Work Samples and Content Validity for Frontline US Hiring

Work sample content validity is the most workable way for a US employer to justify a frontline hiring test, because it asks you to show that the test looks like the job rather than to prove a statistical correlation across hundreds of hires. A warehouse filling 40 roles a year cannot run a credible predictive study, but it can document the tasks that matter and build a short exercise that mirrors them. That is why task-mirroring assessment has become the default for hourly hiring.

Key Takeaways

• Content validity means the selection procedure is a representative sample of the content of the job, a standard set out at 29 CFR 1607.14(C).
• It cannot support a test of a trait or construct such as intelligence, aptitude, personality or judgment. Those need a different validation strategy.
• A criterion-related study needs enough people to reach statistical significance. Content validity has no such sample requirement.
• Physical agility and fitness tests are generally not medical examinations, so long as they measure task performance and add nothing medical such as heart rate or blood pressure readings.
• Federal posture on disparate impact shifted in 2026, but validation records remain the practical defense against state regulators and private suits.

What Content Validity Means Under the Federal Standard

Content validity is a demonstration that the behaviors performed during the test are a representative sample of the behaviors of the job itself. The Uniform Guidelines on Employee Selection Procedures (29 CFR Part 1607) recognize three validation strategies at 29 CFR 1607.5: criterion-related, content, and construct. Section 1607.14(C) sets the technical standard for the content route.

That section says a procedure measuring a skill or ability should closely approximate an observable work behavior, or its product should closely approximate an observable work product. It goes further: the closer the content and context of the test are to real work behaviors, the stronger the basis for content validity. As the test drifts from the work, the claim weakens.

One hard limit is worth knowing first. Section 1607.14(C)(1) says a content strategy is not appropriate for procedures that purport to measure traits or constructs such as intelligence, aptitude, personality, judgment and spatial ability, nor for skills an employee will learn on the job. If your test measures a general mental quality rather than an observable task, you need a criterion study instead.

Why the Content Route Fits Frontline Hiring

Criterion-related validation is sample-hungry in a way that rules it out for most hourly employers. Section 1607.14(B) requires the user to judge how many people a meaningful study needs, and a fairness investigation additionally requires an adequate sample in each group to reach statistical significance. A care provider with 30 staff cannot reach those numbers.

Content validity moves the burden somewhere an employer can meet it: a documented job analysis and a test that visibly does the same thing as the job. The evidence is a description of the work and a record of how the exercise was built from it, not a correlation coefficient. That is also why a short, well-built exercise beats an off-the-shelf battery on paper trail as well as on prediction, a point examined in more depth in IQ Testing vs Work Sample Tests: Research Findings.

How to Build a Work Sample Test That Holds Up

1. Run a Job Analysis First

Section 1607.14(C)(2) expects a job analysis covering the important work behaviors, their relative importance, and any work products they generate. Where a behavior is not observable, the analysis should identify the observable aspects and the resulting product. In practice this means sitting with supervisors and incumbents, listing what the job does in a shift, and marking which tasks are frequent, difficult, or costly when done badly. Keep those notes; they are the foundation of the file.

2. Sample the Task, Not the Trait

Choose the critical behaviors that make up most of the job, then build the exercise from them. A picking test should involve picking. A dispatch test should use a real call script and system screen. The temptation is to substitute something easier to score, such as a logic puzzle standing in for troubleshooting, and that is exactly what the Guidelines treat as weak.

3. Standardize Administration and Score With Anchored Rubrics

Every candidate should get the same instructions, materials, time limit and setting. Scoring should use behavioral anchors, meaning written descriptions of what a 1, a 3 and a 5 look like, rather than a general impression. Anchors are what make two raters agree, and section 1607.14(C)(5) says reliability should concern the user and be estimated statistically where feasible.

4. Train the Raters

Two supervisors scoring the same exercise should land within a point of each other. Get there by having several raters score one sample, comparing results, and talking through disagreements until the anchors are read the same way. Repeat when new raters join. Platforms that handle structured scoring keep the rubric in one place so it does not drift between locations, which is part of what AI skill assessments and a maintained skill tests library are for.

5. Pilot for Adverse Impact Before Rollout

Run the exercise on current staff or a first applicant cohort and check pass rates by group before making it a gate. If one group passes at a much lower rate, ask whether the cause is a genuine job requirement or an artifact of the test, such as reading-heavy instructions for a manual role. Fixing that at pilot stage is cheap; fixing it after 18 months of rejections is not.

