Opinion
ACTION I No. 96-CV-0374 (FB) (RML), ACTION II No. 02-CV-0256 (FB) (RML).
April 20, 2007.
Esther Gwen Tamburo, Esq., United States Department of Justice, Employment Litigation Section, Washington, DC, for the Plaintiff.
Lawrence J. Profeta, Esq., The City of New York Law Department, New York, NY, for the Defendants.
Michael E. Rosman, Esq., Center for Individual Rights, Washington, DC, for Intervener Brennan, et al.
Matthew B. Calangelo, Esq., NAACP Legal Defense Fund Educational Fund, Inc., New York, NY, for Intervenor Arroyo, et al.
Emily J. Martin, Esq., American Civil Liberties Union Women's Rights Project, New York, NY, for Intervenor Caldero, et al.
MEMORANDUM ORDER
Familiarity with the Court's Memorandum and Order of September 11, 2006 ("September 11th M O"), United States v. New York City Bd. of Educ., 448 F. Supp. 2d 397 (E.D.N.Y. 2006), is presumed; in the September 11th M O, the Court (1) denied the motion of the New York City Board of Education ("the Board") to enter its settlement agreement ("the Agreement") with the United States of America as a consent judgment; (2) certified a class "comprising all custodial employees whose layoff-protection rights have been adversely affected by the grant of seniority benefits to beneficiaries who are non-victims of discrimination"; and (3) made the following declarations regarding the validity of the Agreement under Title VII and the Fourteenth Amendment:
(a) As for the testing claims, all beneficiaries who received relief, except Luis Torres, are entitled to the seniority benefits provided by the Agreement as to transfer rights and temporary care assignments; as for layoffs, however, the grant of seniority benefits to beneficiaries who are not victims of discrimination violates Title VII and the Fourteenth Amendment.
(b) As for the recruiting claim, the grant of seniority benefits for any purpose to black, Hispanic and Asian males who were not victims of discrimination violates the Fourteenth Amendment. As for the women, they are entitled to the seniority benefits provided by the Agreement as to transfer rights and temporary care assignments; as for layoffs, however, the grant of seniority benefits to women who are not victims of discrimination violates Title VII.
(c) [Ciro] Dellaporte is not a member of a protected class; therefore, he is not entitled to any relief.Id. at 447. In addition, the Court identified three issues requiring an evidentiary hearing:
(a) which, if any, of the beneficiaries, in addition to Lloyd Bailey, Joseph Christie, Belfield Lashley, Gilbert Rivera, Peter Robertin, Felix Torres and Mayra Zephrini (Cintron), are actual victims of discrimination and received the relief to which they were entitled;
(b) whether the results of Exam 8206 in respect to Luis Torres satisfy the evidentiary standards for establishing discrimination under Title VII and the Fourteenth Amendment; and
(c) whether John Mitchell and Eric Schauer were denied transfers in favor of particular individuals who impermissibly received retroactive seniority.Id. at 447-48.
On September 25, 2006, the Brennan Intervenors moved for reconsideration, arguing (1) that — in addition to Luis Torres — Pedro Arroyo, Kevin LaFaye and Fidel Seara are not entitled to any benefits under the Agreement if the Court concludes that there was insufficient evidence that the 8206 Exam had a disparate impact on Hispanics; and (2) that the certified class should be redefined. On September 26, 2006, the Caldero Intervenors moved for (1) reconsideration of the Court's conclusion that the seven black and four Hispanic males designated as recruiting-claim beneficiaries in the United States' Relief Chart were not entitled to the benefits they received under the Agreement, and (2) clarification that the Agreement's award of permanent employment status has not been found unlawful. On November 21, 2006, the Court held an evidentiary hearing on the issue of whether the 8206 Exam had a disparate impact on Hispanics.
John Brennan, James G. Ahearn, Dennis Mortenson and Scott Spring.
Janet Caldero, Celia I. Calderon, Martha Chellemi, Andrew Clement, Kristen D'Alessio, Laura Daniele, Charmaine Didonato, Dawn L. Ellis, Marcia P. Jarrett, Mary Kachadourian, Kathleen Luebkert, Adele A. McGreal, Marianne Maousakis, Sandra D. Morton, Maureen Quinn, Harry Santana, Carl D. Smith, Kim Tatum, Frank Valdez and Irene Wolkiewicz.
