Meeting of May 16, 2007 - on Employment Testing and Screening
I appreciate the opportunity to present my views and experience concerning the practical advice and counseling available to employers considering skills testing or other objective selection procedures. As I will explain, advising on this topic presents real questions and difficult issues for employers. Because the vast majority of employers considering skills testing are looking for non-discriminatory and objective ways to find qualified employees, they would benefit from additional guidance from the Commission on how to do so.
When I counsel clients in this arena, I have to explain that test validation presents a classic “rock and a hard place” quandary. The 1978 Uniform Guidelines on Employee Selection Procedures (“Guidelines”) represent the “rock” and the “hard place” is really a “trilemma” of “hard places.” To start with, employers cannot technically comply with the standards set forth in the Guidelines in their current, obsolete form, because industrial psychologists will not likely be persuaded to abandon the state-of-the-art validation technology in favor of a decades-old methodology presented in the Guidelines. The Guidelines are frozen in time and are not going to change – more on that later. Therefore, employers are left with choosing among three “hard places”: (1) investing substantial resources to develop objective testing procedures which can be validated under modern, state-of-the-art, mathematically sound techniques (at one extreme); (2) adopting wholly subjective decision-making procedures (at the other extreme); or (3) wandering into an uncertain middle-ground, where the employer seeks to use an objective skills-based test, but the scope, projected usage, or budget priorities won’t support the investment of the time and money required for a state-of-the-art validation project. Unlike many choices employers face in deciding whether to comply with regulations, each of these choices carries very real downsides, and no one choice presents a clearly preferable option.
Any discussion of test validation has to begin with a mention of the Guidelines. The Guidelines provide highly technical standards for developing and validating tests, which have not been updated since 1978. While the Guidelines, when written, were intended to be “consistent with generally accepted professional standards,”2 as early as 1982, the General Accounting Office noted that the Guidelines had already been criticized as outdated by the American Psychological Association (“APA”). The APA publishes the technical standards used in the Guidelines and in the industry, and has updated the technical standards several times since the adoption of the Guidelines in 1978. The obvious disconnect is that the industrial psychology field now uses current, state-of-the-art validation standards most recently revised in 1999, while the Guidelines remain frozen and still reflect the 1978 standards. That is a regulatory “rock” facing any employer who considers testing.3
The first “hard place” – validation using state-of-the-art techniques – is simply not available to many employers. Validation that is fully current and compliant with modern industry standards is a time consuming and expensive proposition. Skilled professionals applying modern standards conduct thorough job analyses, complex portability studies, and assessments of statistical bias, and based on that highly sophisticated analysis, craft appropriate tests. No doubt state-of-the-art validation is worth the time and money for tests involving highly sophisticated inquiries or large numbers of applications – yet many, many legitimate opportunities for skills testing do not fall into those categories.
For good reason, employers can feel relatively confident that a test validated under the latest technical standards would survive scrutiny. However, I am careful to advise clients that undergoing this validation process does not eliminate the possibility of costly litigation; rather, it provides employers with a potentially legally valid defense to adverse impact that may or may not relieve them of associated liability. Moreover, they do not preclude the very expensive battle-of-the-expert type litigation that we sometimes see in testing cases.
The second “hard place” represents the other end of the spectrum – making employment decisions without any objective measure, and often based upon wholly subjective criteria. Unstructured, subjective employment decisions do not lend themselves to validation under the Guidelines or current validation standards. If validation of skills testing is economically inaccessible for some employers or some jobs, some employers will simply not test at all. Yet, unstructured, subjective employment decisions are obviously not desirable. Indeed, a number of substantial class action cases are built upon a foundation of attacking subjective decision-making.
Finally, employers may elect the third “hard place” – using an objective “professionally developed ability test” to make selection decisions without advance validation that would meet modern, state-of-the-art validation standards. Employers considering this approach must be willing to walk into the statutory “space” suggested by Section 703(h) (“professionally developed ability test” not adopted as subterfuge for discrimination) and 703(k)(1)(A)(ii) (“. . . job related for the position in question and consistent with business necessity”) without a prior state-of-the-art validation study, and hope it can defend its test based on the language of the statute should a challenge arise. The best hope for this middle-ground approach are those tests seeking to measure job skills and duties, rather than those seeking to uncover more ephemeral qualities like personality, character traits, or aptitude. All tests must, of course, be job related in order to comply with Title VII – the challenge will be defending their use in the face of adverse impact. Employers in this middle-ground may craft a job-related test, or may purchase a “professionally developed” off-the-shelf test, or may work with a professional to develop a test short of full validation. The problem with this approach is that if a tests result in adverse impact, experts will be ready to pounce on the lack of “validation.” I must advise clients that without advance validity studies, carrying the burden of proving job-relatedness and business necessity is not a walk in the park.
