The Uniform Guidelines on Employment Selection Procedures, on the Occasion of the Guidelines’ 40th Anniversary

Richard J. Fischer, Guest Author, McLean Consulting

No one who works with tests and other employment practices wants their employer to be sued because of a discriminatory practice, so their acting to mitigate potential employer liability is part of their job, even if not explicitly stated in a job description. So, how a practice, such as test use, hiring or compensation, can be challenged as illegal, and the applicable standard under which a claim could happen, should be common knowledge.

Not always, and what we know, or think we know, can be confounded by at least some misinformation. For instance, discrimination doesn’t have to be intentional. It can be, sure, but in every testing discrimination claim involving the Uniform Guidelines on Employment Selection Procedures (1978) and in many claims brought under Title VII of the Civil Rights Act of 1991 (1991), no intentional discrimination is alleged at all. Each claim was identified statistically based on the test or other employment practice having caused adverse impact—that one or more demographic groups were underselected, paid less, or otherwise disadvantaged at a substantially higher rate than the top performing group as a result employer use of the practice. So, an employer can go out of their way to treat all candidates, employees, or examinees equally and still be held liable under the disparate impact theory of discrimination.

The Uniform Guidelines is especially prone to myth and misinformation. It’s a legal standard, that’s true. Although the Guidelines is not federal law, it is still routinely enforced for discrimination—and with the full force and effect of law, that’s also true, just as if the Guidelines was law. But, the Guidelines is enforced only against Federal contractors found to have discriminated on the contract, not against all employers, and enforced only by the U. S. Labor Department’s Office of Federal Contract Compliance Programs (OFCCP).

The point is that too many employers, and the supposed experts retained to advise those employers, aren’t always sure of how a test or practice can be illegal, or how the Uniform Guidelines fits in, or if it does, or even what the Guidelines is. Others seek legal defensibility by acting on at least some myth, which is a problem because you don’t know it’s myth you’re relying on. Let’s review the Guidelines and how a practice can be challenged as illegal while debunking some of the more common myths along the way to finally put a particular myth to rest.

Maybe. Hopefully.

The Supreme Court Introduces Disparate Impact Theory

The Civil Rights Act of 1964 first made it illegal for employers to discriminate on the basis of traits such as race, religion, color and gender, which created the so-called disparate treatment theory of discrimination. Then came Griggs v. Duke Power Company (1971) and the Supreme Court introducing the disparate impact theory of discrimination as the second way to challenge an employment practice.

To put Griggs in context, many employers had a long history of discrimination. Duke Power, for instance, kept Blacks in labor jobs, which paid the least. After the 1964 Civil Rights Act made intentional discrimination illegal, Duke required non-labor job candidates to have a high school diploma and to pass two general intelligence tests. Anyone could compete for any job, but the three requirements acted to still keep Blacks out of the higher-paying non-labor jobs, so Willie Boyd and 14 other Black laborers sued under Title VII of the 1964 Civil Rights Act. The plaintiffs then picked Willie Griggs as lead plaintiff, because he was the youngest member of the group and thus had the most to gain from a favorable ruling.

And that’s what happened. Several district and appeals court decisions later, the Supreme Court held that a practice with adverse impact was illegal under Title VII unless the practice is shown to “measure the person for the job and not the person in the abstract.“ In other words, a test which adversely impacts Black examinees as a group, for instance, is discriminatory unless the employer can then show that the test is job related and consistent with business necessity (i.e., valid):

The (1964 Civil Rights) Act proscribes not only overt discrimination, but also practices that are fair in form, but discriminatory in operation . . . (G)ood intent or absence of discriminatory intent does not redeem employment procedures or testing mechanisms that operate as built-in headwinds for minority groups and are unrelated to measuring job capability.(Griggs v. Duke Power, 1971)

Griggs is noteworthy for having introduced disparate impact theory, which requires no employer intent to discriminate, and because the Court awarded “great deference” to the Guidelines in the decision—actually, to the then-version of the Guidelines EEOC (1970, 1966) had written and was using at the time. Great deference directs the courts to defer to the Guidelines when assessing validation of the challenged exam, direction, which the Court affirmed in Albemarle Paper v. Moody (1975). So, federal courts reference the Guidelines in a Title VII claim involving a test to show how they assessed validity but decide the case based on Title VII law.

