Legal Update - February 2010

Gathered by Richard Tonowski, Equal Employment Opportunity Commission (EEOC)

SELECTION CASES

G’day, mate.  You’re out 19 million quid.

Outback Steakhouse is a national organization of Australian-themed casual restaurants.  The franchise in Colorado entered into a settlement with the EEOC valued at $19M.  The issue was non-promotion of women to managerial positions.

As with several other recent EEOC selection cases, the substance of the suit was on the looseness of the selection procedure.  In EEOC’s view, the problem was a “tap on the shoulder” initial step.  There were no applications.  Managerial prospects were identified by their supervisors.  The procedure was more formalized in later steps, including tests that were not at issue in the charge.  For EEOC and the charging parties, too few women got the call to advance. 

The company denied any discriminatory practices and argued that the disproportionate lack of women had to do with lack of interest in managerial careers, which could involve long and irregular hours.  Outback commissioned a survey to back up its assertion.  EEOC objected to its technical adequacy: conducting the survey after suit was filed and after many of the affected women had left the company, non-random sampling, and unique scoring and statistical analysis.

This is one type of case that EEOC seems to be pursuing recently.  The central features are appreciable under-representation of a demographic group and a hazily-defined selection procedure.  While the employers are likely to claim lack of interested and qualified applicants, EEOC will suggest that in the alternative excessively subjective decision-making is likely being influenced by stereotyping

Shot down before take-off.

It isn’t often that the court finds that an employer’s testing effort is so bad that there’s no need for a trial, but it happened to New York City.  Then it got worse.  At the heart of the case is determination of the point where severe, persistent adverse impact indicates malignant neglect tantamount to purposeful discrimination.

Federal Judge Nicholas Garaufus followed up his granting summary judgment against the New York City Fire Department (FDNY) entry-level selection testing last summer with separate rulings on liability and relief.  There are now three orders regarding this case (United States of America and The Vulcan Society, Inc. et al. v. The City of New York et al., E.D.N.Y., No. 07-cv-2067).

The first order was issued on July 22, 2009, granting summary judgment to the plaintiffs regarding disparate impact.  The court’s conclusion was that the plaintiffs had established their prima facie statistical case and the city failed to raise a triable issue in its defense that the tests were job-related.  At issue were Exam 7029 (administered from 1999 through 2002) and Exam 2043 (administered from 2002 through 2007).  The pass rate for whites on Exam 7029 was 89.9%, while that for blacks was 60.3%; for Hispanics, the pass rate was 76.7%.  For Exam 2043, the pass rates were whites was 97.2%; for blacks, 85.4%; for Hispanics, 92.8%.  Plaintiffs presented evidence that the disparities to be statistically significant.  (It might be noted that not all of the disparities would fail the 80% Rule.)  Plaintiffs also presented calculations on what these disparities meant regarding the numbers of blacks and Hispanics who could not move forward in the selection process.  Also examined was the relative placement of the race/ethnic groups on the resulting rank-ordered lists.  Blacks and Hispanics were disproportionately toward the bottom of the lists.

The city apparently did not challenge the numbers.  Rather, the city argued that the statistical calculations of significance and shortfall assumed perfect parity among demographic groups in capability and preparedness when such was unrealistic, that large numbers of applicants made the significance tests unreliable, and that the 80% Rule should be used to evaluate any adverse impact.  The court found that large sample sizes (and low p values) would be expected to make statistical significance tests more reliable, not less.  Regarding the “perfect parity” assumption, the court found that it was an appropriate use of statistical significance tests to show that it was unlikely that the disparities could have occurred simply by chance; whether those disparities were due to differences in capability and preparedness is properly raised in arguments of job-relatedness and business necessity (which, according to Footnote 18, the court regarded as interchangeable terms); these arguments do not rebut the disparities themselves, which are shown to be statistically and practically (given the magnitude of the alleged shortfall in black and Hispanic hiring) significant.  Finally, there was no reason to exclude statistical significance tests in preference to the 80% Rule.  The court found that the Uniform Guidelines (29 C.F.R. § 1607.4D) contemplated alternative statistical tests, not just the 80% Rule to show adverse impact.

Having found that plaintiffs had established a prima facie case, the court turned to the matter of business necessity.  Job analysis was conducted for Exam 7029 and used again for Exam 2043.  The job analysis resulted in 18 abilities and 21 task clusters.  Abilities were linked to clusters.  However, only 9 of the 18 abilities were deemed testable.  Panels of incumbent firefighters were used as item writer for the two tests.