Addressing this “trilemma” must begin with an acknowledgment that a general mistrust of tests seems to be a prevalent attitude among some in the equal employment opportunity field. This attitude is due for re-examination. While tests no doubt are subject to abuse (for which the disparate treatment doctrine is available), they do provide employers with tools to make objective, non-discriminatory selection decisions. How can that be a bad thing? Tests enable employers to comply with the spirit, purpose, and letter of Title VII – to make employment decisions free from consideration of race or any other protected characteristics. I would urge the Commission, and my colleagues in the equal employment opportunity community generally, to give facially valid, piece-of-the-job tests the benefit of the doubt rather than the intense scrutiny that requires mathematical proof of a technical and scientific level.
In keeping with that spirit, I invite the Commission to consider the problem and to craft guidance on how employers can comply with Title VII under the middle-ground approach previously discussed.4 All employers engage in a cost-benefit analysis when thinking about implementing selection procedures. For many employers, state-of-the-art validation is not the most effective operational decision. Indeed, the words in Title VII –“job related” – are not as demanding and exacting as the Guidelines or modern validation techniques suggest. Title VII allows employers to use “professionally developed ability test[s],” so long as those tests are “not designed, intended or used to discriminate” based upon a protected characteristic. What is missing is guidance on how to do that.
While the Guidelines interpret Title VII, they do not expressly address the phrase “professionally developed ability test.” There is room for additional guidance regarding this specific provision of Title VII, which should address compliance techniques short of state-of-the-art validation, or “short form validation,” for lack of a better word. Indeed, the Guidelines themselves provide that employers who cannot or need not employ full validation techniques can justify a selection procedure “in accord with Federal law.”5 This further guidance will allow employers to avoid the all-or-nothing approach that promotes using subjective selection procedures, and could encourage less elaborate objective techniques consistent with the purpose and requirements of Title VII.
Certainly, there are types of test validation that require application of more rigorous technical standards. I would suggest that criterion-related or construct validity not be undertaken without professional advice and sophisticated mechanisms to assess validity. However, content validation, which essentially requires verification that a test measures the actual skills and duties necessary to perform a specific job, lends itself well to less intense or onerous standards. Additional EEOC guidance need not be a departure from settled law regarding content validity; rather, it could be based on existing factors that measure content validity, such as those outlined in the often-cited Guardians case from the Second Circuit and followed in many jurisdictions. These factors include: (1) job analysis; (2) competent test construction; (3) job skills and duties reflected in test content; (4) correlation between job skills and duties and test content; and (5) a scoring system that correlates to job success. In a nutshell, if a test looks like it is testing skills necessary to do the job, employers ought to be able to defend whatever adverse impact occurs with something short of mathematical certainty and exacting precision. Recall that the employer is objectively seeking qualified employees with piece-of-the-job testing strategies and is doing so without regard to the characteristics protected in Title VII.
Additional guidance would not undermine the Guidelines or contemporary validation techniques. To the contrary, validation pursuant to mathematically sound standards still offers a “safer harbor” to employers who use them, as compared to those employers who, for pragmatic reasons, use a less scientifically developed selection procedure. While courts make clear that they are not bound by the Guidelines, they do routinely give deference to the Guidelines, and do look favorably on employer defenses if an employer elects to pursue state-of-the-art test validation techniques. Moreover, there is no reason that EEOC guidance could not be developed with testing professionals who could suggest middle-of-the-road validation strategies.
Section 703(h) of Title VII provides, in pertinent part, that it shall not be “an unlawful employment practice for an employer to give and to act upon the results of any professionally developed ability test provided that such test, its administration or action upon the results is not designed, intended or used to discriminate because of race, color, religion, sex or national origin.” The Guidelines provide a great deal of technical specificity regarding validation procedures, without providing any simplified process for validation. Notably, the Guidelines do not address the specific phrase, “professionally developed ability test” in Section 703(h) of Title VII, leaving room for further guidance from the Commission regarding that phrase.6 This additional guidance should provide employers with tools to craft valid, objective tests to measure concrete job skills and duties, but not job aptitude or personality.