Willie Griggs’ involvement in the case made him a civil rights icon. No one ever heard of Willie Boyd.

The Federal Enforcement Agencies Issue the Uniform Guidelines

In response to Griggs, the four federal enforcement agencies—the EEOC, U. S. Civil Service Commission, which since rebranded as OPM, and the U. S. Departments of Justice and Labor—separately issued their own standards to remedy discrimination. The result was sporadic enforcement, probably not surprising because each agency was applying their own standards, and no agency had actual authority to compel compliance or remedy any discrimination. The four agencies came together in 1978 to co-issue the Guidelines to operationalize the disparate impact theory of discrimination introduced in Griggs.

Still, none of the agencies could enforce the Guidelines. The U. S. Labor Department then stepped forward to legally adopt the Guidelines to administer Executive Order 11246 (1965), which had made discriminating in federal contracting illegal. The Labor Department already had legal authority to enforce discrimination on federal contracts from the Executive Order 11246, which had created the Department’s Office of Federal Contract Compliance Programs. What EO 11246 didn’t provide was an enforcement standard, so OFCCP went through the process detailed in the Administrative Procedure Act (1946) to accept the Guidelines as that standard.

Federal agencies routinely go through the process to issue regulations as binding to help that agency implement federal laws specific to the agency. The proposed regulation becomes binding on the population the agency had preexisting legal authority to regulate, with the agency then able to enforce the now binding regulation against that population for violations. To do that, OFCCP (or any federal agency seeking to formally adopt regulations as binding) advertised the proposed regulation in the Federal Register, allowed time for public comments, and then incorporated those comments into the final regulation. For OFCCP, that population is “covered” federal contractors. A covered contractor has more than 50 employees and holds a contract or subcontract worth at least $50,000.

When enforcing laws against discrimination, OFCCP cites the particular discrimination found that violate 41 CFR 60-3, which is where the Guidelines is cited in the Code of Federal Regulations. The CFR contains the full set of binding regulations; it is the U. S. Code which contains federal laws.

How a Plaintiff Builds a Disparate Impact Claim

Under the Guidelines or Title VII, which makes the very same disparate impact discrimination unlawful against all employers as does the Guidelines, which is enforceable only against federal contractors, a plaintiff makes a “prima facie” showing that a test or employment practice caused adverse impact against a group like women or Blacks. In a Guidelines claim against a federal contractor, that plaintiff can only be OFCCP. Adverse impact doesn’t mean the test or practice is illegal. Not at all. It’s just the initial statistical indicator that discrimination has happened “on its face,” which is what prima facie means.

The burden then shifts to the employer to defend the claim, which requires the employer to “justify” use of the test by showing acceptable validation through any of the four local validation strategies contained in the Guidelines: criterion-related (Guidelines 15.B.9.), content (Guidelines 15.C.6), and construct validity (Guidelines 15.D.8.), or by transporting existing validity of the subject test across users (Guidelines 7.B.). Some employers use multiple strategies to show the challenged test is valid. That’s fine, as long as validity is acceptably shown using at least one Guidelines strategy.

Should the employer provide acceptable validation—and “acceptable” is always open to dispute in any adversarial proceeding given the central importance of validity to both sides in a disparate impact claim—the burden then shifts back to the plaintiff. To prevail, the plaintiff must then offer a suitable alternative that the employer declined to use in favor of using the higher-impact test being challenged. At least, in theory, because case law on suitable alternatives is essentially nil.

