Are you testing legally?
Testing is a widely used method of assessing the work personality or job skills of an applicant or employee. From the standpoint of what is legal, testing is also one of the more nettlesome phases of human resource management.
You should never assume that you must use testing. Many employers engage in no formal testing whatsoever, and flourish without it. Therefore testing is like a pension or profit sharing plan - - - you are under no obligation to develop and implement one, but once you do numerous legal issues come into play, and there are good reasons to have it.
The numerous laws governing employee relations, from federal and state legislation and regulations prohibiting discrimination to common law theories such as invasion of privacy and defamation, challenge employers to adopt testing practices that both carry out the testing objectives and avoid legal liability when an examinee raises a challenge in court.
A few cases have addressed the question of whether the use of personality tests had an adverse impact on protected individuals. See, e.g., Colbert v. H-K Corp., 4 Fair Empl. Prac. Cas. (BNA) 529, 530 (N.D. Ga. 1971) (intelligence and personality tests upheld as reasonably related to job performance) 8
Generally, adverse impact claims have risen in the context of ability and aptitude tests. Although public sector employers such as fire departments and police departments frequently appear in this litigation, private sector employers can also be involved. The Seventh Circuit held that an employer's aptitude test had a disparate impact on Hispanic job applicants because there was no significant correlation between an applicant's test score and his ability to perform the duties of an entry-level manager. Melendez v. Illinois Bell Tel. Co., 79 F.3d 661, 665-669 (7th Cir. 1996). The plaintiff's expert in this case testified that the aptitude tests could "predict a person's job performance only 3 percent better than chance alone." Id. at 665.
There are different methods by which a job applicant or employee may show adverse impact. The most common approach is the comparison between the percentage of minority applicants successfully passing a personality or aptitude test and the percentage of majority applicants successfully passing the tests.
In order to establish a prima facie case, the applicant or employee must establish adverse impact with a significant sample of minority applicants. If the sample size is small, the court may reject the statistic because a slight shift of a few scores would greatly alter the overall disparity. However the court may accept a small sample if the complaining employee has other evidence of discrimination such as a history of minority exclusion. See, e.g., Rogers v. Int'l Paper Co., 510 F.2d 1340, 1356 (8th Cir.), rev'd on other grounds, 423 U.S. 809 (1975) (small minority sample enough to establish adverse impact when evidence of past discrimination shown).
Along with an adequate sample size, the complaining employee must also establish that the adverse impact is "substantial." Using the applicant-statistic approach, for example, the court examines the ratio of the majority pass rate to the minority pass rate. See 29 C.F.R. § 1607, i.e., the Uniform Guidelines of Selection Procedure (the "Guidelines").
The Guidelines were developed by the Office of
Federal Contract Compliance and the Equal Employment Opportunity
Commission. They do not have the force of law but are given deference by
most courts. The Guidelines provide that to establish an adverse impact
claim, the minority pass rate must be less than 80 percent of the majority
pass rate. 29 C.F.R. § 1607.4(D) ("[a] selection rate for any race, sex, or
ethnic group which is less than four-fifths … (or 80 percent) of the rate
for the group with the highest rate will generally be regarded by the
federal enforcement agencies as evidence of adverse impact, while a greater
than four-fifths rate will generally not be regarded as evidence of adverse
Once an employee establishes a prima facie case of disparate impact under Title VII, or the Human Rights Act, the employer must show that the test is job-related. Does it measure what it is supposed to measure? Does the measurement relate to the job? See, e.g., Graffam v. Scott Paper Co., 870 F. Supp. 389, 399-403 (D.Me. 1994) (selection criteria chosen to implement reduction in force were job related, where psychologist testified that all criteria used except that of specific technical skills described job behaviors required in managerial and technical jobs), aff'd, 60F.3d 809 (1st Cir. 1995). In order to meet the job-relatedness requirement, the employer must show the scored test relates to the job at issue, i.e., the test must be validated. In other words, the test scores must correlate with some other measure of what the test is used for.
There are three recognized methods by which an
employer may show validation: (1) criterion validation; (2) content
validation; and (3) construct validation. 29 C.F.R. § 1607.5 (A).
Criterion validation compares success on the test with success on critical or important job duties. Criterior related tests measure traits or characteristics deemed relevant to future job performance, e.g., an intelligence test. According to the Guidelines, a criterion validation study is "technically feasible" when: (1) a substantial number of individuals are included in the validation study; (2) a range of scores on the selection procedure and job performance measures can be obtained that is sufficiently representative of the normally expected ranges; and (3) reliable and valid measures of job performance are available. See 29 C.F.R. § 1607.5. If these three conditions are present, an employer is more likely to obtain a statistically significant result.
Employers can use content validation when the test measures skills a job requires; for example, a typing exam for a typist position would prove content validation. See 29 C.F.R. § 1607.14 (C). This form of validation is often utilized for aptitude testing, but is not applicable to personality testing because such tests do not measure skills or job knowledge.
Construct validation necessitates a relationship between satisfactory job performance and a specific construct or trait. Examples of construct-validated tests would be those measuring as honesty and integrity, or human relations skills required to perform effectively as a customer service representative, as well as accurate measurement of the trait by the test. See 29 C.F.R. § 1607.14(D). This method of validation is the most appropriate for personality testing, because it focuses on the link between the particular trait and projected performance on the job.
To minimize liability with respect to testing, consider these points:
Is the test really necessary? What is the objective your organization is attempting to achieve? If you believe the test can enhance your screening process, you should document the business necessity for the test.
Ask your test developer to hold you harmless and indemnify your organization for discriminatory challenges to the test and any damages that may result from the test’s usage. This indemnification should include paying the employer’s reasonable attorneys’ fees incurred in defending any legal challenge to a test.
Even if the test being used has been validated, maintain statistics to monitor for adverse impact. If adverse impact occurs, take steps to eliminate the impact or consider alternative screening measures.
Administer the test on a provisional basis while monitoring the test for adverse impact. If impact occurs, follow the procedure in the preceding paragraph before the test as a formal screening device.
The federal Americans with Disabilities Act, prohibits all testing at the pre-employment stage that pertains to mental and physical disabilities or conditions. This means that mental and physical examinations are illegal before the offer of hire is made (although highly sensitive safety positions are exceptions, and drug testing is permitted).
What legal issues must an employer consider, when administering tests to applicants or employees? Although there is a formidable assortment of laws out there, basic notions of fairness apply whenever you engage in testing. Here is a checklist of questions. Y(yes) or N(no) are indicated where appropriate:
Drug Testing (GO TO THE DRUG TESTING SECTION)