When the compliance evaluation begins, the compliance officer is running numbers through a computer program to determine if there is an imbalance in outcomes such as hiring, pay rates, promotions, or terminations that would indicate a need for further, more thorough, investigation. A statistical tipping point such as two or more standard deviations from the norm or an established degree of disparity in a pay snapshot lead the agency to examine the employment event further to determine whether or not there is discrimination. A contractor's self-audit follows a similar pattern. In both instances the original indicators are a rough measure of possible discrimination and often, indeed, usually require further refinement.
The process of refinement requires a determination of whether the right people are in the potential affected class as well as a determination of whether the right people are in the favored group offered for comparison. Moving individuals out of either or both categories can change the calculation of standard deviations and thus, change the outcome of the case. OFCCP compliance evaluations will often drop the discrimination allegation if the statistical tipping point is not met after refinement. If the case is not dropped after the indicator falls out, OFCCP would be limited to pursuing individual isolated instances of apparent discrimination, in which case the analysis of who is similarly situated should be even more narrowly tailored.
This article discusses the meaning and importance of determining who is similarly situated in hiring. The issue of who is similarly situated in termination, promotion and compensation cases will be addressed in next month's publication.
Similarly Situated in Hiring Cases
There are several factors that go into determining whether the right people have been included in the discrimination analysis. The compliance evaluation usually covers a set period of time. If discrimination is suspected that period can begin up to two years prior to the issuance of the scheduling letter. In hiring cases, individuals who applied prior to the review period, including those who applied prior to the review period but hired in the review period, are typically removed from OFCCP's analysis. People who applied during the review period but after the last hire that falls within the review period are also typically eliminated from the analysis. Most analyses include all applicants who applied during the review period and were considered for a position that was filled during the review period. If there are applicants who do not meet these criteria they probably will not be included in OFCCP's discrimination analysis in either the potential affected class or the alleged favored group.
OFCCP does not usually focus on the identity of the decision maker in determining who is similarly situated with respect to hiring. In hiring cases, the agency is usually looking for a pattern over time rather than an isolated hiring event in its analysis of discrimination. This is why everyone hired during the review period into a particular job is analyzed together.
Contractors, however, may want to look at the specific hiring events to see how many of the alleged disfavored group were actually available for hire when openings occurred. For example, assume the contractor had 10 hiring opportunities during the review period. Further assume, that you could only be considered for positions for which you specifically applied. By looking at the specific hiring events, you could determine whether females were fairly evenly distributed across all 10 hiring events or whether there were only two hiring events where there were female applicants in the applicant pool. Assume females were hired in the two hiring events where females were actually in the applicant pool and males hired in the other 8 hiring opportunities where no females applied. This presents a different picture of the likelihood of gender discrimination than if females were present in all the applicant pools for each of the 10 hiring events. Reviewing the specific hiring events, thus, may present a different picture of whether females were actually similarly situated to males with respect to the likelihood of being hired.
Method of Application
An individual's method of application can be a factor in determining who is similarly situated. For example, if a company requires all applicants to submit an application form in order to be considered for a position and never accepts resumes, applicants who submit resumes would be eliminated from the pool since they are not similarly situated with respect to the method of application. However, if the company makes exceptions to this policy, the similarly situated pool expands to include everyone who submitted resumes and/or application forms. Making exceptions to company policy can easily expand the pool available for comparison and make a huge difference in a contractor's potential liability.
A related issue arises when contractors hire individuals for job B who only applied for job A. If the contractor only hired applicants for job A who applied for job A, the pool would be limited to those who were similarly situated in that they all applied for job A. However, if the contractor hires an applicant who did not apply for job A but rather applied for job B it expands the pool for comparison because now the agency must consider the other job B applicants who could possibly have been tapped for the job A position. The expansion of the pool of potential applicants can dramatically impact the analysis in either direction. In other words, it might produce an indicator of discrimination or it might eliminate an indicator of discrimination. The important thing for a contractor is to avoid producing indicators of discrimination.
In hiring cases, which make up the largest percentage of OFCCP cases, the requirements for the position are another factor in determining whether the pools in the equation are similarly situated. It may be tempting to look at the stated qualifications for the position and start removing individuals from the applicant pool based on whether they meet the stated qualifications. I recommend that the analysis begin by looking at the qualifications of all of the people hired into the position during the time period under review to determine if the contractor actually applied its stated criteria in making the hires.
