The U.S. Office of Federal Contract Compliance Programs (OFCCP) released two new FAQs this week.  The new guidance documents address three topics: validation of tests used by contractors when selecting workers for employment, OFCCP's use of “practical significance” during compliance evaluations, and how contractors can determine whether to include project-based or freelance workers in their affirmative action programs (AAPs).

Validation of Employee Selection Procedures

One FAQ released this week discusses the validation of employee selection procedures and, most notably, how OFCCP will assess the validity of selection tools built on AI and other “‘new technology’ screening devices.”  Much of this FAQ recapitulates the Uniform Guidelines on Employee Selection Procedures’ (UGESP) basic requirements for assessing the validity and legality of employee selection procedures.  The FAQ also clarifies that AI-powered tools are subject to same requirements as other employee selection procedures.

In adverse impact inquiries, OFCCP first examines differences in selection rates between protected groups in order to determine whether an adverse impact exists.  The FAQ lists the statistical tests that the agency typically uses to assess whether a gap in selection rates is wide enough to raise an inference of adverse impact.  This FAQ only mentions in passing the concept of “practical significance” in determining whether a gap in selection rates is significant, but a separate FAQ also released this week (discussed further below) signals that the OFCCP will look beyond bright-line statistical and numerical tests when assessing whether an adverse impact exists.

If there is an adverse impact, an employer must validate the selection procedure in accordance with the UGESP through either content, criterion-related, or construct-validation studies.  The FAQs underscore that the UGESP generally requires validation to be performed at the employer’s own local facilities unless they can demonstrate that a validation study conducted on workers from other organizations can be reliably transported to the contractor’s workforce.  This means that employers generally cannot use off-the-shelf aptitude and other similar tests without validating those tests on their own workers—a key requirement for employers considering implementing commercially available tools that use machine learning and algorithmic modeling.

Lastly, the new guidance makes clear that a contractor must validate an employee selection procedure in accordance with the UGESP if the procedure has an adverse impact, “[i]rrespective of the level of technical sophistication” that the selection procedure involves.  In other words, employers should expect OFCCP to assess recruitment and hiring tools that leverage data mining, machine learning, and algorithmic modeling under the same standards it uses to assess paper-and-pencil examinations, physical fitness tests, and other traditional selection procedures.

Practical Significance in Adverse Impact Analysis

Another FAQ discusses how OFCCP uses the concept of “practical significance” when assessing adverse impact and the availability of alternative selection procedures in disparate impact cases.  OFCCP describes practical significance as referring to “whether an observed disparity in employment opportunities or outcomes reflects meaningful harm to the disfavored group.”  In contrast to statistical significance, “[t]he concept focuses on the contextual impact or importance of the disparity rather than its likelihood of occurring by chance.”  OFCCP “may consider practical significance as part of a holistic evaluation” of an adverse impact case alongside statistical significance and other available evidence.

While the UGESP and the original Q&A document that accompanied it make brief reference to the concept of practical significance, OFCCP’s new guidance acknowledges, “[t]here is no universally accepted measure of practical significance in the EEO field. What is considered practically significant depends on the employment opportunity at issue and the specific facts of the case.”  The lack of specificity in this description is in line with the historical use of practical significance by courts and regulators, which have been inconsistent both in whether they use practical significance and, where they do use the concept, the standards by which it is assessed.  As employers increasingly leverage “Big Data” in their employee selection processes, we expect considerations of practical significance to become increasingly prominent.

Freelance Workers in AAPs

In essence, the new guidance reaffirms the rule that AAPs should include employees only, and not independent contractors.  Generally, workers who “perform temporary, flexible jobs and are paid per service provided or task completed” are considered outside the employer/employee relationship.  The new FAQ then refers readers to the OFCCP’s more extensive FAQ on the distinction between employees and independent contractors, which the OFCCP assesses using the U.S. Supreme Court’s “Darden factors” to determine independent contractor status.  Generally, under the Darden factors, where a company controls the manner and means of a worker’s performance of his or her work, that worker is deemed an employee rather than an independent contractor.   

Contractors should pay close attention to this issue not only for affirmative action planning purposes, but also to ensure compliance with other federal and state laws such as the Americans with Disabilities Act, the Consolidated Omnibus Budget Reconciliation Act, the Employee Retirement Income Security Act, employment tax laws, and others.

Conclusion and Takeaways

None of OFCCP’s new guidance is groundbreaking, but it does underscore the need for contractors—and, indeed, employers more generally—to tread cautiously when deploying new employee selection tools or making decisions concerning which workers to include in AAPs.  The distinction between employees and independent contractors has become increasingly difficult to draw with the rise of the gig economy, so contractors must make a careful assessment of workers’ classification when preparing AAPs.

OFCCP’s decision to assess algorithmic tools under the same standards as paper-and-pencil tests may seem counterintuitive, but this approach is consistent with current case law.  While technology has changed radically in recent years, the UGESP has remained constant since the Equal Employment Opportunity Commission, Department of Justice, Office of Personnel Management, and OFCCP first promulgated it more than four decades ago.  In the interim, both the nature and methods of employee selection have evolved considerably, as has the social science regarding test validity and how practical considerations can be used to assess whether a test is unfairly biased.

For the moment, however, the UGESP remains on the books as it appeared 40 years ago, and OFCCP’s new guidance makes it clear that the agency will continue to apply the UGESP to new methods and forms of employee selection.  That means employers must be mindful of whether practically meaningful differences in selection rates exist and, if so, whether the selection procedure can be validated in accordance with the UGESP.