Additional guidance from the EEOC would be consistent with existing case law regarding test validity. While many courts, including the Supreme Court in Griggs v. Duke Power, noted that the Guidelines are entitled to deference, courts also address the validity of selection procedures by considering and interpreting the phrase “professionally developed ability test” not expressly addressed in the Guidelines. 7 Under Title VII, if a facially neutral testing procedure results in adverse impact, the procedure must be “job related for the position in question and consistent with business necessity” to withstand scrutiny.8 While courts have frequently looked to the Guidelines to determine whether a selection procedure is job-related and consistent with business necessity, the Guidelines should not provide the only method to determine compliance with Title VII, particularly because the Guidelines contain obsolete technical validation standards.9 Indeed, while many courts afford the Guidelines deference, a few courts have determined test validity without reference to the Guidelines.10
The Court in Griggs noted that “[p]roponents of Title VII sought throughout the debate to assure the critics that the Act would have no effect on job-related tests. Senators [co-managing] the bill on the Senate floor issued a memorandum explaining that the proposed Title VII ‘expressly protects the employer’s right to insist that any prospective applicant, Negro or white, must meet the applicable job qualifications. Indeed the very purpose of [T]itle VII is to promote hiring on the basis of job qualifications, rather than on the basis of race or color.”11 If the Guidelines are seen as completely occupying the field of testing under Title VII, and if the Guidelines make validation practically inaccessible for many employers seeking to implement facially neutral selection procedures, one can ask whether they are promoting or detracting from a fundamental purpose of Title VII. Stated otherwise, is our fear of adverse impact effectively pushing employers towards subjective selection techniques, a place often thought of as a breeding ground for stereotypes and preconceptions?
Certainly, Title VII seeks to encourage employers to use neutral job-related criteria to make employment decisions. Indeed, the Griggs court noted that “[n]othing in the Act precludes the use of testing or measuring procedures; obviously they are useful. What Congress has forbidden is giving these devices and mechanisms a controlling force unless they are demonstrably a reasonable measure of job performance. . . . What Congress has commanded is that any test used must measure the person for the job and not the person in the abstract.”12
In order to accomplish this end, Congress expressly permitted the use of a “professionally developed ability test,” provided that it is not intended or used to discriminate. As used in Title VII, this phrase is not an end-run around the job-relatedness and business necessity requirements. As the Supreme Court noted in Connecticut v. Teal, “reliance on [Section] 703(h) as offering the employer some special haven for discriminatory tests is misplaced. . . . After examining the legislative history of [Section] 703(h), we concluded that Congress, in adding [Section] 703(h), intended only to make clear that tests that were job related would be permissible despite their disparate impact.”13 Now, however, proof of job-relatedness, via mathematically sound validation standards, has become so onerous that it has become a practical impossibility for many employers. While “professionally developed ability test[s]” are not outside the rubric for validation procedures set forth in the Guidelines, there is room for additional interpretation of that phrase. This further guidance could provide the middle ground between the “special haven for discriminatory tests” and the overly-burdensome, state-of-the-art validation standards.
There are three types of validation methods – criterion validation, content validation, and construct validation. The Guidelines explain the general and technical standards for all three types of validation in great detail. Both criterion and construct validation require a complex analysis of intangible aspects of job performance. Content validation measures the concrete relationship of a test to the actual skills and duties required of a particular job. Though it is difficult to imagine how criterion or construct validation could be simplified, content validation lends itself more easily to a simplified process.
The court in Guardians Association of the New York City Police Department v. Civil Service Commission of the City of New York distilled the requirements for content validation under the Guidelines into five elements.14
The first two concern quality of the test’s development: (1) the test-makers must have conducted a suitable job analysis, and (2) they must have used reasonable competence in constructing the test itself. The next three attributes are more in the nature of standards that the test, as produced and used, must be shown to have met. The basic requirement, really the essence of content validation, is (3) that the content of the test must be related to the content of the job. In addition, (4) the content of the test must be representative of the content of the job. Finally, the test must be used with (5) a scoring system that usefully selects from among the applicants those who can better perform the job.15
Although the Guardians court “distilled” these five elements from the Guidelines, application of the factors need not rely on technical validation standards. While many employers cannot afford to hire professionals trained in state-of-the-art validation techniques, these employers might be able to apply the factors outlined in Guardians to develop or purchase an objective, job-related selection procedure.