The city was in procedural trouble from the start, having failed to submit two expert declarations until after the close of discovery.  The court refused to allow one and limited the other.  The court then proceeded to examine the city’s business necessity in light of the Second Circuit’s landmark opinion, Guardians Association of the New York City Police Department, Inc. v. Civil Service Commission (630 F.2d 79, 2d Cir. 1980).  The analysis was as follows:

Job analysis.  As in Guardians, the city had condensed a large list of tasks into a more manageable number of task clusters, and then proceeded to link abilities to task clusters.  Guardians had pointed out that there would be no way of determining what abilities were important for what specific tasks.  Moreover, statements by the subject matter experts (SMEs) involved in linkage indicated that some did not understand the abilities.  Nor did the city ascertain whether the abilities and the tasks to which they were applied could be learned on the job.

Test construction. The city did not use professional test developers to construct the test items.  Input from SMEs was essential for the content of the questions; input from professionals was needed to write the questions.  There was no pilot-testing of the draft tests.

Direct relationship to the job.  The city presented evidence that the nine cognitive abilities that it found to be testable were job-related.  Plaintiffs showed that questions allegedly measuring the same ability correlated more strongly with other abilities.  Factor analysis indicated that the tests had one general cognitive factor and a week second factor that apparently did not correspond to anything in the test plan.  Accordingly, it was not clear what the content of the tests was.  However, the court allowed that some of the cognitive abilities being tested were relevant.

Representativeness.  The court found that the city had done little beyond stating that one of the task clusters had cognitive demands.  Of the nine non-testable abilities, two involved oral abilities.  There was no explanation for the omission of the other seven. Non-cognitive abilities were ignored in the job analysis.  O*NET and other sources indicated that non-cognitive abilities were important for the job, and there were means to test these that were available during the time of the disputed tests.  Reading level of the tests was not checked by the city; it was checked by the plaintiffs and alleged to be too high.

Scoring.  The city set cut-off scores for Exam 7029 based on anticipated need alone—something that Guardians said not to do.  The cut-off for Exam 2043 was a default based on civil service rules.  Plaintiffs presented figures indicating that a substantial number of applicants who took both tests had passed one but not the other, further indicating the arbitrary nature of the cut-offs.  Ranking of applicants who took the two tests could vary radically.  Because of bunching up of scores (as in Guardians) missing a single question could alter ranking considerably.

The second order came on January 13, 2010.  Here the court found that the city had deliberately discriminated against blacks and Hispanics, maintaining “a pattern and practice of intentional discrimination against blacks, in violation of the Fourteenth Amendment to the United States Constitution, Title VII of the Civil Rights Act of 1964, and State and City Human Rights Laws” (p.2).  The court repeated and amplified some of its conclusions in the summary judgment order.  These are worth noting, but the fight here (assuming that the city appeals) is going to be on the legal issues. 

The court was obviously tired of more than 30 years of severe under-representation of blacks in FDNY and what it viewed both as sloppy testing and sloppy legal defense.  The decision references extensively a similar case from 1973.  The remedies from that case included a hiring quota and development of new tests.  The city obtained federal money for constructing new tests but subsequently canceled the development contract for financial reasons. Hiring quotas ended in 1977.  Blacks were about 4% of the fire department in 1971; in 2002, 2.6%; at the start of the current litigation in 2007, 3.4%.  Under-representation was far greater in FDNY than in other large city fire departments or in other NYC agencies.  For example, in 2009 blacks were 3.4% of firefighters, but 18% of rank-and-file police officers.

But the decision follows procedure and reasoning that seems problematic.  To begin, the court’s action is based on the city’s development and use of tests between 1999 and 2007 which “had discriminatory effects on certain minority applicants” (p. 2).  While it was certainly known that there was adverse impact, adverse impact per se is not discrimination.  There was no finding that the use of the tests was unlawful until 2009—after use of the tests was discontinued.  That does not absolve the city of its liability for use of the tests, but it does make one wonder how use of a presumably neutral selection procedure—a presumption that distinguishes disparate impact from disparate treatment—got converted to deliberate discrimination.  To find that an employer engaged in a pattern or practice of discrimination is to say that unlawful discrimination was standard operating procedure.  Perhaps the clearest statement of the court’s position is as follows: “The City’s willingness to treat black applicants differently—to tolerate adverse outcomes against one race that it would not tolerate against another—is, if not the textbook definition of discriminatory intent, its nearly indistinguishable synonym” (p. 49).  The problem would seem that disparate treatment becomes a nearly indistinguishable synonym for disparate impact.  The adverse impact is clear; the city’s differential toleration by race of adverse impact is not.  While the court specifically acknowledged that there are circumstances when adverse impact is not discriminatory, under some circumstances it is not only discriminatory, but deliberately so.  That concept is hardly new, but it has generally involved the employer’s use of a seemingly neutral procedure as pretext for discrimination.  Here, the adverse impact per se seems to be the deliberate discrimination. 

