Ricci and Briscoe as Disparate Impact Cases
UPDATE: Seven African-American testtakers in Ricci have moved to intervene in Ricci, which is back at the district court for implementation of the Supreme Court decision. Also, African-American testtakers have filed disparate treatment and disparate impact discrimination charges with the EEOC. All this reported in the Connecticut Employment Law Blog, www.ctemploymentlawblog.com/2009/11/articles/decisions-and-rulings/black-firefighters.
The main thrust of Ricci focused on the disparate impact issue and its implications will likely be worked out in Briscoe v. City of New Haven, a disparate impact case brought against the City because it has now used the test scores challenged in Ricci. In Ricci, the City argued that its decision not to use the test scores was made to avoid the risk of disparate impact liability to the African-American testtakers who would not be promoted if the test scores were used. The Court conceded, as did all the parties, that the use of the test scores would have resulted in a disparate impact on African-American testtakers. Using the “pass rate,” or cutoff score that was set for the test, less than 80% of the minority testtakers passed. More important, the actual use of the test among those who passed would have excluded from immediate promotion all the African Americans and all but two of the 22 Hispanic testtakers. (Three African Americans might have some chance for promotion if new openings occurred in the future during the life cycle of the test.) With that prima facie case of disparate impact discrimination conceded, the focus moved to the business necessity and job relatedness affirmative defense and the plaintiff’s surrebuttal possibility of showing that an alternative was available that served the interests of the City but resulted in less impact.
Given the posture of the case – using the risk of disparate impact liability as a defense to a disparate treatment claim – the City had the burden to prove that it would not be likely to carry its burden of proving the test’s business necessity and job relatedness or that disparate impact plaintiffs would likely be able to prove an alternative way promote to promote firefighters that had less impact. The Court rejected the arguments that the City had to prove it would actually lose such a disparate impact case or that its good faith belief sufficed. Instead, the City had to have a “strong basis in evidence” for believing it would be liable for disparate impact discrimination. In other words, it should be somewhat easier for the City to win the issue of its potential risk of disparate impact liability than it would be if disparate impact plaintiffs actually had to prove the City liable for disparate impact discrimination.
The major focus of the Court turned to whether the City showed it had a strong basis in the evidence that the test was neither “job related for the position in question [nor] consistent with business necessity.” Long story short, the City had spent a lot of money hiring a test consultant who did an analysis of the lieutenant and captain jobs and then drafted a written test asking questions about information relevant to the jobs. Instead of jumping from job analysis to test writing, the next step should have been, at least under the EEOC’s Uniform Guidelines on Selection Procedures, deciding what kind of test should be constructed, if any, and what strategy should be used to validate the test. Under the classic approach established by the professional standards of industrial psychologists, a pen-and-pencil written test for a job that did not require the workers to take such tests as part of the job would have to be validated using criterion-related validation.
Criterion-related validation would require giving the test, hiring all the testtakers, doing a subsequent evaluation of their job performance and then running a correlation coefficient to determine whether the test scores correlated with their subsequent job performance scores. Since criterion-related validation is not used out in the real world, the City’s test consultant simply argued that the written test was content valid, i.e., that the written test was a sample of the lieutenant and captain jobs even though test taking was not part of the job.
Relying on content validation, even though it technically was inappropriate for this written test, would not be not risky if the City had been sued for disparate impact discrimination. Basically, the lower courts have upheld pen-and-pencil tests as content valid even though the job did not require workers to take pen-and-pencil tests as long as the defendants had spent some real money constructing and administering the tests and the test questions asked about the job. So, given the judicial abandonment of the professional test validation standards and if business necessity and job relatedness were the only issues, the City would have prevailed on its affirmative defense. With the test being held to be validated, the City could use it. So, on the issue of the validity of the written test, it is no surprise that the Supreme Court found as a matter of law that the test was job-related and justified by business necessity. If that was the only issue, the City would lack a “strong basis” in evidence that it faced disparate impact liability if it had used the test results.
