Sunday, August 25, 2013

Kroger and Kronos: Chaos and Disorder

In classic Greek mythology, Kronos (also known as Cronus) was the leader of the first generation of Titans. Cronus was usually depicted with a sickle or scythe and the Greeks considered Cronus a cruel and tempestuous force of chaos and disorder.

Kronos, the company, is a U.S.-based workforce management software and services company. According to the company, tens of thousands of organizations in more than 100 countries - including more than half of the Fortune 1000 - use Kronos to control labor costs, minimize compliance risk, and improve workforce productivity.

In August 2006, Kronos acquired Unicru, Inc., a company specializing in software used to assess and hire hourly workers. At the time of the acquisition by Kronos, Unicru had as customers more than 140 leading companies and brands, including SuperValu, Kroger, Toys "R" Us, Best Buy, CVS, Borders, Lowe's, Caribou Coffee, and Marquis Healthcare.

The Unicru assessment consists of a number of statements, to which an applicant must answer “strongly disagree,” “disagree,” “agree,” or “strongly agree.” It includes statements such as the following: “You have confidence in yourself”; You are always cheerful”; “You try to sense what others are thinking and feeling”; “You always say whatever is on your mind”; and “It is easy for you to feel what others are feeling.”

Kroger, Kronos and the Unicru assessment are being investigated by the Equal Employment Opportunity Commission (EEOC) for compliance with labor and employment laws, including the Americans with Disabilities Act (ADA). The investigation has been ongoing for more than five years and has generated a number of district court and appellate court decisions as Kronos has sought to avoid disclosing information about the Unicru assessment and its impact on persons protected by the ADA.

Cloning Employees and Institutionalizing Biased Hiring Practices

According to Kronos, the Unicru assessment is an artificial intelligence test that uses neural networks to “learn” the characteristics of a customer’s “best” employees.  As stated by Kronos’ Chief Scientist and the developer of the Unicru assessment, Dr. David Scarborough, in chillingly Orwellian terms, "[o]ur system allows you to clone your best, most reliable people."

First used for engineering and industrial applications during the mid-1980s, neural networks evolved from early artificial intelligence research. Modeled on the function of the human brain, a neural network attempts to imitate human reasoning. Large amounts of data are fed into the network, which looks for relationships and reaches conclusions.

"There are a couple of dangers," states Jai Shekhawat, CEO of Chicago-based Fieldglass Inc., which develops software for managing workers. "Is something a correlation--a predictor--or merely a coincidence? At best, [these methods] are complementary to human judgment, not a substitute for it."

Notwithstanding such dangers, Kronos customers like Kroger are substituting this “coincidence” for human judgment. Based on the prospective employee's answers on the application, the Unicru assessment categorizes the applicant as red, green or yellow. In most cases, red is usually an automatic discard, or, as Dr. Scarborough stated “[m]anagers are strongly discouraged from hiring first quartile (“red”) applicants …”

There is no evidence that the Unicru assessment determines whether an employer’s hiring practices are biased or discriminatory. For example, if the Unicru assessment had been utilized fifty years ago, many companies’ “best” employees would have the personality traits of white males – persons of color, women and those with disabilities need not have applied.

The Unicru assessment embeds and industrializes existing stigma, bias and discrimination in the hiring process. As stated by Cynthia Dwork and Deirdre K. Mulligan in a recent Stanford Law Review article:
While automated decisionmaking systems “may reduce the impact of biased individuals, they may also normalize the far more massive impacts of system-level biases and blind spots.” Rooting out biases and blind spots in big data depends on our ability to constrain, understand, and test the systems that use such data to shape information, experiences, and opportunities.
As a “blind” tool that “learns” from the employer, the Unicru assessment replicates the existing bias of the employer and applies it on a massive scale. All applicants have their test responses fed through a discriminatory filter that is the Unicru assessment (a filter that is biased both on its own and in conjunction with its “learned” behavior). Hiring decisions are being made by Kroger and other Kronos customers based on this deeply flawed process.

Illegal Medical Examination

The ADA prohibits the use of pre-employment medical examinations. At the pre-offer stage, an employer, like Kroger, is only entitled to ask about an applicant's ability to perform the essential functions of the job. The ADA's prohibition against pre-employment examinations seeks to ensure that the applicant's disability is not considered prior to the assessment of the applicant's qualifications.

EEOC guidance provides a seven-factor test for analyzing whether a test or procedure qualifies as a “medical examination,” including:
  • whether the test is designed to reveal an impairment of physical or mental health such as those listed in the Diagnostic and Statistical Manual of Mental Disorders (“DSM”); and
  • whether the test is interpreted by a health care professional.
According to the guidance, the presence of any one of the seven factors is enough to support a finding that the test is a medical examination and the Unicru assessment meets the two factors listed above. 