A plaintiff not able to offer a suitable alternative loses their case right there. A test the employer can successfully defend is legal and can continue to be used, despite the test’s adverse impact or degree of impact. A test that the employer cannot justify is discriminatory, which happens more than you’d think, which is surprising given the value of validity to any assessment: “The employer is then liable for monetary damages to compensate those screened out by the illegal practice.

In reality, an employer facing a disparate impact challenge first responds by vigorously attacking plaintiff statistical analyses, offering alternative statistics and analyses to make adverse impact go away: No adverse impact, no disparate impact claim. Employers also routinely retain psychometric experts to help convince OFCCP, or a private plaintiff in a Title VII claim, that validation is acceptable technically. When inequitable hiring or compensation is alleged, employers similarly try to make the adverse impact go away, while working to attribute the significant hiring or pay difference to intervening variables like education, experience, bad plaintiff data or inappropriate analyses, rather than to discrimination.

The Disparate Treatment Theory of Discrimination

A disparate impact claim centers on the unintended discriminatory consequences of employer use of the challenged practice based on the practice having caused adverse impact. In contrast, a disparate treatment claim alleges intentional discrimination, as in Ricci v. DeStafano (2009). In Ricci, the City of New Haven feared a disparate impact claim and the associated bad press when no Black examinees scored high enough to be promoted, so they simply voided all results. Ricci and his co-plaintiffs, each of whom would have been promoted based on their test scores, then sued under Title VII. The issue before the Court was whether an employer could discriminate against one group to avoid discriminating against another group.

The Supreme Court eventually said no, but by a 5–4 margin in a case that would seem to be a very clear instance of intentional discriminated based on examinee race. Chief Justice Roberts nicely summarized the issue during oral arguments by asking defense counsel if the city would have tossed out everyone’s results if no Whites had passed.

To Investigate or Not Investigate Suitable Alternatives

One of the more common, let’s debunk this myth in its own section.

When a test causes adverse impact, the Guidelines require the employer to then justify test use by showing: (i) acceptable validation (Guidelines 3.A.) for the test, and (ii) that suitable alternatives were investigated as part of that validation (Guidelines 3.B.). The Guidelines consider the latter so essential that validation is not acceptable on that basis, alone, and the challenged test discriminatory unless the employer investigated suitable alternatives, psychometric evidence aside.

Aamodt (2016) advises accordingly: “Did the employer search for alternatives with equal validity but less adverse impact? . . . If a selection procedure results in adverse impact, but is found to be job related, the Uniform Guidelines still requires employers to search for an alternative method that would have equal validity but less adverse impact.” In brief, a valid test for which no suitable alternatives. Attorney Santiago (2008) similarly notes that: “Ultimately, each employer is responsible for . . . considering suitable alternatives which may lessen any disparate impact.” In their extensive discussion specifically focusing on the kinds of alternatives that could be considered as suitable under Guidelines 3.B., Sady, Dunleavy and Aamodt (2013) state that:

it is incumbent on the employer to . . . (b) evaluate the availability of alternative selection procedures that are equally valid for the intended purpose and result in lower adverse impact. As the UGSEP (3.B.) state

Not necessarily so. It is test validation, alone, that justifies use of the challenged test because no employer has any legal requirement to consider suitable alternatives; at no point in the burden-shifting process in a disparate impact claim is the employer required to proffer a lower impact alternative. That’s the plaintiff’s burden to win the challenge, should the employer justify exam use of the subject exam. That was clear from the start in cases like Albemarle (1975), which predates the Guidelines, and in consistent post-1978 case law. Congress then acted accordingly to codify in federal law, the Civil Rights Act of 1991, that the suitable alternatives requirement rests on the plaintiff. (The suitable alternatives issue is irrelevant in a Title VII disparate treatment claim.)

The legal requirement aside, employers should act to lessen assessment adverse impact to promote test fairness. That’s good psychometric practice. Test developers routinely consider item difficulty, differential functioning, item discrimination and distractor analysis, and use multiple steps to refine and field test draft items, to have an item pool sufficient to construct a valid exam with desired psychometric properties, including minimizing adverse impact. In the same vein, validity is a psychometric requirement, not a legal one; validity is required legally only to justify use of a test with adverse impact.