But, even if a defendant proves its written test is job-related and justified by business necessity, the plaintiff has one more bite at the apple. That is by showing that the employer refused to adopt an available alternative that had less disparate impact and served the employer’s legitimate needs. The Court did discuss this and, in light of the weak “available alternative” authority that has been developed in the lower courts, it is not surprising that it found that the City lacked a “strong basis” in evidence that potential disparate impact plaintiffs would prevail on their available alternative surrebuttal claim. While shifting the weighting of the written and oral scores to reduce reliance on the written part of the exam would be an alternative that the Court indicated would have reduced the test’s disparate impact, that was not enough to establish that the test was illegal. The Court found that the City produced no evidence to show that a different weighting would be equally valid in determining who was qualified. In other words, an alternative that reduced the impact of the test was a necessary but not sufficient condition for this surrebuttal stage. There would also have to be a strong basis in the evidence that a change in weighting the scores would have adequately served the employer’s interests. While there was some testimony before the City when it was making its decision whether or not to use the test scores that the City of Bridgeport had reduced the disparate impact of its written exams by changing its weighting with an oral score, the City did not have before it any evidence that doing this would serve the City’s legitimate interests. This is a bit debatable since the claim was made that the revised system used by Bridgeport worked just fine. All in all, applying the prevailing lower court authority to this pen-and-pencil exam, it was likely that the City would not have a strong basis in evidence that it would be liable if it used the test scores. In sum, the Court pretty much applied the prevailing authority from the lower courts regarding job-relatedness, business necessity and available alternatives in deciding that the City would have won, not lost, a disparate impact challenge if it had used the test scores for promotion.
This is where the recent, post-Ricci disparate impact claim, Briscoe v. City of New Haven, comes in. Plaintiff, an African American firefighter who took the test at the same time as the Ricci plaintiffs, claims that he would have been promoted if the weighting were changed and that the weighting of 60/40 favoring the written component had a disparate impact on African Americans. Further, he claims the City knew, even before it decided to use a written test for promotions, that using this 60/40 weighting would have a disparate impact. Nothing was done about the weighting issue because it was a longstanding feature of the collective bargaining agreement between the City and the firefighters’ union even thought the existence of that contract obligation would not be a defense to a disparate impact claim. Briscoe’s claim of impact is that, of the 77 candidates for the lieutenant position, he scored the highest on the oral part of the exam, but his overall score using the 60/40 weighting left him 24th on the list and thus not eligible to be promoted. His complaint alleges:
“The City did not believe that the 60 percent weighting that it required was job related, and it knew that the weighting would have a disparate impact on African-American candidates: for example, on the lieutenant exam immediately preceding the 2003 exam, the African-American candidates as a group performed substantially better than the white candidates on the oral exam, but they were scored much lower overall because of the 60 percent weighting given to the written test.”
The job-relatedness of the written test, as decided in Ricci, is irrelevant to Briscoe’s claim because he focuses on the impact and job-relatedness of the weight given to the written test scores, not the test itself, when those scores were combined with the oral scores to rank order those who would be promoted. The claim relies on Connecticut v Teal, where the Court upheld a disparate impact challenge to part of an overall selection process – the use of a written test – even though there was no disparate impact resulting at the end of the whole selection process. In Teal, African-American plaintiffs had flunked a written exam and so were not eligible to continue further in the selection process. Because of the impact that the written test scores had in the overall process, they could challenge the test even though when the hiring decisions were finally made there was not a disparate impact on African Americans among those hired.
Briscoe tries to escape Ricci entirely by making his disparate impact based on fact that the liability of the City is triggered by events that all occurred before the test was administered, before the consequences in terms of the scores of different groups were known and before the City decided not to use the test results because it knew the racial consequences of using the scores. Thus, the argument will be whether or not Ricci has any relevance because Ricci’s focus was at a time after the test had been given, its racial impact was known, including the racial composition of those who would be promoted if the scores were used. In Ricci, Justice Kennedy seems to emphasize that the case focuses on what happened once the test was given.
“Nor do we question an employer’s affirmative efforts to ensure that all groups have a fair opportunity to apply for promotions and to participate in the process by which promotions will be made. But once that process has been established and employers have made clear their selection criteria, they may not then invalidate the test results, thus upsetting an employee’s legitimate expectation not to be judged on the basis of race. . . . Title VII does not prohibit an employer from considering, before administering a test or practice, how to design that test or practice in order to provide a fair opportunity for all individuals, regardless of their race.”