Since the Unicru assessment is based on the five-factor model (FFM) of personality it meets the first factor listed above. As set out in previous posts -  ADA, FFM and DSM and Employment Assessments are Designed to Reveal an Impairment - assessments based on the FFM are designed to reveal an impairment of mental health, such as those listed in the DSM.

As to the second factor, whether the test is interpreted by a health care professional, the individuals who developed the Unicru assessment are psychologists, most of whom are members of the APA. In developing the Assessments, the psychologists establish the rules by which the assessments are to be interpreted (i.e., how the responses to the questions are to be scored, including whether the applicant receives a green, yellow or red rating).

According to the APA Model Act for State Licensure of Psychologists, “[t]he practice of psychology includes … (a) psychological testing and the evaluation or assessment of personal characteristics, such as intelligence; personality; cognitive, physical, and/or emotional abilities; … [and] (f) provision of direct services to … groups for the purpose of enhancing … organizational effectiveness, using psychological principles, methods, and/or procedures … for making decisions about the individual, such as selection …”

EEOC guidance states that psychologists are among the “variety of health professionals [that] may provide documentation regarding psychiatric disabilities” for ADA purposes. Accordingly, the psychologists who developed the Unicru assessment are "health care providers" for purposes of the ADA.

(Not) Walking the Talk

Kroger's Policy on Business Ethics states:
We are committed to a policy of equal opportunity for all associates without regard to race, color, religion, gender, national origin, age, disability or sexual orientation.
Kroger has six core values: Honesty; Integrity, Respect; Diversity; Safety; and, Inclusion. In a June 13, 2011 press release announcing the appointment of Kroger's chief diversity officer, Kroger's CEO is quoted as saying:
“Diversity is a core value at Kroger. We take our commitment to diversity seriously, both because it is right and because it makes us better at our business. When our decision-making is inclusive and reflects the diversity of our customers, we make better decisions.”
For job applicants with mental illness, there is no respect, no inclusion, no diversity, no honesty and no integrity. In the more than twenty years since passage of the ADA, there has been little positive movement in de-stigmatizing mental illness in the workplace (please see Mental Illness and Issues of Employment). People with mental illnesses identify employment discrimination as one of their most frequent stigma experiences. In its use of the Unicru assessment, Kroger, wittingly or not, continues the disturbing pattern of employment discrimination against citizens of the United States with mental illness.

Failing Customers and Shareowners

Kroger's Policy on Business Ethics also states:
As a retailer providing millions of Americans with their daily food and as a publicly owned company, The Kroger Co. has a special obligation to comply with the law and deal ethically with customers, suppliers, associates, and shareowners. 
Psychiatric medications are among the most widely prescribed and biggest-selling class of drugs in the U.S. In 2011, Americans spent $18.2 billion on antipsychotics to treat depression, bipolar disorder and schizophrenia, $11.0 billion on antidepressants and $7.9 billion on treatment for ADHD, according to IMS Health, which tracks prescription-drug sales. These three categories of prescription drug sales accounted for approximately 11.6% of all prescription drug sales in the U.S. for 2011

Kroger is the fifth-largest pharmacy operator in the United States, operating retail pharmacies in over 1,948 stores. During fiscal 2011, Kroger pharmacists filled over 146 million prescriptions at a retail value of approximately $7.3 billion. Assuming 11.6% of Kroger prescription drug sales were for antipsychotics, antidepressants and ADHD medications, prescription drugs for persons with mental illness accounted for approximately $847 million of Kroger prescription drug sales in 2011, some two-thirds of the amount of Kroger’s operating profit for that year.

Persons who have their prescriptions filled at Kroger, their family members and other loved ones also shop at Kroger for other products and services. Those persons, their family members and other loved ones provide a material percentage of Kroger’s overall revenue each year. How does Kroger repay this customer loyalty? By utilizing an unlawful pre-employment assessment to eliminate from consideration for employment persons with mental illness.

Why should persons with mental illness, their family members and other loved ones continue to shop at Kroger? Good question. 

Kroger's continuing use of the Unicru assessments calls into question Kroger's "special obligation to comply with the law and deal ethically" with its shareowners, As previously noted, Kroger and Kronos have been engaged in litigation with the EEOC for more than five years over legality of the Unicru assessment. To be precise, the five years of litigation have primarily addressed the unwillingness of Kroger and Kronos to provide information requested by the EEOC in order to conduct its investigation into the Unicru assessment. Two appellate courts, the latest in September 2012, have ruled decisively in favor of the EEOC and its right to investigate a broad set of nationwide and historical data from Kroger and Kronos.