The Guidelines Can Be Updated, But Won’t Be

A final word. The Guidelines provides that “New strategies for showing the validity of selection procedures will be evaluated as they become accepted by the psychological profession” (Guidelines 5.A.). Again, in theory, because I doubt the Guidelines will ever be revised in practice or that there is a pressing need. Advocates of validity generalization, though, including employment lawyers such as Copus (2006), are especially vocal in their demand that OFCCP allow VG theory under the Guidelines. VG holds that cognitive measure is inherently valid across jobs and job settings (Schmidt and Hunter, 2003) without need for local validation. Under VG theory, there is no need to justify use of any cognitive test because such tests are already, automatically valid, as shown by meta-analysis. So, VG, they claim, would have been included in the Guidelines as the fifth acceptable strategy had it been more fully developed in 1978.

OFCCP has legal authority to enforce the Guidelines but no authority to change or deviate from the Guidelines, so the agency continues to reject VG. Rightly so, because no federal court has ever accepted VG in any of the many Title VII claims in which it has been offered; after hearing evidence from VG experts in each case, courts across circuits have then rejected VG as acceptable each and every time. Local test validation is here to stay to justify challenged test use, whether in a claim brought by OFCCP under the very same Guidelines that was issued in 1978 or in a private plaintiff’s Title VII claim. At least, until a federal court accepts VG, and that will never happen, either.

References

Aamodt, M. (2016). Conducting background checks for employment selection. SHRM-SIOP Science of Human Resources Series. Alexandria, VA: Society for Human Resources Management.

Administrative Procedure Act (1946). Public Law 79–404 (contained in Title 5 of the U. S. Code starting at Section 500; enacted June 11, 1946).

Albemarle Paper v. Moody (1975). 422 U. S. 405, 425.

Civil Rights Act of 1964 (1964). Public Law 88-352; 78 Stat. 241.

Copus, David (2006). An Open Letter Challenging OFCCP on Ability Tests (April, 2006). Morristown, NJ office: Ogletree Deakins Law Firm.

EEOC (1966). Guidelines on Employment Selection Procedures. Washington, DC: Equal Employment Opportunity Commission.

EEOC (1970). Guidelines on Employee Selection Procedures. 35 Federal Register 12, 333-12,336.

Executive Order 11246 (1965). Signed September 24, 1965 (revised January 4, 2002).

Griggs v. Duke Power Company (1971). 401 U.S. 424.

Ricci v. DeStafano (2009). 557 U.S. 55.

Sady, K., Dunleavy, E., and Aamodt, M. (2013). UGESP Series Post #5: Suitable alternatives and reasonable effort: What is an alternative and what is reasonable? Washington, DC: DCI Consulting.

Santiago, L.A. (2008). Employment tests and selection procedures: Are you following EEOC’s best practices? Holland and Knight Labor, Employment and Benefits Information Sheet. New York, NY: Law Firm of Holland and Knight.

Schmidt, F.L. and Hunter, J.E. (2003). History, development, evolution and impact of validity generalization and meta-analysis methods. In K.R. Murphy (Editor), Validity generalization: a critical review (pp. 288-350). Erlbaum: Mahwah, NJ.

Uniform Guidelines on Employment Selection Procedures (1978). Federal Register, 43 (August 25, 1978), pp. 38290-38315; Code of Federal Regulations, 41 CFR 60-3

About Richard J. Fischer, PhD

Richard Fischer was OFCCP’s first psychometric expert, analyzing psychometric and statistical evidence to provide a defensible expert opinion on the legality of federal contractor tests and selection practices, helping to recover significant compensation for hiring discrimination victims. Now working with the Department of Homeland Security supporting Federal Emergency Management Agency human capital programs, Rich holds the PhD in Educational Measurement and Statistics (Psychometrics) from Columbia University, and an MA in Instructional Technology from Columbia University’s Teachers College.