Legal Issues In Testing
The path leading to screening and selecting competent, motivated employees is filled with pitfalls and landmines. It is obvious that for decades, if not centuries, selection decisions were made by employers on whatever basis an employer chose—the attractiveness of the candidate, his or her resume, the schools that candidates had attended, a recommendation from a friend, handwriting, or any other notion that the employer might have had about what predicted job success. This lack of a systematic approach led to both poor hiring decisions and active discrimination in hiring. At various times in our country’s history, this discrimination adversely affected every conceivable racial, religious, and ethnic minority as well as women and persons with physical and mental handicaps.
Over the past 50 years, a number of social movements concerned with such discriminatory practices, including those focused on minority rights, women’s rights, the rights of persons with handicaps, and others, have promoted the adoption of a variety of legislative and regulatory remedies on the national, state, and local level that have attempted to address these practices. The most far reaching are the US Civil Rights Act of 1964 (especially Title VII), as amended in 1972, the Civil Rights Act of 1991, the American with Disabilities Act of 1990, and the creation of the Equal Opportunities Commission in 1994. These laws and regulations are the root cause of the many legal dangers inherent in the screening and selection of employees. We are not lawyers and this chapter does not offer legal advice; rather, it is intended to provide a context in which to understand the impact of these laws and regulations on assessment and selection. It is imperative that employers seek and use competent legal advice from attorneys who focus in this specialized area of the law and choose an attorney knowledgeable about the various laws that regulate employment actions.
In our opinion, there is little question that the net effect of these anti-discriminatory efforts have been positive and, as a consequence, there is far less active discrimination in hiring than there has been in the past. But what often occurs instead is inadvertent discrimination as a result of using screening measures that produce adverse impact. Adverse impact refers to the outcome of any personnel process that leads to the disproportionate rejection of members of protected groups. It includes psychological tests and any other screening process that lead to proportionately fewer members of various groups being advanced to the next level of selection.
For example, many jobs require a college degree as a condition of employment. Since a higher percentage of minority candidates do not have such a degree, this requirement can have an adverse impact in that a higher percentage of minority candidates are disqualified than Caucasian candidates. Thus, while there is no deliberate intent to discriminate in setting this requirement, the requirement of a college degree is indeed discriminatory. How can such a discriminatory requirement be justified? The best defense for invoking such a requirement is to demonstrate that the requirement is job-related; that is, to show that college graduates perform better on this job than those with less education—having a college degree is a valid indicator or predictor of success. Thus it is incumbent on every employer, in a situation addressing adverse impact, to evaluate whether or not each and every one of the selection criteria are valid; that is, they have demonstrable relationships to job success.
The Role Of Tests In Pre-Employment
Since their development almost a century ago, psychological tests have been advocated as the best way to improve applicant screening and selection. Psychological tests, it is argued, are standardized, objective measures of people, far less open to any bias than other procedures, and more likely to lead to a data-driven, valid, screening and selection process.
However, in the 1971 Griggs v. Duke Power Company landmark case, the US Supreme Court ruled that, since the use of a standardized intelligence test coupled with the requirement of a high school diploma systematically excluded African Americans from consideration, the use of such measures violated Title VII of the Civil Rights Act of 1964. The Court noted that the inclusion of these screening measures was intended to reduce discrimination, but that such “good intentions” were not sufficient; what mattered was the adverse impact on the protected minority group. Thus, if a screening instrument causes adverse impact, it is illegal to use such a test unless there is good evidence to prove that the test is valid!
The Griggs v. Duke Power decision has been augmented by several additional Supreme Court decisions, including Albemarle Paper Company v. Moody (1975) and Connecticut v. Teal (1982), among others. These later decisions supported and extended the earlier prohibition against using any screening device without demonstrable criterion-related validity. Further, tests with the least adverse impact must be used, and adverse impact at any stage of a multiple-stage screening process constitutes discrimination. In another later decision, Washington v. Davis (1976), the Court decided that, since the verbal communication test used by the Washington, DC police force as a screening device had criterion-related validity, it could be used despite its adverse impact.
Measuring Adverse Impact
The 1991 Civil Rights Act established a standard for determining whether or not a selection procedure has adverse or disparate impact. Specifically, a claim of disparate impact requires proof the selection procedure has a significant adverse impact on a protected group, such as minorities. The 1978 Uniform Guidelines on Employee Selection adopted by the Equal Opportunities Commission (EEOC) adopted the 80 percent rule, a test of statistical significance. This rule establishes that, as compared to the group with the highest rate of selection, if less than 80 percent of applicants from any protected class are accepted, there is prima facie evidence of adverse impact. Thus it is important for employers to maintain careful records in order to evaluate whether or not adverse impact has occurred.
There is an implicit paradox inherent in applying the 80 percent rule. Although there is a prohibition against inquiring about race, age, religion, or national origin, it is a consideration of these very factors that must be taken into account in determining the presence of adverse impact. The various laws and regulations are silent as to how to resolve this dilemma. Obviously some determination must be made in establishing the necessary data set, but it is up to the employer to determine the basis for these data. One recommended solution is to obtain this information from each candidate after the selection decision is made, although there are obvious administrative problems inherent in implementing this procedure.
These various Supreme Court decisions highlight the need for employers to validate their selection processes, both to enhance the effectiveness of these processes and as a safeguard against litigation. Validation studies need to be conducted for each class of jobs involved and employers should be aware that the same test profile may not be applicable to every job or job class.
Obviously, there is no way to prevent an unhappy applicant or employee from litigating, but the kind of solid validation provided by the benchmarking process outlined in Chapter Five will provide reasonable grounds for mounting a defense in any such litigation. Employers who fail to conduct such validation research do so at their own risk.
The underlying issue is that real differences exist between various groups. It is well beyond our competence to attempt an explanation for these differences, but they clearly do exist. For example, men have greater upper body strength than women. But the two distributions overlap; there are some women with greater upper body strength than some men. In jobs that actually do require greater upper body strength, such as fire fighters, it would be illegitimate to simply exclude women from the applicant pool (as was once the case). Rather, we must use testing for such strength and select those applicants with the necessary level of upper body strength. This example highlights both the need for objective testing or evaluation of candidates and the need for such testing to be based on a valid job-related requirement.
What is both ironic and paradoxical is that the use of psychological tests, initially heralded as a safeguard against bias, has become the target of accusations of bias. The paradox is that, if employment decisions are not based on objective test results, then they will be based on human judgment, a much more fallible process. But at the same time, we must recognize that the law of unintended consequences again holds sway. Despite the best of intentions, psychological testing can produce adverse impact, and only through validation research can this use be justified. Clearly, there is a need for selecting tests with clear content validity and routinely conducting criterion-related validation studies, through benchmarking or more formal predictive validation.
The preceding information is excerpted from the Technical Manual for JobCLUES® ©Copyright 2005 by Psichometrics International and is therefore protected by the copyright held therein.