The Federal Government complaint did not include disparate treatment; the interveners (Vulcan Society and individuals) had a complaint that did, regarding blacks only.  It is on this part of the complaint that the court was now ruling.  The court dismissed state law claims against the fire department and the administrative services department that had prepared the tests; the city, not its governmental units, is the proper entity to name in a suit.  Title VII claims against Mayor Michael Bloomberg and Fire Commissioner Nicholas Scoppetta were dismissed; Title VII does not impose liability on individuals.  The city moved to dismiss the remaining charges for failure to state a claim that could be addressed by the court.  The court used its discretion to convert a motion to dismiss to a motion for summary judgment and ruled against the city.

The court applied the familiar McDonnell Douglass burden-shifting formula and found that the plaintiffs had established a prima facie case of disparate treatment by virtue of having established a prima facie case of disparate impact.  This was further bolstered by “extensive historical, anecdotal, and testimonial evidence” (p. 31).  There was also the disparity between FDNY and other fire departments and other NYC uniformed services.

Some statistical disparities that are so egregious that they speak to intent; as the court noted, use of statistics in disparate treatment cases is long established.  But what is different here in contrast to other disparate treatment cases involving statistics (such as Teamsters, cited frequently by the court; see citation below), is that in other cases there would be reliance on a clearly discriminatory practice causing the statistical disparities; here the discriminatory process is reliance on scores from a (presumably) neutral mechanism. 

The reference to FDNY statistics outside the time of the alleged discrimination in this case, as well as reference to the racial composition of other fire departments and other NYC agencies might have been expected to draw objections on relevance from the city.   Here, the court found that the city brought forward nothing to contradict the statistically-based prima facie case.  The court discounted statements by the test developers that they did not have discriminatory intent; it did not help the city that they also indicated that they were not familiar with Guardians.  The court also discounted the city’s efforts to recruit more black applicants and providing exam opportunities to go from paramedic to firefighter. 

“On the other hand, a showing that the Exams were constructed properly—that is, that they test for relevant skills and properly differentiate between better and worse candidates—would be highly relevant to the City’s defense because it would support an inference that the City’s actual intent in enforcing the pass/fail and rank-ordering policies was to select the best candidates, and that the interveners’ prima facie showing of disparate impact is merely the unfortunate by-product of a legitimate, neutral policy” (pp. 38-39; emphasis in original).  So, does the city win or lose on disparate treatment depending on how it fared on disparate impact?  If the city really believed that its testing was technically adequate, however ill-founded that belief, does a finding of inadequacy after the tests were discontinued establish that discrimination was standard operating procedure?  Does the court assume that the persistence of adverse impact indicates that the selection procedures must be purposefully maintaining adverse impact, as evidenced by its citing International Brotherhood of Teamsters v. United States (431 U.S. at 340 n.20, 1977; as quoted in the current case):

Statistics showing racial or ethnic imbalance are probative in [disparate-treatment cases] because such imbalance is often a telltale sign of purposeful discrimination; absent explanation, it is ordinarily to be expected that nondiscriminatory hiring practices will in time result in a work force more or less representative of the racial and ethnic composition of the population in the community from which employees are hired.

Our ordinary expectations may have diminished since 1977 with regard to elimination of racial and ethnic under-representation simply by the use of nondiscriminatory hiring practices.  In any event, Teamsters was about requests for assignments being denied or ignored on the basis of race and ethnicity, thus creating statistical racial imbalances.  It was not about impact resulting from a presumably neutral mechanism. 

Finally, while the mayor and fire commissioner escaped personal liability due to qualified immunity as city officials, the court found that they were deliberately indifferent to the adverse impact against blacks.  Apparently following mixed-motive doctrine, the court stated that the city had the burden of putting forth evidence that the city was not motivated by discriminatory purpose at all (emphasis in the decision).  The city did not do so, and therefore the interveners were entitled to judgment on their equal protection and § 1981 claims as a matter of law.  Since state law claims are analyzed the same as Title VII claims, it followed that the city was in violation of these as well.