At anytime over the past five years, Kroger could have ceased using the Unicru assessment, if only as a risk mitigation strategy for its shareowners. As noted in the Challenges to Pre-Employment Assessments posting, in July 2011, CVS and the Rhode Island Civil Liberties Union (ACLU) entered into a voluntary settlement addressing the ACLU’s complaint challenging CVS’s use of a pre-hire questionnaire that the ACLU claimed could have a discriminatory impact on people with certain mental impairments or disorders. Pursuant to the settlement agreement, CVS agreed to permanently remove the questions at issue from its online application.

Each day Kroger continues to use the Unicru assessment, there are thousands more potential plaintiffs with claims against Kroger. By now, the aggregate number of potential plaintiffs numbers in the millions - with each job applicant over the past 5+ years having a number of claims against Kroger.

Under the ADA, Kroger may use a third party like Kronos to undertake the assessment of Kroger job applicants. The use of a third party, however, does not insulate Kroger from any claims arising from the assessment usage. Under the ADA, Kroger is responsible (and liable) for any failures on the part of Kronos and the Unicru assessment to comply with the provisions of the ADA.

Any comfort Kroger or its shareowners take in the indemnification provided by Kronos should be tempered by the recognition that such indemnification may prove illusory. Kronos and its insurers may not have the capital necessary to indemnify Kroger and its shareowners for all claims arising from Kroger's continuing use of the Unicru assessment. Please see Damages and Indemnification Challenges for Employers.


  1. The Five-Factor Model of personality is a normal-range personality model. Measures designed to assess FFM traits are explicitly NOT designed to assess impairment or mental disability. They are not prohibited by the ADA.

    1. In her seminal review of the personality disorder literature published in 2007, Dr. Lee Anna Clark stated that “the five-factor model of personality is widely accepted as representing the higher-order structure of both normal and abnormal personality traits.” Dr. Clark was one of the eleven members of the Work Group that selected the Five Factor Model of personality serves as a basis for categorizing and diagnosing personality disorders in the Diagnostic and Statistical Manual of Mental Disorders (DSM-5).

      There is a rich literature concerning the FFM as a structural model for describing and understanding disorders of personality, including those within the DSM. As of 2002, there were over fifty published studies relating the FFM to personality disorders.

      Since that time, quite a number of additional studies have expanded on this research base and provided further empirical support for understanding the DSM personality disorders in terms of the FFM domains. For example, in a study published in 2003 titled “The five-factor model and personality disorder empirical literature: A meta-analytic review,” the authors analyzed data from 15 other studies to determine how personality disorders are different and similar, respectively, with regard to underlying personality traits. In terms of how personality disorders differ, the results showed that each disorder displays a FFM profile that is meaningful and predictable given its unique diagnostic criteria.

      In a study published in March 2012 by the Journal of Abnormal Psychology, the authors tested whether the summation of relevant personality trait scores for borderline personality disorder, using data derived from a measure of the FFM, would result in a construct that corresponds to the DSM-IV (the then current version of the DSM). The study found that DSM-IV and FFM borderline personality scores were significantly correlated and generated highly similar patterns of relations.

      By its design, a test that measures so-called “normal” personalities or seeks to identify individuals with “stable personality traits” screens out persons who may not possess “normal” personalities and/or “stable” personality traits – persons like military veterans suffering from post-traumatic stress disorder, college students diagnosed with bipolar disorder and mothers who have major depressive disorder – persons whom the Americans with Disabilities Act (ADA) are designed to protect.

      Pre-offer employment tests using the Five Factor Model are discriminatory because they constitute unlawful pre-offer medical examinations and they screen out persons with mental disorders, both of which violate the ADA. In Karraker v. Rent-a-Center, the court held that the use of the Minnesota Multiphasic Personality Inventory (MMPI) to evaluate applicants for promotion constituted an unlawful “medical examination” under the ADA because the MMPI “is designed, at least in part, to reveal mental illness and has the effect of hurting the employment prospects of one with a mental disability.” Similarly, pre-offer employment tests utilizing the Five Factor Model are designed to reveal mental illness and requiring particular scores on tests based on the Five Factor Model will have the effect of hurting the employment prospects of a person with a mental disability. Like the MMPI, and as evidenced by its adoption in DSM-5 (as noted above), the Five Factor Model is a widely-accepted and frequently-used means for categorizing and diagnosing certain mental disabilities.


Because I value your thoughtful opinions, I encourage you to add a comment to this discussion. Don't be offended if I edit your comments for clarity or to keep out questionable matters, however, and I may even delete off-topic comments.