New models of psychopathology do not change intention and use of normal-range personality measures and have little bearing on existing EEOC guidance

Melson-Silimon et al. (Reference Melson-Silimon, Harris, Shoenfelt, Miller and Carter2019) refer to Equal Employment Opportunity Commission (EEOC) Enforcement Guidance (2000) that clarifies when tests might be considered medical examinations and thus prohibited under the ADA. The seven central points of EEOC guidance relate to the original development and use of the test(s) under consideration. Interpretation of these central criteria is not impacted by recent developments relating to the AMPD or other models of psychopathology because they do not change how normal-range personality tests have been developed or how they are used in employment decision making. Tests of normal personality, in the forms currently used in the majority of selection settings (see below), are (1) not administered by health care professionals; (2) not interpreted by health care professionals; (3) not designed to reveal an impairment of physical or mental health; (4) not invasive; (5) not measuring an employee’s physiological responses to performing a task; (6) not normally given in a medical setting; and (7) not used with medical equipment. Having a better understanding of personality dimensionality and taxonomic structure of normal and abnormal personality does not change how normal-range personality tests have been designed or are used in organizations.

Traits and behaviors are not impairments, and not every impairment is a disability

Recent enforcement guidance (EEOC, 2008) specifically states that “traits or behaviors are not, in themselves, mental impairments.” Individual items on normal-range personality tests that inquire about certain behaviors or experiences (e.g., “I get stressed out easily”) are not automatically considered indicators of mental impairments, even if they might be related to impairments for some candidates who answer them in the affirmative. The same is true for high levels of desirable traits, which the focal article states “can be maladaptive at extreme levels” (p. 12). The central question is whether a test was designed to—and actually can—measure the underlying trait at a level of precision where an impairment arises (normal-range personality tests cannot). Moreover, current EEOC guidance makes clear that not every impairment rises to the level of a disability under the ADA: “The determination that a particular individual has a substantially limiting impairment should be based on information about how the impairment affects that individual and not on generalizations about the condition” (EEOC, 2008).

Reviewed case law does not preclude the use of normal-range personality tests pre-offer or abnormal-range personality tests post-offer

Among the four cases reviewed in the focal article, one pertained to the use of a normal-range personality test (Thompson v. Borg-Warner Protective Services Corp., 1996), and three pertained to the use of abnormal-range (clinical) personality tests. The former was judged not to be a violation of the ADA and decided in favor of the defendant. The latter three (Barnes v. Cochran, 1997; Karaker v. Rent-A-Center, 2005; and Soroka v. Dayton Hudson Corporation, 1991) all involved the improper use of clinical personality tests or medical evaluations that were conducted before an offer of employment or promotion had been extended. None of these cases preclude the use of either type of personality test for purposes for which they have been developed or professionally endorsed.

Privacy rights need to be balanced with liability for negligent hiring

Normal-range personality test questions do not reach the same threshold of invasiveness as actual medical procedures (such as the drawing of blood, urine, or breath) or even questions found on medical history questionnaires (e.g., medication use). Melson-Silimon et al.’s (Reference Melson-Silimon, Harris, Shoenfelt, Miller and Carter2019) review of the Soroka v. Dayton Hudson Corporation case relating to privacy rights has no connection to the ADA, but it is based on right to privacy considerations in state law.

The applicability of this case to properly implemented normal-range personality testing is very limited. First, the test used in this case was not a normal-range personality test, and thus it provides a poor comparison for the majority of tests used in pre-employment, select-in scenarios. Second, the items of the 1943 edition of the MMPI that were cited as invasive are not found on current versions of the same test—and certainly not on normal-range personality tests. Third, and most importantly, the defendant in this case made other, even more elementary mistakes in using personality tests for hiring purposes, which provided the basis for plaintiffs to sue under California law—failing to conduct a job analysis. Using any psychological assessment tool (be it a personality test, cognitive ability test, or an in-person interview) without first establishing job requirements and specifications is an egregious professional error, independent of whether the tool poses a potential threat under EEOC-related laws and regulations. Job analysis provides a basis for establishing job relatedness, which was relevant in this case, as the plaintiff alleged that questions on the MMPI were invasive and not job related.

In other cases, the courts have upheld the use of personality tests in employment decision making, noting that the privacy rights of the individual must be balanced with rights of other employees, the public, and legal responsibilities of the employer to avoid negligent hiring. In McKenna v. Fargo (1978), the court found that the city had significant interest to identify applicants who were not suitable for firefighter jobs and that this justified even intrusive test questions (Shaffer & Schmidt, Reference Shaffer and Schmidt1999). Of course, the balance of the organizational interest versus applicant privacy consideration will shift even further in the presence of job analyses that demonstrate the link between personality traits and essential functions of the job; in cases where test questions are significantly less invasive than those found on outdated measures of psychopathology; and in occupational settings where the consequences of negligent hiring are even more significant (e.g., pilots, law enforcement officers; see Butcher, Front, & Ones, Reference Butcher, Front, Ones and Butcher2018).