Finally, the court put forth an outline of the proposed remediation on January 21, 2010.  The remedial action is to be finalized in a conference to be held in February 2010.  Remediation involves relief for firefighter applicants affected by the city’s discriminatory examinations and the implementation of a fair and job-related test.  With regard to the latter issue the city’s replacement for the tests in this case with Test 6019 came under discussion.  The city can continue to use it, but subject to review and with liability for any disparate impact.  The court noted that submissions from the plaintiffs indicated that the new test had some of the same scoring issues as the previous tests.  The court declined to impose hiring quotas.

Guardians was a key precedent for the current case.  Besides providing a classic commentary on content-related validation, it overturned a district court’s finding of purposeful discrimination in the NYC police department.  Perhaps history will repeat itself with this fire department case regarding what constitutes purposeful discrimination.  However, in re-reading that old case this writer was struck by the same thing that struck Judge Garaufis in his review of fire department testing in this case: history seems to have been repeating itself with the city’s use of vague arguments about statistical significance and lack of ability in the applicant pool, questionable issues in test development, and scoring that left huge numbers of applicants separated by just a few points.  Of course, the comments here are all based on the views put forth by the courts; the city may have more of an explanation.  Unfortunately for the city, there was not enough explanation put forth to avoid summary judgment on all issues.

The court distinguished this case from Ricci in its July order.  The case presents the converse of Ricci.  Here presumably NYC should have pulled the plug on its tests.  Did it have a strong basis in evidence to do so?  For the court, the answer seems affirmative; the city should have been aware of Guardians and that its efforts did not measure up to the standards described there.  The New Haven and New York decisions would seem to put employers in an interesting bind where either upholding or voiding test results with adverse impact can be taken as deliberate discrimination.

Ricci, continued.

Bridgeport

The City of Bridgeport, CT settled a discrimination suit involving 12 firefighters (white, with one Hispanic).  The promotional test for lieutenant had been held in December 2006 and results promulgated in September 2007.  However, the city decided to change the weights of the test components in July 2008 from 50% written, 45% oral, and 5% seniority to 75% written and 25% oral.  The re-weighting was more favorable to non-whites.  The formerly high-ranked whites sued.

The settlement gives the plaintiffs promotions, back pay and seniority, and partial legal fees.  The cost to the city was reported at over $100K. 

During the New Haven hearings that led up to the civil service board declining to certify the test results involving Ricci et al., it was mentioned that Bridgeport weighted test components differently from New Haven and seemed to have better non-white representation. 

I/O Solutions

Chad Legel, the president of I/O Solutions (company that developed the promotional tests in Ricci) tells his side of the story in a letter to the editor in The Industrial-Organizational Psychologist, January 2010 on the SIOP website (www.SIOP.org).

OTHER COURT CASES

Equal opportunity sexual harassment.

The Eleventh Circuit considered en banc a case of sexual harassment and unanimously reversed summary judgment for the employer (Reeves v. C.H. Robinson Worldwide, Inc., 11^th Cir., No. 07-10270; 1/20/2010). While the case does not break new ground, it provides a definitive statement on what is or isn’t harassment.

Reeves worked as a transportation sales representative for Robinson, a shipping company.  Her job was phone-intensive with carriers, customers, truck drivers, and dispatchers.  She was the only woman in her area of cubicles.  The court noted that Reeves was a veteran of the Merchant Marine and was no stranger to coarse language.  The loud language and pictures in her work area could be described as gross, in this writer’s opinion.  She complained about it to her co-workers, supervisor, and two company executives, to no avail.  Eventually she quit and sued the company.  District court granted summary judgment to the company, finding that the behavior in question was not directed at Reeves in particular, but at all employees in the area, male or female.  The Court of Appeals reversed, and then re-heard the case en banc, again reversing the district court.

The court stated four principles for sexual harassment discrimination:

1. The plaintiff must show the employer discriminated on the basis of the plaintiff’s membership in a protected group, and that the offensive conduct was severe or pervasive enough to alter the terms or conditions of employment;
2. Title VII is not a civility code;
3. Workplace conduct cannot be viewed in isolation, but cumulatively, and in its social context;
4. The plaintiff can prove hostile work environment by showing discrimination directed at the plaintiff’s protected group, even if not individually singled out in the offensive conduct.