Contractors will often have established hiring criteria but one can only know if these criteria were actually applied as stated by looking at whether the individuals hired actually satisfied these criteria. For example, a contractor might indicate that an applicant must be a licensed electrician in order to be qualified for the job but actually hired someone who had completed the necessary course work and apprenticeship who was in the process of pursuing his or her license but had not yet actually obtained his or her license. By examining the actual hires you know that the similarly situated applicants would not be limited to licensed electricians but rather to all applicants in both the favored and disfavored groups who were similarly situated to the hired applicant in that they had completed similar course work and apprenticeship and were pursuing their license. If the contractor had not hired anyone without an electrician's license everyone without such a license could be eliminated from the applicant pool since they would not be similarly situated to the favored class hires. This presumes, of course, that the licensing requirement itself was not discriminatory.
In making such adjustments to the applicant pool or the comparator favored class pool it is important to make the adjustments on both sides of the equation. For example, assume you had included only licensed electricians in the qualified applicant pool and afterwards discovered a hire who is not a licensed electrician. In this situation as noted above you would add back in everyone who has qualifications that match the least qualified hire. Since you would be adding potential hires back into the equation you should really rerun the IRA with the appropriate individuals in the pool for all groups since not only might previous indicators be resolved but a different affected class might emerge. At this stage, similarly situated means at least as qualified as the least qualified hire.
The fact that these individuals are added back in does not mean that the likelihood of someone without the license being hired is the same as the likelihood of someone with the license being hired. This issue of the likelihood of hire should be further analyzed. Assume for the sake of this example that the issue being evaluated is gender discrimination. One way to account for the difference in likelihood of hire between the licensed and not licensed would be to compare the selection rates of license female electricians to the selection rates of licensed male electricians. In a separate equation, compare the selection rates of unlicensed female electricians to the selection rates of unlicensed male electricians. In these equations, the concept of similarly situated means similarly situated with respect to having or not having an electrician's license. By running these comparisons you can detect whether it is more important for a female to possess the license than it is for males to possess the license. You can also see whether the license requirement had any impact at all on the likelihood of hire. There may be other ways to account for the difference in likelihood of hire but the contractor should make sure that OFCCP has accounted for this difference.
If the tipping point indicator is resolved through refinement of the applicant pools, OFCCP may still look for isolated instances of apparent discrimination. Once the statistical indicator is gone the inferences permitted by the indicator are also gone. This means that the individual alleged to have been discriminated against must be even more closely examined against the favored comparator. For example, assume there is no longer a statistically significant disparity in male and female hires but OFCCP is concerned that an apparently more qualified female applicant was not hired while apparently less qualified males were hired. Since there is no statistical disparity that permits an inference of discrimination, the female and male comparator(s) will have to be examined closely to see if they were actually similarly situated to each other. Questions such as whether they applied at the same time and whether they were examined by the same hiring official(s) become important in ways they would not have been in a pattern and practice statistical case.
Understanding how to use the concept of similarly situated at each stage of the analysis can make all the difference between being in violation and being in compliance. Both the contractor and OFCCP should be examining this question over and over again through the evaluation or self-audit process. The question of who is "similarly situated" is the question of whether we are comparing apples to apples or apples to oranges. You do not want to be cited for discrimination because you are being evaluated using apples to oranges criteria.
The following suggestions may be helpful:
- Make sure you completely understand who is in the OFCCP applicant pool and who is in the hiring pool. Get a list of who these people are.
- Have the agency explain its analysis of each stage of the hiring process. Since individuals are retained or excluded at each stage of the hiring process, it is important to understand each addition or subtraction to or from the applicant and hiring pools.
- Make sure that any addition or subtractions are made from both the favored and disfavored group.
- Where policy is not followed, examine the impact of such deviations on the size of the potential applicant pools. Convey the importance of following company policy to limit potential discrimination.
- Make sure that OFCCP is not conflating failure to follow your policy with discrimination.
- If you disagree with the agency about who is included in its analysis, take the time to carefully explain what the proper pools for comparison are and who should be included.
- When statistical indicators drop out remember cohort analysis will often require that the comparators were head to head in availability for the employment opportunity not that they simply applied during the same review period.
The more comfortable you get with playing with comparisons of similarly situated groups, the easier it will be for you to determine if discrimination actually occurred and the easier it will be to identify any flaws in OFCCP's analysis.