Personality measures can be used in both select-in and screen-out applications

Melson-Silimon et al.’s (Reference Melson-Silimon, Harris, Shoenfelt, Miller and Carter2019) discussion of how and why normal-range personality measures could potentially run afoul of the ADA is based on a fundamental assumption that is unsupported and mischaracterizes how these tests are commonly used: “Because I-O psychologists generally utilize a ‘bottom-up’ approach with personality tests, such that low standing on desirable traits and high standing on undesirable traits are used to ‘select out’ applicants (Carrigan, Reference Carrigan2007), we assume this type of selection here” (p. 124). The source for this assumption cited in the focal article (Carrigan, Reference Carrigan2007) actually does not address the issue of select-in versus screen-out uses of personality measures. Although it briefly notes the historical use of screen-out approaches prior to World War II, the majority of that article actually addresses and endorses the use of pre-employment tests “with the intent of selecting the most qualified applicants from among applicant pools” (Carrigan, Reference Carrigan2007, p. 39) to “forecast an applicant’s likelihood to succeed on the job” (p. 40). This characterization is consistent with the observations of the present authors—based on several decades’ experience with personality measures in work settings—of how normal-range personality measures are typically used. When used in employee selection, these tests are much more likely to be used with “select-in” rather than “screen-out” approaches.

When organizations attempt to identify job relevant personality characteristics among applicants (e.g., conscientiousness, stress tolerance), high-scoring individuals are selected for employment. In talent management systems, relevant personality characteristics (e.g., ambition, interpersonal sensitivity) are measured to identify individuals with high potential. Organizations that utilize developmental approaches relying on strengths and positive psychology (cf. Asplund, Lopez, Hodges, & Harter, Reference Asplund, Lopez, Hodges and Harter2007; Peterson & Seligman, Reference Peterson and Seligman2004) also focus on high scores on desirable traits. The same is true for predicting leadership, entrepreneurship, or innovation. Before and during employment, applications of normal-range personality measures aim to identify the best or the most suitable, highlighting differences among higher scoring individuals. The purpose and capability of normal-range personality inventories does not include flagging impaired or disordered individuals.

Of course, some organizations also screen out for job relevant personality characteristics (e.g., irresponsibility, dishonesty). In pre-employment screening, job-related measures such as integrity tests or normal-range conscientiousness scales can be utilized. As EEOC guidance makes clear, integrity tests and normal-range conscientiousness-type measures are appropriate for pre-offer screening purposes. Screening out may also call for assessing job-relevant personality pathology, especially for public safety and high stakes jobs. In that case, an assessment strategy that includes abnormal-range personality measures (e.g., negative affect, disinhibition) is appropriate after a conditional job offer. In sum, understanding how personality measures are typically deployed in organizations is essential for providing guidance on appropriate use of normal- and abnormal-range personality measures.

Normal-range personality measures are not curvilinearly related to job performance

The focal article questions the usefulness of personality assessments by stating: “recent studies suggest that even high levels of ‘desirable’ traits like high conscientiousness can be maladaptive at extreme levels” (p. 124), and “…at its extremes, conscientiousness reflects obsessive-compulsive tendencies … that reduce the likelihood of these positive outcomes, including multiple dimensions of job performance (Carter et al., Reference Carter, Dalal, Boyce, O’Connell, Kung and Delgado2014; Le et al., Reference Le, Oh, Robbins, Ilies, Holland and Westrick2011)” (p. 127). This characterization of research on potential nonlinear relations in the personality domain, especially around generalizably valid conscientiousness measures, is very selective and troubling.Footnote ¹

The linearity of conscientiousness–work behavior relations has been documented in many more samples than the two sources cited in the focal article. Robie and Ryan (Reference Robie and Ryan1999) found no evidence of nonlinearity for conscientiousness–job performance relations (five samples). Walmsley (Reference Walmsley2013; Walmsley, Sackett, & Nichols, Reference Walmsley, Sackett and Nichols2018) examined 123 validity studies and reported a “general lack of meaningful curvilinear effects for each Big Five scale in relation to overall job performance” (p. ii). For conscientiousness, the quadratic term provided only small incremental validity over the linear term for overall-, intrapersonal-, interpersonal-, leadership-, and work skill performance (ΔR ² was .034, .032, .032, .027, and .031, respectively). The most recent examination (Nickel, Roberts, & Chernyshenko, Reference Nickel, Roberts and Chernyshenko2019; four samples), concluded that high conscientiousness had no detrimental effects on health, well-being, job satisfaction, citizenship behaviors, or counterproductivity. This work also illuminates the unusual findings of the Carter et al. (Reference Carter, Dalal, Boyce, O’Connell, Kung and Delgado2014) study cited by the focal article, which seem to relate to the use of an ad hoc, de facto abnormal-range conscientiousness measure.Footnote ² The accumulated evidence to date is in line with the psychometrically informed expectation that normal-range conscientiousness and other personality trait scales do not assess the abnormal ranges of the respective constructs well and consequently are not related to decrements in desirable or increases in undesirable work behaviors. There is no downside to scoring high on conscientiousness. There is no downside to selecting high-scoring individuals in selection or talent management programs when normal-range conscientiousness measures are used.