Either severity or pervasiveness is sufficient to establish a violation.  Evaluation includes frequency; severity; physically threatening or humiliating, or mere offensive utterance; and unreasonable interference with an employee’s work performance.  The plaintiff must prove that the conduct was subjectively severe or pervasive, a perception that must also be objectively reasonable.  Sexual language and discussions that are truly indiscriminate do not per se establish a violation.  Harassment with sexual connotations is not necessarily sexual harassment.  Nevertheless, a member of a protected group cannot be forced to endure objectionable conduct that is gender-specific in the workplace, just because the workplace may be otherwise rife with generally indiscriminate vulgar conduct.

In the current situation, it was enough for Reeves to hear on a daily basis language from her co-workers that indicated that they viewed women negatively.  Evidence that objectionable conduct was directed at the protected group may constitute actionable harassment even if not explicitly directed at the plaintiff.  The context could reasonably be taken as indicating that the conduct was not simple teasing, offhand comments, or isolated incidents.

The claim the Reeves’ co-workers were using such language before Reeves arrived is irrelevant.   Once Reeves was there, she should not be disparately subjected to degrading conditions to which her male co-workers were not subjected.  (Note that the issue is not disparate exposure to the conduct, but on the disparate effect of the conduct.)

In summary, a reasonable jury could find that Reeves was discriminated against, and so to the jury the case will go unless settled.

In another take on equal opportunity for being victimized, it might be noted that men are gaining on women in filing sexual harassment complaints, including complaints of being harassed by other men.  Gesaman (2010) reported that EEOC has noted a growing trend in male-on-male harassment.  The EEOC does not always track the gender of the harasser, but charges filed by males between 1992 and 2008 went from 8% to 16% of all sexual harassment complaints.  EEOC recently settled a case of male-on-male harassment in Phoenix against The Cheesecake Factory for six employees totaling $340K. [Source: Gesaman, K. (2001) Abuse of power.  Newsweek, 1/13/2010.  Retrieved 1/14/20110 from www.newsweek.com/id/230677.]

Recession?  What recession?

There’s one organization that is seeing an increase in business during hard economic times. EEOC recorded 93,277 new discrimination charges in Fiscal Year 2009 (ending 9/30/2009), the second highest total in its history.  Race and retaliation tied at 36% of the total, followed by sex at 30%.  The agency resolved 85,980 charges and garnered $376M in monetary relief.  EEOC also filed 281 lawsuits and resolved 321 cases; these figures include only substantive issues and not procedural actions such as subpoena enforcement.

While there is a general feeling that a bad employment situation makes for more discrimination charges, it is not clear if other factors (more workers’ awareness of EEO rights, more diverse workforces, more bad behavior by employers) are driving the increase.

NOTED IN PASSING

Federal Employee EEO Process Overhaul.

The EEOC has announced long-awaited changes to the federal employee process, which operates differently from the private sector process.  Deputy Legal Counsel Peggy Mastroianni indicated that the proposed changes, published for comment in the Federal Register on December 21,

Caregivers as a Protected Class?

This is the title of a report authored by Stephanie Bornstein and Robert Rathmell of the Center for WorkLife Law, University of California Hastings College of Law.  The report details the development of 63 local laws in 22 states that go beyond state and federal EEO laws to give protection specifically to home caregivers.  Twenty-three of the laws allow for recovery of monetary damages.  Family responsibility discrimination (FRD) can create significant liabilities.  A case noted in the report involved a woman (a medical salesperson) who was fired after rescheduling an appointment because her daughter was ill.  Under a Chicago law, she was awarded $300K in compensatory and punitive damages plus attorney’s fees.  The EEOC had issued guidance in 2007 touching on aspects of FRD as covered by federal sex and disability discrimination laws.  The report indicated that in the burgeoning area of FRD suits, plaintiffs are mostly bringing charges under existing, non-FRD law.  As noted here previously, one area of expanded legal coverage is “sex plus” theories that combine a recognized protected class (sex, usually female) with a characteristic associated with the protected class but not a class itself (in this case, caregiver).

FRD laws tend to use language from local housing anti-discrimination ordinances that prohibit discrimination based on familial or parental status.  ·

The study is available at www.worklifelaw.org/pubs/LocalFRDLawsReport.pdf.

“Inflexible Leave” Policy under EEOC Attack.

Gordon Gibb of the Disability Insurance Resource Center commented on several EEOC suits involving employers with policies that terminate employees if they are not fully able to return to their jobs after a specified period of disability leave.

EEOC and DOJ in Enforcement Alliance.

Thomas Perez, Assistant Attorney General for Civil Rights, announced on December 18 that the Department of Justice and the EEOC were working toward a memorandum covering joint enforcement.