Indeed, the Guardians court noted that adherence to validation standards in all cases can act as a bar to employers’ ability to utilize tests. “The danger of too rigid an application of technical testing principles is that tests for all but the most mundane tasks would lack sufficient validity to permit their use.”16 The court explained that even a typing test for a typist position has “substantial, though not complete, validity.”17 Because a typing test results in quantifiable results, and because it tests precisely for what is required on the job, a typing test may appear valid even without necessitating numerous experts’ application of technical standards to deem it so. However, as the court noted, “[t]yping is one of the few activities that a test-taker can perform in virtually the same manner as he will be required to perform on the job. That is obviously an ideal testing situation, but not one that is frequently available.”18
While a typing test presents an obvious example of a test capable of facial validation, many types of content-specific tests are capable of validation without application of technical standards. The factors described in Guardians present an excellent rubric for this type of facial validation. Because state-of-the-art validation is simply not an option for many employers, there should be a way to use a “professionally developed ability test” without having to hire highly-skilled professionals or look to the outdated Guidelines for comfort. As the Guardians court and many others have noted, “the Guidelines should always be considered, but they should not be regarded as conclusive unless reason and statutory interpretation support their conclusions.”19
This sentiment leaves open the possibility that the Commission can – and should – develop additional guidance interpreting the phrase “professionally developed ability test” consistent with the five factors outlined in the Guardians decision. A simplified approach to content-validation based on the five factors will allow many employers to do what they want but cannot afford to do – comply with the requirements of Title VII to “give and act upon the results of any professionally developed ability test provided that such test, its administration or action upon the results is not designed, intended or used to discriminate” based on protected characteristics.20 This part of Title VII is available for facially valid and sensible skill and qualification testing, and, I would submit, with some guidance from the Commission, should be made accessible for objective and rational selection devices and procedures – a core value in Title VII.
Millions and millions of selection decisions are routinely made around the country. If some employers are willing to invest effort in making those decisions in an objective, race- and gender-neutral fashion through tests which seek out necessary skills, I suggest that the Commission should reach out and meet those employers “half way” by making the “hard place” between scientific validation and unstructured subjectivity a more comfortable place to be.
1. I am the Chair of the Employment Law Practice Group of Wilson Sonsini Goodrich & Rosati. I practice in the firm’s Palo Alto, California office. I have practiced in the employment law field for over 30 years, and had the privilege to serve as a Commissioner of the E.E.O.C. from 1984-1987. In my practice, I advise and represent employers in all aspects of employment law.
I would like to thank my colleague, Tamara Fisher, for her research and assistance with preparing this statement.
2. 41 C.F.R. § 60-3.5(c).
3. Commission staff does interpret Section 60-3.5(c) to mean that they may apply the Guidelines as well as “current” generally accepted standards. See Part 4, Section 1.B of the Regional Attorneys’ Manual.
4. Indeed, it is clear that the Commission has contemplated providing additional guidance in the EEOC Compliance Manual. The Manual currently reserves two chapters – one on the Guidelines and one on “Selection Charges Which Are Not Covered by” the Guidelines – both of which are designated as “Future Sections” that have not yet been published. Several sections in the Compliance Manual reference these “Future Sections.” The occasion of this hearing may be a good time to draft those sections, providing additional guidance to Commission staff and employers who wish to use selection procedures compliant with Title VII.
5. 41 C.F.R. § 60-3.6(b)(2).
6. Interestingly, the 1966 version of the EEOC’s Guidelines did explicitly interpret the phrase “‘professionally developed ability test’ to mean a test which fairly measures the knowledge or skills required by the particular job or class of jobs which the applicant seeks, or which fairly affords the employer a chance to measure the applicant's ability to perform a particular job or class of jobs. The fact that a test was prepared by an individual or organization claiming expertise in test preparation does not, without more, justify its use within the meaning of Title VII.” Griggs v. Duke Power, 401 U.S. 424, 434, n.9 (1971).
7. 401 U.S. at 433-434.
8. Section 703(k)(1)(A)(i).
9. In 1982, the Uniform Guidelines were “criticized as outdated. The American Psychological Association's Committee on Psychological Tests has informed the federal enforcement agencies that in its view ‘the Guidelines reflect a reliance on and use of measurement theory that does not represent the current state of research and theory in psychological testing.’” Report by United States General Accounting Office, EMPL. PRAC. GUIDE (CCH) ¶ 5062, at 3288 (July 30, 1982); see also A SCIENTIST-PRACTITIONER'S VIEWPOINT ON THE ADMISSIBILI TY OF BEHAVIORAL AND SOCIAL SCIENTIFIC INFORMATION, 5 Psychol. Pub. Pol'y & L. 194, 197 -198 (1999) (“many would argue that the research in scientific psychology should deem the guidelines obsolescent”).
10. See, e.g., Washington v. Kroger Co., 671 F.2d 1072, 1077 (8th Cir. 1982) (math test for grocery checkers facially valid); Eison v. City of Knoxville, 570 F. Supp. 11, 13 (E.D. Tenn. 1983) (physical tests for police officers valid).
11. Id. at 434.
12. Id. at 436.
13. 457 U.S. 440, 452 (1982) (emphasis in original).
14. 630 F.2d 79, 95 (2d Cir. 1980).
16. 630 F.2d at 90.
17. Id. (emphasis added).
19. Id. at 91.
20. Title VII, Section 703(h).
This page was last modified on May 11, 2007.
Return to Home Page