Matching the Role to the Right Work Sample

• Warehouse and distribution - Work Sample: Pick 10 items from a pick list, scan, stage on a pallet; Scoring Anchor: Items correct out of 10; scan errors; time against incumbent median
• Food service - Work Sample: Assemble three orders to spec during a simulated rush, including one modification; Scoring Anchor: Spec accuracy per order; modification handled correctly; hand-wash and glove steps observed
• Facilities and janitorial - Work Sample: Clean a defined area using supplied chemicals and a label-reading step; Scoring Anchor: Dilution read correctly from label; sequence followed; area passes a checklist inspection
• Care and direct support - Work Sample: Role-play a resident refusing assistance, then write the shift note; Scoring Anchor: Dignity behaviors present; escalation point identified; note factual and legible
• Light manufacturing - Work Sample: Assemble a small part from a work instruction, then inspect two units for defects; Scoring Anchor: Assembly matches instruction; both defects found; no missed lockout step
• Call center and support - Work Sample: Handle a recorded customer issue, log it in a sample ticket form; Scoring Anchor: Issue correctly identified; policy applied; ticket fields complete and accurate

Every anchor above points at something a second rater could verify from the same evidence, which is the test of a usable rubric. Reliability and validity questions of this kind are treated at greater length in Ensuring Accuracy: Validity and Reliability in Skill Assessments.

Physical Ability Tests and the Disability Line

A physical work sample is generally not a medical examination, which matters because the Americans with Disabilities Act (ADA.gov) restricts pre-offer medical examinations under 42 U.S.C. 12112(d). EEOC guidance places physical agility tests, which measure the ability to perform actual or simulated job tasks, and physical fitness tests, which measure performance of physical tasks such as running or lifting, outside the medical examination category. The condition is that the test adds nothing that could itself be considered medical, and the guidance names heart rate and blood pressure measurement as examples that cross the line.

So a lift-and-carry exercise set at the weight and frequency the job genuinely involves is a work sample. Attach a blood pressure cuff or a range-of-motion screen and you have converted it into something the ADA regulates differently. Keep the two apart, hold any genuine medical screen until after a conditional offer, and publish how candidates can request an accommodation.

What Changed Federally in 2026, and What Did Not

On June 9, 2026, the Department of Justice Office of Legal Counsel issued a memorandum opinion, Constitutionality of Disparate-Impact Liability Under Title VII, concluding that the EEOC's disparate-impact guidelines, expressly including the Uniform Guidelines, are unconstitutional and inconsistent with Title VII. The opinion implements Executive Order 14281, and the EEOC has aligned its enforcement posture, treating ordinary selection procedures as presumptively job-related.

What did not change matters more for planning. The Uniform Guidelines remain codified at 29 CFR Part 1607. An OLC opinion is executive-branch legal advice: it does not bind federal courts, amend Title VII, or overrule Griggs v. Duke Power Co., 401 U.S. 424 (1971). Disparate-impact liability is statutory at 42 U.S.C. 2000e-2(k), and private plaintiffs keep an independent right to sue. State law is moving the other way, with California, New York, New Jersey and Illinois codifying disparate-impact protections. Federal posture has receded; exposure has relocated rather than disappeared, and validation records remain the practical defense.

Related Resources

• AI skill assessments
• skill tests library
• automated reference checks
• HR and pre-employment resources
• IQ Testing vs Work Sample Tests: Research Findings
• Ensuring Accuracy: Validity and Reliability in Skill Assessments
• Uniform Guidelines on Employee Selection Procedures (29 CFR Part 1607)

Frequently Asked Questions

Can I use the same work sample across several locations?

Yes, if the job analysis shows the same major work behaviors at each site; the Guidelines allow jobs with substantially the same major work behaviors to be grouped for validity purposes. Where a site genuinely differs, such as a cold-storage facility, document the difference and adjust the exercise.

Does a work sample replace reference checking?

No. A work sample shows what a candidate can do under observation, while references speak to how someone worked over time. Running structured automated reference checks alongside a scored exercise gives two independent sources rather than one.

Do we still need to measure adverse impact after the 2026 federal changes?

Yes. Private plaintiffs retain an independent right of action under Title VII, several states have codified disparate-impact protections, and agencies rescinding related rules expressly preserved the data-retention expectation on the basis that the same data can evidence intentional discrimination. Stopping measurement removes your evidence, not your exposure.

Where can I read more on assessment design?

Righteo publishes practitioner material on testing and hiring process design in its HR and pre-employment resources.

This article is general information for US employers and is not legal advice. Confirm your obligations with employment counsel in the states where you hire.

Content Validity Rewards the Unglamorous Work

Content validity rewards employers who do the unglamorous work of writing down what a job involves and then testing that. It needs no large sample, it produces evidence a reader can follow, and it tends to predict performance better than a generic battery. Start with one high-volume role: document its critical tasks, build a 20-minute exercise from them, write anchors for each score point, and pilot it on current staff before it gates a single applicant.