substantiation, not something the personal score can answer.
Sources: CM-604 Theories of Discrimination; Section 15: Race and Color Discrimination
How is an access barrier different from a group-rate signal?
A personality assessment can raise a disability-access question even when no group selection-rate comparison is available. The U.S. Department of Justice’s *Algorithms, Artificial Intelligence, and Disability Discrimination in Hiring* guidance gives personality assessments as an example of hiring tests and says they should measure relevant job skills rather than limitations unrelated to the job. The central issue here is the match between what the test requires a person to do and what the employer intends to assess. That inquiry concerns disability access and the testing conditions; it does not require first establishing a demographic rate pattern.
The format can introduce demands that are incidental to the intended skill. A test may require a person to perceive visual material, manipulate an interface, or respond through speech, even though the relevant job skill is something else. Under the ADA.gov guidance, if a testing method screens out a person because of a disability, the employer may need to provide an accessible alternative or make an adjustment so the test measures the skill it is meant to assess rather than the disability-related barrier. An alternative format or adjustment should preserve the relevant assessment purpose; the guidance’s point is not that every testing condition must be identical, but that the method should not measure an unrelated limitation in place of the job-relevant skill.
This is different evidence from a race, sex, or ethnicity rate comparison. Group rates can indicate a pattern in a selection process, but they do not establish that a particular person encountered a disability-related access barrier. In the other direction, a person’s difficulty with an inaccessible format does not by itself establish a group-rate disparity. Both concerns can coexist, or either can arise without evidence for the other. They answer different questions and call for different information: one about outcomes across groups, the other about whether the test’s content and response demands let a person demonstrate the relevant skill.
If the format obstructed your ability to show a relevant skill, ask whether an accessible format or adjustment is available and explain which part of the response method created the barrier. That request focuses on the testing conditions and the intended skill, without claiming that a group-rate analysis has been done. Conversely, not encountering an access barrier does not answer whether selection rates differ across groups. Keeping the inquiries separate makes each concern clearer and avoids treating the absence of one kind of evidence as an answer to the other.
Sources: Algorithms, Artificial Intelligence, and Disability Discrimination in Hiring
What can you reasonably do with your own report now?
First clarify what the result was used for: private reflection, an initial screen, ranking, or another employment decision. If it affected an application, ask what job purpose the score was intended to serve and what rule turned it into a decision, such as a threshold or combination with other information. If the format made it difficult to demonstrate the relevant skill, ask about an accessible condition or adjustment. These are focused questions about your experience and the assessment’s role; they do not establish whether group selection rates differed or whether an employer’s process was fair.
If your remaining question is personal—such as how you tend to plan, collaborate, or handle change—the [Work Pattern Report](/assessment) can turn a vague work-pattern question into observations across its ten continuums. It is a low-stakes reflection aid with no norm, cutoff, or hiring validation, so it cannot evaluate an employer’s fairness, explain an application decision, or recommend a job. It is separate from process records, job analysis, and any legal assessment. For more on reading personality assessments responsibly, see [Personality Report topics](/topics).
Sources: Work Pattern Report
Questions readers ask
Does a ratio above 80% prove a personality assessment is fair?
No. The four-fifths ratio is a screening rule of thumb, not a fairness certificate or legal definition. Other statistical or practical evidence may matter, and the ratio does not establish that a score use is supported for the job.
Can my personality report show whether a company’s assessment had adverse impact?
No. A personal report concerns one person’s responses. Adverse impact concerns group selection rates in a defined employment process, which requires information about who was considered and advanced, how scores were used, and the relevant job and decision stages.
Sources and notes
- Questions and Answers to Clarify and Provide a Common Interpretation of the Uniform Guidelines on Employee Selection Procedures
Official joint-agency technical assistance interpreting the Uniform Guidelines: adverse impact definition, rate and impact-ratio steps, limits of the four-fifths rule, total-process-first approach, and validity evidence expectations when an impactful procedure continues in use.
- Designing an Assessment Strategy
OPM describes the four-fifths method as a widely used rule of thumb for comparing passing or selection rates and explains the basic lower-rate / higher-rate calculation in an assessment context.
- Section 15: Race and Color Discrimination
EEOC guidance gives a personality-test example involving management potential and explains job-performance validation and the possibility of a less discriminatory alternative that meets the employer's needs.
- Personality Testing in Personnel Selection: Adverse impact and differential hiring rates
The accessible abstract reports that results from 398 bus-operator candidates indicated potential adverse-impact and differential hiring-rate issues could depend on the selection decision method and designated group assessed.
- Algorithms, Artificial Intelligence, and Disability Discrimination in Hiring
ADA.gov specifically includes personality assessments among hiring tests and says testing should measure relevant job skills rather than unintended sensory, manual, or speaking limitations; accessible tests or adjustments may be needed.
- CM-604 Theories of Discrimination
EEOC compliance guidance describes modifying components to remove impact from a job's total process and, if impact remains, supporting the procedure with job-related evidence; it also discusses suitable alternative procedures or uses.
- Work Pattern Report
The live product is a 100-item self-report across ten work-pattern continuums, with no norm, cutoff, or validation for hiring; it can support low-stakes reflection on one's own work patterns only.
Apply it to your work
Turn a vague work-pattern question into specific observations
From this guide: If you are using a report to reflect on recurring work friction, separate your own tendencies from conclusions about an employer’s selection process.
A personal report cannot establish adverse impact or explain a hiring decision. It can help you examine how you approach decisions, planning, feedback, collaboration, and change. The Work Pattern Report offers a low-stakes way to put those tendencies into more specific observations. Use it for reflection on your own work patterns, while questions about selection rates and job-related use remain questions about the employer’s process.