Kelly Cahill Timmons*
Abstract
Vendors of pre-employment personality tests promise that their products will help employers identify the ideal candidate for each job in a cost-effective manner, while improving diversity. Some vendors tout the algorithms underlying their assessments, claiming that artificial intelligence will remove unconscious bias from the hiring process. They report that their tests have no disparate impact based on race, sex, or national origin.
What vendors do not address, however, is the impact their personality tests have on individuals with mental disabilities, who are protected from discrimination under the Americans with Disabilities Act. In fact, they claim that they do not have and cannot obtain information on that impact, because the ADA prohibits asking job applicants whether they have a disability.
This Article contends that test vendors and employers must stop dodging the question of the effect of personality tests on individuals with mental disabilities. An adverse effect is likely: most modern personality tests are based on the Five Factor Model of Personality, and psychological studies have demonstrated correlations between the five factors and various mental disabilities. Personality test vendors should research whether and to what extent their products tend to screen out applicants with disabilities—while employers cannot ask about disability, that does not mean that test vendors cannot do so outside the employment context. In the absence of statistical evidence of group impact, applicants with disabilities should be able to assert individually focused disparate impact claims.
Although employers can use tests with a disparate impact on persons with mental disabilities if the tests are job-related and consistent with business necessity, the evidence is mixed regarding how well personality tests predict job performance. Vendors often claim that their tests satisfy the ADA’s affirmative defense because their algorithms compare the personality traits of each applicant to those of existing top performers in a particular job. However, the danger of algorithmic bias in hiring—noted by several legal scholars with respect to race, sex, and national origin—is of particular concern in the disability context. Given the underrepresentation of individuals with mental disabilities in the workplace, seeking to replicate the personality traits of a company’s top performers may perpetuate exclusion and inequality.
The ADA was intended to ensure that individuals with disabilities are not barred from jobs that they can perform, and a low score on a personality test does not necessarily indicate an inability to perform a job’s essential functions. In light of the individualized focus of the ADA, applicants with mental disabilities concerned about being screened out by a personality test should request as a reasonable accommodation an alternative to the test or the opportunity to opt out of the test.
*Associate Professor, Georgia State University College of Law. The author would like to thank Roland Behm and GSU Provost Wendy F. Hensel for their inspiration and insight.