One concept every HR and Talent Acquisition leader should understand is adverse impact.
Adverse impact occurs when an employment practice results in substantially different selection outcomes for members of a protected group, even when there was no intention to discriminate.
That last point is important. Adverse impact does not necessarily mean somebody deliberately discriminated against candidates, nor does identifying a statistical disparity automatically establish unlawful discrimination. It means the outcomes of a selection process show a meaningful difference that warrants further investigation.
For organizations using the Predictive Index, the critical question therefore isn't simply, "Does PI cause adverse impact?" A better question is: "How are we using PI within our overall selection process, and what do the outcomes tell us?"
Imagine an organization has 1,000 applicants for a group of similar positions. After a particular stage in the selection process, 50% of applicants from one group progress compared with 35% from another.
One commonly referenced method for initially examining this difference is the four-fifths, or 80%, rule.
Dividing the lower selection rate by the higher selection rate gives us:
35% ÷ 50% = 70%
Because 70% is below 80%, the difference could indicate adverse impact and should be investigated.
The 80% rule should not be treated as a simple legal pass-or-fail test. Sample size, statistical significance, the applicant population, the specific selection procedure and other factors can all matter. Organizations should involve appropriate HR, legal and industrial-organizational expertise when evaluating potential adverse impact.
But the concept gives us an important starting point: look at the outcomes of your selection process, not simply the intentions behind it.
Consider a typical hiring process:
Application → Résumé Screening → PI → Interview → Final Selection
If an organization discovers that one demographic group is progressing at a substantially lower rate than another, knowing the final hiring numbers isn't enough. We need to understand where that difference is occurring.
Perhaps the disparity begins during résumé screening. Perhaps it emerges during interviews. Perhaps hiring managers are interpreting behavioral data inconsistently. Perhaps the organization has created an unnecessarily restrictive PI Job Target. Perhaps someone has introduced an arbitrary "fit score" or cutoff that automatically eliminates candidates.
It could also be a combination of several factors.
This is why discovering adverse impact should be the beginning of an investigation rather than the conclusion of one.
The goal should be to examine the complete selection process and understand which components are contributing to the difference in outcomes.
The Predictive Index Behavioral Assessment helps organizations understand an individual's behavioral drives and how those drives may influence the way they naturally approach work.
That information can be extremely useful in hiring.
But PI does not tell us everything we need to know about a candidate.
A behavioral profile does not tell us whether someone has the technical expertise required for a job. It does not tell us whether they have successfully managed a team, negotiated a contract, built a financial model or developed a sales strategy. Those are competencies, skills and experiences that need to be assessed through other parts of the selection process.
This is one of the most important principles for using PI responsibly:
PI should provide one lens into the candidate, not become the entire hiring decision.
A strong selection process combines behavioral information with other job-relevant evidence, including structured interviews, competencies, experience, technical knowledge, work samples and other appropriate measures.
One of the most important safeguards when using PI is establishing the behavioral requirements of the job before evaluating candidates.
A PI Job Target should answer the question:
"What behaviors does this job actually require for someone to perform successfully?"
It should not answer:
"What kind of person would the hiring manager like?"
That difference is significant.
Suppose a hiring manager says they want someone who is "highly dominant, outgoing and fast-paced." Before translating those preferences into a Job Target, we should understand why the job requires those behaviors.
Does the person genuinely need considerable independence and decision-making authority? Does the role require frequent persuasion and interaction with unfamiliar people? Is the environment genuinely fast-moving, with constantly changing priorities?
If the answer is yes, those behavioral requirements can be connected to the realities of the job.
If the answer is simply, "That's what our best salesperson looks like," we should dig deeper.
The objective isn't to replicate an existing employee's personality. It is to understand the behavioral demands created by the work itself.
Another potential problem arises when organizations begin using behavioral characteristics as proxies for capabilities.
For example:
High Dominance does not mean someone is a good leader.
High Extraversion does not mean someone is a strong communicator.
High Patience does not automatically mean someone is empathetic.
High Formality does not automatically mean someone possesses excellent attention-to-detail skills.
These behavioral drives may influence how someone approaches certain activities, but they don't demonstrate competency.
Consider "strong communication skills," a requirement found in countless job descriptions.
What does that actually mean?
For a salesperson, it might mean persuading skeptical customers. For an engineering leader, it could mean explaining complex technical concepts to non-technical colleagues. For a people manager, it might mean listening carefully and delivering difficult feedback. For someone working in compliance, it could mean producing extremely precise written communication.
Each requires communication competency, but the behavioral demands could be very different.
