When you are an employer, knowing who is defined as an “applicant” matters. The Equal Employment Opportunity Commission (EEOC) and the Office of Federal Contract Compliance Programs (OFCCP) focus on systemic hiring discrimination and obtain large monetary settlements every year.
Within the U.S. Department of Labor, the EEOC investigates charges of discrimination brought against employers, and the OFCCP conducts roughly 4,000 audits of federal contractors and subcontractors every year.
And here’s a well-kept secret: The agencies do not have to prove the employer intentionally discriminated; rather, under the disparate impact theory, they must demonstrate that the employer’s hiring process negatively affected a group at a statistically significant level. So, identifying and strategically presenting who is an “applicant” — who should count against the employer in these statistical analyses.— is critical in EEOC investigations and OFCCP audits.
So, just who is an applicant?
The Internet Applicant Rule is the single greatest idea ever contrived by the Department of Labor. It defines the type of applicants federal contractors and subcontractors must include in these analyses — and therefore who can be removed — as follows:
● Candidates who submit an expression of interest
● Candidates who meet the basic qualifications
● Candidates who are considered (their substantive qualifications reviewed)
● Candidates who do not expressly (or who the employer may infer) remove themselves from consideration
Similarly, for employers who are not federal contractors or subcontractors, to fall within the definition of “applicant” in the Uniform Guidelines on Employee Selection Procedures, candidates must submit an expression of interest, be considered, and not remove themselves from consideration.
Big Numbers Are Bad Numbers
Many employers allow candidates to apply online, and for some employers, that is the primary or only way candidates can apply. With this technology, employers are seeing significant increases in their candidate pools. But they should also be thinking about how these large numbers may affect potential liability in EEOC investigations and OFCCP audits. (In systemic discrimination investigations, the agency often asks for three to five years of applicant data; for the OFCCP, the agency starts with one year and may expand to two years if it sees statistical indicators.)
In statistical analyses, big numbers are bad numbers. The more applicants in the analysis, the more likely the employer is to tip over into statistical significance; a standard deviation greater than 1.96 is considered statistically significant.
In the example below, the original data contained all expressions of interest (everyone who applied) for a total of 10,000 candidates for one position. Often, employers remove candidates who were not considered and who withdrew — and, in the case of federal contractors, candidates who did not meet the basic qualifications; this allows them to present their applicant data in a better light. It is this decrease in applicant numbers in the above example that brought the employer below statistical significance.
How to Decrease Large Applicant Numbers
Develop Strategic Disposition Codes under the Internet Applicant Rule
To take advantage of the rule, employers should develop strategic disposition codes that mirror elements of the rule to identify who does not meet the definition of “applicant” and can be removed. For example, if the candidate does not return calls, takes another job, or wants too much money, the candidate is not considered an applicant because he or she is not willing to do the job, and the employer can infer that the candidate will remove himself or herself from consideration. Similarly, if the employer does not review the candidate’s substantive qualifications because the job was filled and the candidate applied too late, the individual is not an applicant under the rule because the employer did not consider him or her.
Implement Best Practices
To minimize the number of applicants, employers should consider implementing the following best practices:
● Close requisitions when filled. Closing requisitions when filled will reduce the number of candidates applying.
● Make one hire per requisition. One hire per requisition also makes the data cleaner to defend since all candidates in the requisition were considered for that specific position. “Evergreen,” or “continuous,” requisitions, on the other hand, are difficult to defend because it is not clear who was considered for each position. When this happens, the agencies can count the candidate multiple times, which leads to larger numbers.
● Do not move between requisitions. Similarly, moving candidates between requisitions suggests that the employer considers candidates for more than one position and allows the agencies to count them multiple times. Instead of moving them, consider inviting them to apply for another suitable requisition.
● Use data management techniques (DMTs). DMTs allow employers to consider some of the candidates, and those not reviewed do not count against the employer. This technique is especially helpful for entry-level, high-volume positions. To use this technique, look at the candidates in batches, such as 10 or 20 at a time, by a qualification-neutral means, such as timing — first received or most recently received. Once you have enough successful candidates to move on to the next phase, stop reviewing. Once the position is filled, disposition the remaining candidates who were not reviewed as “Never Considered – Data Management Technique.”
Using these suggested best practices and the Internet Applicant Rule can help employers reduce their applicant pools, thus minimizing EEOC and OFCCP liability.●
Jennifer Seda, JD, is a principal in the Denver, Co., office of Jackson Lewis P.C. She oversees the preparation of many of affirmative action plans each year. Over the past several years, she has defended multiple OFCCP audits, including on-site and corporate management (“glass ceiling”) reviews. For questions, contact Jennifer at firstname.lastname@example.org.