Minimalize Your Legal Risks While Targeting Your Best Hires

September 17, 2015

by Greg Barnett, PhD

 

Target recently agreed to settle an EEOC complaint that cost them a whopping 2.8 million dollars. Whenever a large company like Target is involved in a lawsuit of this nature, it is worth taking notice and understanding what went wrong.

The EEOC complaint focused specifically on the company’s hiring for non-exempt, upper-level professional roles. The investigation found sufficient evidence that three of Target’s pre-employment assessments disproportionately screened out applicants based on race and gender. One of the tests also violated the Americans with Disabilities Act as it was deemed to be a form of pre-employment medical exam (with some reports suggesting that it was an assessment interpreted by psychologists).

While Target claims no intentional wrongdoing, the primary fault was found in the company’s use of assessments that weren’t accurately measuring characteristics that were clear necessities for performing the job or failing to document how the assessment tools were measuring job required skills, abilities, etc.

While no company wants to intentionally discriminate against others, they do want to hire the very best talent for the roles they’re looking to fill. The EEOC is clear that companies can use assessment tools that can lead to unfair outcomes if (and only if) they measure characteristics of potential employees that are required to perform the job.

When using assessment tools that may lead to unfair selection ratios, companies must show job relevancy in order for the tools’ use to be valid. For example, many manufacturing positions are filled after candidates are screened using a physical strength assessment. While some physical strength tests will lead to a disproportionate number of women being eliminated from the candidate pool, the assessment is legal to use if the requirements – like regularly lifting 150 pounds on an ongoing basis – are relevant to the role.

This means that it is especially important to identify the requirements for a job before you implement any selection process. Use a good job analysis process to identify what is required for success in a job role before using selection tools. Many companies offer easy to use job analytic technology that is geared for their own assessment tool or tools. At PI, we strongly recommend that every client use our PRO (Performance Requirement Options) behavioral job analysis tool before ever using the Predictive Index for hiring. The PRO asks subject matter experts (people who know the job well) to identify the key behaviors required for success in a job role, taking into account not just the responsibilities of the job, but also the work environment and culture. Formally, the output provides a clear outline about what is needed for success in the job so that selection decisions are based on behavioral requirements. Informally, the PRO process has a big benefit for our clients because it provides an opportunity to have a discussion about the role and the behaviors that are best suited for success. These discussions can be valuable because they both align various internal stakeholders, but also help build much better clarity around what a person is expected to do in a job. Using job analysis may take a little extra time but it is time well spent when considering the benefits of hiring better talent and the costs of not getting it right.

It is important to carefully monitor the performance of all assessment tools to ensure that they aren’t leading to adverse impact (i.e. unfairness) against a protected group. Target probably should have more carefully tracked their entire hiring process to identify what percentage of different protected classes were remaining in the candidate pool as they were going through various stages of the assessment process. Had they been actively monitoring this, they probably would have noticed the exclusions and been able to make earlier course corrections either to the tools they were using or the criteria they were basing their decisions on.

While looking at assessments, there is often an important balance between risk and fairness. While personality and behavioral tools tend to be fair to all people who take it (e.g., there are no differences on gender, age, or ethnicity), tools like cognitive ability tests can do a better job of identifying top talent. Instead of shying away from higher risk assessments, the best approach is to first understand what is required for success on the job (e.g. job analysis) and then build a balanced selection system which maximizes validity and minimizes risk.

Take-aways:

  • Always conduct a job analysis that helps identify what is required for the job
  • Use that analysis as an initial basis for making choices about the selection tools you are using
  • Monitor your candidate pool and pass rates to make sure protected groups are not receiving unfair treatment
  • Consider identifying a selection strategy that balances traditionally fair assessments with higher risks tools

 

 

Comments

  1. I don’t see how the PI could be justified as testing skills necessary for a job. The speed required to “pass” the test is tested under artificial conditions. I doubt that you have statistical data that could prove that speed on the PI correlates with speed in every day work tasks. On the PI, a matter of seconds on a question can make the difference between passing and not passing. In every day work, this may not translate into a significant difference in the ability to complete tasks.

    In addition, people who make take a little more time on a task may have compensating strategies that they can use on the job that would not apply to the PI. For example, they might prioritize tasks or request clarification of the level of “perfection” needed on a task.

    Absent of such statistics, it seems like that a company using the PI to screen job candidates is violating the ADA.

    1. Hi Jay,

      Thank you for your perceptive questions! The PI Cognitive Assessment to which you are referring measures general cognitive ability, but it is not intended to assess all of the necessary skills for a job. Nevertheless, higher cognitive ability is related to higher job performance in nearly all job roles, which is why cognitive assessments are such useful and flexible tools for many companies!

      Regarding your observation about speed, it should be noted that all standardized testing is done under artificial conditions. (For example, very few real-life situations are presented as multiple choice questions). Test designers and psychometricians have the responsibility to design a standardized assessment that provides comparable scores for all test takers, but which also relate to appropriate interpretations and applications by the test users. This is the practice of psychometric validation, which provides many of the statistics to which you are referring. In the case of the time component of the PI Cognitive Assessment, speeded information processing falls into the domain of general cognitive ability, as defined empirically by psychologist John Carroll. This is why many (but certainly not all) cognitive assessments have a time component. The questions on the PI Cognitive Assessment are simple enough that most people would be able to answer all of them correctly, given enough time, so the timed conditions help differentiate between different levels of cognitive ability.

      While it is true that a matter of seconds could impact people’s scores, all respondents are provided with the same time limit (with the exception of extended time accommodations), thus ensuring no one respondent has an advantage over another. You may also be interested in knowing that the PI Cognitive Assessment is not a pass/fail test! Test users see how well a respondent matches the cognitive demands of a given job role, and this is reported on a sliding scale. This helps to ensure that getting one or two questions right or wrong does not change a candidate’s overall suitability for a job. The cognitive scores are also just one of many factors a company considers in the hiring process.

      Appropriate use of the PI Cognitive Assessment or other cognitive assessments does not constitute a violation of the ADA, and test users are able to provide many reasonable accommodations for candidates, including extended time versions that use score equating to yield comparable scores. The relevance of cognitive ability for a given job role is determined by the test user through a PI Job Assessment or other valid standard-setting methodology, and the utility of cognitive ability for selecting candidates is typically demonstrated through criterion validation studies. These steps help ensure not just validity, but fairness too. These studies and analyses make sure that candidates are compared against an accurate job target, that the assessment is relevant for selecting candidates, and that the assessment is not used inappropriately with populations or applications for which it was not intended. By adhering to these psychometric best-practices, PI’s researchers ensure fairness and protection for candidates while providing accuracy and standardization for test users making hiring decisions.

      Thanks again for your great comments!

Leave a Reply

Your email address will not be published. Required fields are marked *