The line in the posting

An employer deletes the bachelor's-degree requirement from a job posting. Who ends up working there?

“There are many ways people can learn the skills necessary to perform the job. And if an employer decides everyone must have a four-year degree, when, in fact, many people could successfully perform the job without one, that can operate as a discriminatory barrier that leads to an unjustified, disparate impact.”

Jenny R. Yang, former Chair, U.S. Equal Employment Opportunity Commission, PBS NewsHour, July 7, 2026

States, the federal government, and many private employers keep striking the bachelor's-degree requirement from job postings. The movement has a name — skills-based hiring: screen on what the work entails, not on a credential that might proxy for the capability. The case is easy to make. A degree requirement excludes many adults before anyone reads a résumé, and it excludes some groups at higher rates than others, because attainment is unevenly distributed.

Changing the posting may or may not change hiring. That depends on who applies, who gets interviewed, who gets an offer — and who decides. The posting shapes only the first stage: application. But a different applicant pool can still move what follows, so the answer is not obvious either way.

You are the hiring manager. Three dials, one per stage. Set them and watch how far down the process your change travels — and where it stops. Two settings near the bottom of the page produce identical data for opposite reasons, and that pair is worth reaching.

Before you touch anything

The employer deletes the degree requirement from the posting. What happens to the four-fifths ratios among the people actually hired?

Commit to an answer. The controls unlock once you do, and the page will show you what you said next to what the arithmetic does.

The pool moved. The hiring moved a fraction of it.

With the requirement in force, Hispanic adults reach the hired group at 0.52 of the White non-Hispanic rate and Black adults at 0.69 — two of the four groups below the four-fifths threshold. Delete the requirement and the candidate pool reaches 1.00 for each group: the allocative fix does what an allocative fix does.

Among hires the ratios go to 0.66 and 0.78. They rise, and they stop short of the line. The count of groups below four-fifths among hires is 2 before and 2 after. That count is the figure to watch, and it is the fourth card in the strip above.

What that result rests on. The reveal sets the third dial — reviewer preference — to 85%, meaning a candidate without the credential stays competitive at 15% of one who holds it. Attainment rates are Census; the reviewer is modeled. That setting is an assumption rather than a finding, and the dials let you change it.

Applicant pool

0.0%

of applicants hold no bachelor's degree

People hired

0.0%

of hires hold no bachelor's degree

Workforce after 10 years

0.0%

of the cohort holds no bachelor's degree

Below four-fifths among hires

of 4 groups compared with the White non-Hispanic rate

Four scenarios, then the dials

Run them in order. Each sets all three dials at once, and each number on the page updates. The fourth is the one that works, and it reads differently once you have felt the second.

The three dials — the sandbox, once you have been through the four. One dial per stage.

Requirement removed Allocation

Share of openings posted without a degree requirement.

0% of openings

Applicant response Supply

Of qualified people without a degree, the share who learn the requirement is gone and apply.

100% respond

Reviewer preference Valuation

How much reviewers still favor degree-holders once both are in the pile. 0% = the credential carries no weight.

0% preference

Where the composition changes

No bachelor's degree Holds a bachelor's degree

The workforce is a stock, not a flow

Each year about 8.6% of this cohort turns over — the federal rate. Even a hiring change that lands fully takes roughly eight years to move the workforce halfway. Process > snapshot.

Who is in the pool at each stage

How representation is computed, and every underlying number

Two mechanisms, one dataset

Pin a scenario, then move the dials and try to reproduce the same observable result — the two numbers an outside analyst could measure. If two mechanisms produce identical data, that data cannot separate them.

Pinned

Nothing pinned yet.

Current

What to take away

Allocations

How the organization matches people to jobs, projects, and opportunities.

The degree requirement is an allocative rule. It sorts who is eligible before anyone evaluates anyone, and it does so identity-neutrally — which is precisely why it produces disparate impact. Any criterion correlated with identity will. Deleting the line from the posting is an allocative fix, and it acts in this row.

Valuations

How the organization evaluates and rewards the people it lets through.

A reviewer who still favors the credential is valuing it, whatever the posting says. Valuations of merit drive allocations of opportunity, so a preference that outlives the policy keeps sorting candidates after the rule that used to do the sorting is gone. No edit to a posting reaches this row.

DGP determines the fix

Match the intervention to the mechanism that generated the disparity.

An allocative fix addresses an allocative DGP. Applied to a valuative one it widens the pool, leaves the hiring where it was, and the unadjusted gap re-emerges. Matching the wrong solution to the wrong problem wastes resources and can produce new inequities. That is the sequence you just ran.

The three steps a disparate-impact claim runs through

1The complaining party shows the practice causes a disparate impact. That is the four-fifths ratio in the panel above.

2The employer must show the practice is “job related for the position in question and consistent with business necessity.” A degree is a proxy for skill, not a measure of it — this is where a credential requirement is exposed.

3The complaining party may still prevail by naming an alternative employment practice with less impact that the employer refuses to adopt. Assessing the skills directly is that alternative, and it is a change at valuation.

Griggs v. Duke Power Co., 401 U.S. 424, 431 (1971); codified by the Civil Rights Act of 1991, Pub. L. No. 102-166, § 105, at 42 U.S.C. § 2000e-2(k)(1)(A)–(C). The third step follows the standard the statute preserves — another device without a similarly undesirable racial effect that would also serve the employer's legitimate interest, Albemarle Paper Co. v. Moody, 422 U.S. 405, 425 (1975). Four-fifths rule: 29 C.F.R. § 1607.4(D). Substituting a criterion is not the same as validating one — a replacement correlated with what it replaced reproduces the same disparity under a new name, and the burden of showing otherwise sits with the employer.

Nobody in this simulation treated anyone differently. The same rule was applied to each applicant, and it still drove selection-rate ratios below four-fifths — because degree attainment is not evenly distributed, and so long as factors correlated with identity are used to screen, disparate impact will follow. Systemic unfairness does not require discriminatory intent. And from the panel above: the same two numbers can come from a preference that outlived the policy or from applicants who never arrived. Those point to different rows of the matrix and different fixes — structured evaluation, or outreach — and postings-and-hires data cannot separate them. Naming that limit is the difference between choosing a fix and guessing at one.