That suggests that when the City was planning to use the written test it was required to look, rather than being restrained from looking, at the potential disparate impact of its subsequent use. In other words, Ricci only comes to bear once a test has been given and the successful testtakers have an established reliance interest in its use.
Assuming that the earlier time frame does makes Ricci irrelevant to his claim, Briscoe focuses his attack on the disparate impact of the weighting of the written and oral scores and not the rest of the test as it was constructed by the test consultant. Again, the complaint alleges that:
“If the oral exam were weighed 60 percent and the written test 40 percent instead of the other way around, the plaintiff would be ranked ninth instead of 24th and would be promotable. . . . If the written test were not considered [at all], . . three African-Americans would be in the top 12, and the plaintiff would be at the top of the list.”
Briscoe looks more deeply at the issue of what caused the disparate impact in the City’s promotion process than the simple recognition in Ricci that, at the end of the day, the City’s selection process caused a disparate impact on the basis of race. Based on Teal, the weighting element of the selection would appear to make out a prima facie case of disparate impact discrimination. That was not addressed in Ricci because that case focused on the justification, the job-relatedness, of the written test itself, not how the test results were to be used.
If Briscoe establishes a prima facie case on the weighting issue, the burden of persuasion shifts to the City to demonstrate that the weighting formula it used was “job related for the position in question and consistent with business necessity.” The only evidence in Ricci about the weighting was whether it was an available alternative, an issue upon which the City failed to carry its burden of persuasion. That does not address the business necessity issue as to the weighting at all. And, even if the City did prove that the weighting formula it used was job related and consistent with business necessity, that should still leave to plaintiff the chance to prove that a different weighting formula was an available alternative that the City knew of and failed to adopt. Just because the City failed to carry its burden on that issue in Ricci, would not seem to bar Briscoe from carrying his burden unless the Ricci Court meant that, as a matter of law, the alternative of changing the weighting was not an available alternative.
While Ricci seemed to adopt the weakened use of the professional test validation standards to justify a written test as job-related for jobs that did not involve test taking, that decision may not foreclose Briscoe. And, of course, if Briscoe is successful, that will have no negative impact on the Ricci plaintiffs, who have been and will be promoted using the results of the test that the Ricci plaintiffs and Briscoe took. If the City failed to carry its burden of proving that the weighting of the written and oral scores was job-related and consistent with business necessity, Briscoe would be entitled to a remedy. Presumably, the City would be ordered to adopt a new promotion process that would either not result in any disparate impact or, if it did, was job-related and consistent with business necessity. Briscoe would then be subjected to the new process and, if he showed himself to be qualified, would get the first opening for lieutenant after the cycle of use of the Ricci test was completed. He should also receive backpay plus front pay until he gets his promotion, with that determined by the difference between his earnings as a firefighter and what he would have earned as a lieutenant. If, however, the City proved the test was valid but Briscoe proved on surrebuttal that an alternative was available, Briscoe would have to show that he was qualified if that alternative was used. If he did, he would be entitled to promotion once the Ricci test had run its course and back and front pay until that happens.
Given the success of the Ricci plaintiffs’ disparate treatment claim and assuming Briscoe will be successful on his disparate impact claim, the City of New Haven would appear to be damned for not considering the disparate impact before it set up the promotion procedure and then also damned for trying to take account of that impact once the test was given and the reliance interests of the successful testtakers had been established. But relief to the Ricci plaintiffs is not in conflict with relief to Briscoe: The Ricci plaintiffs get the promotions earned by their test scores and Briscoe would likely get the next opening of a lieutenant slot after the cycle for the use of the test had ended.
With that result, there would certainly be strong incentives for employers to investigate the potential disparate impact of any employment practices before using it and either modifying it to avoid that impact or be confident that its use can be shown to be job-related with consistent with business necessity with no alternative available to its use. Further, once the employment practice is implemented and there exists identifiable people who have legitimate expectations that their success pursuant to that practice would be followed, then it is likely too late for employers to refuse to honor those expectations. To fail to use the results of the practice would trigger disparate treatment liability unless the employer had a “strong basis in evidence” that it would be liable for disparate impact discrimination. One wonders, of course, whether an employer can prevent the creation of any expectations, or reliance interests, by making it clear from the start that the employer reserves the right not to use the results of the employment practice it uses.