This is why organizations shouldn't simply convert competencies into PI factors.
Instead, ask:
"What does successful performance of this competency actually look like in this job?"
That conversation allows the hiring team to distinguish between the competency required and the behavioral environment in which that competency needs to be demonstrated.
One area that deserves particular attention is the use of hard behavioral-match thresholds.
Imagine an organization decides:
"Candidates who don't achieve at least an 8/10 Job Target match won't progress to interview."
The organization has now done more than use behavioral information to inform a hiring decision. It has created a selection rule that directly determines who is allowed to progress.
That distinction matters.
A candidate who doesn't closely match a behavioral target isn't necessarily incapable of performing the job. The difference may indicate that some elements of the role require greater behavioral adaptation or stretch.
That information can create valuable interview questions.
Instead of automatically eliminating the candidate, the hiring team can investigate.
How has the candidate operated successfully in similar environments before? What strategies have they developed? Is there evidence that they can perform the required behaviors? Does their experience demonstrate that they can successfully manage the behavioral demands of the position?
Used this way, behavioral data makes the interview better, rather than simply making the candidate pool smaller.
This is one of the most valuable ways to integrate PI into a broader selection process.
Suppose the Job Target indicates that the position requires considerable independence, but the candidate's behavioral pattern suggests a more collaborative decision-making style.
Rather than concluding that the candidate is a poor fit, the hiring team can explore the difference.
They might ask the candidate to describe situations in which they had to make significant decisions independently, operate without clear direction or challenge an established approach.
Now the behavioral data has generated a hypothesis, and the interview provides an opportunity to gather evidence.
The same approach can be used for pace, social interaction, structure, risk tolerance and other behavioral demands.
The question becomes less about "Does this person match?" and more about "What should we explore?"
That is a much more sophisticated use of behavioral data.
If an organization using PI identifies potential adverse impact, the first response shouldn't be to immediately conclude that PI caused it — nor should it be to assume that PI couldn't have contributed to it.
The right response is to investigate.
Start by looking at candidate progression through each stage of the selection process.
Where does the disparity first appear?
If it occurs before candidates complete PI, the assessment clearly isn't responsible for that particular difference. If the disparity increases significantly after PI results are introduced into the decision process, that deserves closer examination.
Then examine how PI is being used.
Review the Job Target and how it was created. Examine any match scores, thresholds or cutoffs being applied. Look at whether hiring managers are using PI consistently. Determine how heavily behavioral fit is weighted compared with competencies, experience and other evidence. Look for situations where behavioral characteristics may inadvertently have become proxies for capabilities that should be measured independently.
The organization may determine that the assessment isn't creating the disparity at all. Alternatively, it may discover that the way the organization has chosen to use the assessment needs to change.
Either outcome is useful.
Where concerns are identified, organizations should document the analysis and any remediation, and engage qualified employment counsel and appropriate assessment expertise where necessary.
Organizations shouldn't wait for an adverse-impact concern before examining their hiring data.
A more responsible approach is to monitor candidate progression periodically and understand what happens at each major selection stage.
For organizations using PI, that means knowing not only who was ultimately hired but how behavioral assessment information influenced progression and decision-making.
If candidates are screened out before interview based on behavioral match, that deserves particular scrutiny. If PI is primarily being used to generate interview questions and provide another source of job-relevant information alongside competency and experience data, its role in the process is different.
This kind of monitoring doesn't weaken the value of behavioral assessments.
It strengthens it.
It forces organizations to be clear about what PI measures, what it doesn't measure and how the information should contribute to a defensible, evidence-based hiring process.
The objective of using PI shouldn't be to find people whose behavioral profiles perfectly match a Job Target.
Human performance is more complicated than that.
Instead, organizations should understand what successful performance requires, define the behavioral demands of the role, assess the competencies and experience candidates bring, and then use behavioral data to understand how naturally each candidate may approach those demands.
Sometimes the candidate with the closest behavioral match will be the strongest candidate.
Sometimes they won't.
That's exactly why we use multiple sources of information.
When organizations discover potential adverse impact, the same principle applies. Don't immediately blame an individual assessment, but don't automatically exclude it from scrutiny either.
Follow the data through the entire hiring process. Understand where differences in outcomes emerge. Examine how each selection tool is being used. Make sure the criteria influencing decisions are genuinely connected to the requirements of the job. And change the process when the evidence tells you that you should.
PI can be a powerful component of a hiring process, but its value comes from using behavioral data for what it was designed to do: help us understand people and the behavioral demands of work, while enabling human beings to make better-informed decisions.