Hearings on the other two fact issues identified in the September 11th M O proved unnecessary. With respect to issue (a), regarding actual victims, the Arroyo Intervenors (Pedro Arroyo, Jose Casado, Celestino Fernandez, Kevin LaFaye, Steven Lopez, Anibal Maldonado, James Martinez, Wilbert McGraw, Silvia Ortega de Green and Nicolas Pantelides) stated at an unreported status conference that they do not intend to offer proof of additional actual victims of testing discrimination; as for recruiting discrimination, there is, as explained infra pp. 24-25, insufficient evidence to warrant a hearing on actual victims. With respect to issue (c), regarding Mitchell and Schauer's claims that they were denied transfers, the parties agree that Mitchell lost a transfer to Pedro Arroyo and that Schauer lost a transfer to Kevin LaFaye, see Letter of Michael Rosman (Oct. 31, 2006), ¶ 5; under the holdings of the September 11th M O, the awards to Arroyo and LaFaye, both of whom took a challenged exam, comported with Title VII and the Fourteenth Amendment.
The issues raised in the parties' post-decision motions and at the November 21st hearing are now meet for disposition.
THE 8206 EXAM
As explained in the September 11th M O, an employer seeking to justify an affirmative-action plan in the face of a challenge under Title VII "need point only to a conspicuous imbalance . . . in traditionally segregated job categories," New York City Bd. of Educ., 448 F. Supp. 2d at 424 (quoting Johnson v. Transp. Agency, Santa Clara County, 480 U.S. 616, 630 (1987)), and to survive a Fourteenth Amendment challenge, an affirmative-action plan must be supported by a "strong basis in evidence," meaning evidence "approaching a prima facie case of a constitutional or statutory violation," id. at 435 (quoting City of Richmond v. J.A. Croson Co., 488 U.S. 469, 500 (1989)); a disparity that satisfies the "strong basis in evidence" standard under the Fourteenth Amendment necessarily satisfies the "conspicuous imbalance" standard under Title VII. See id. at 424 ("[A]s Johnson makes clear, the requisite magnitude of the disparity is less [under the "conspicuous imbalance" test than under "strong basis in evidence" test], although just how much less is unclear."). As further explained in the September 11th M O, one way of making out a prima facie case of disparate impact is to show a violation of the EEOC's "80 percent" rule of thumb, under which "a selection rate for any race, sex, or ethnic group which is less than four-fifths (4/5) (or eighty percent) of the rate for the group with the highest rate will generally be regarded by Federal enforcement agencies as evidence of adverse impact," id. at 425 (quoting 29 C.F.R. § 1607.4(D)); alternatively, a prima facie case can be established if "the plaintiff can show a statistically significant disparity of two standard deviations" because "[c]ourts generally consider this level of significance sufficient to warrant an inference of discrimination." Id. (quoting Smith v. Xerox Corp. 196 F.3d 358, 365 (2d Cir. 1999), overruled on other grounds by Meacham v. Knolls Atomic Power Lab., 461 F.3d 134 (2d Cir. 2006)).
At the November 21st hearing, the United States presented testimony from Bernard Sisken, Ph.D., a statistician with extensive experience in employment discrimination. Dr. Sisken, together with another statistician, Leonard Cupingood, Ph.D., prepared the November 1997 report, referenced in the September 11th M O, analyzing the Board's written examinations that the United States relied on when the Board was contesting the claim that the tests had an adverse impact on blacks and Hispanics; that report was received in evidence at the hearing.
To rebut Dr. Siskin's testimony, the Brennan Intervenors presented the testimony of James Sharf, Ph.D., an industrial psychologist with extensive experience in developing job selection tests; from 1974 to 1978, he was the chief psychologist at the EEOC, where he played a major role in negotiating the agency's Uniform Guidelines on Employee Selection Procedures. In January 2003, the Brennan Intervenors retained Dr. Sharf to review and analyze the expert reports prepared on behalf of the United States and the Board; in a report dated January 30, 2003, also received in evidence at the hearing, Dr. Sharf concluded that qualified Hispanics passed the 8206 Exam at more than 80% of the white pass rate and, therefore, the exam did not have a disparate impact on Hispanics.
A. The United States' Evidence
In his 1997 report and at the hearing, Dr. Siskin presented two statistical analyses of the exam. He first compared the percentage of Hispanics and whites that had achieved the minimum passing score of 70, without regard to whether they met other qualifications, such as prior experience, English-language proficiency and licensure requirements. That comparison — 69% (29 out of 42) of Hispanic test-takers, as compared to 85.6% (273 out of 319) of white test-takers — equated to a Hispanic pass rate of 80.61% of the white pass rate. Dr. Sisken testified that this disparity was statistically significant, representing 2.51 units of standard deviation, meaning that the probability that the disparity was due to chance was "[a]bout one in a hundred."
As Dr. Sharf explained, it is "not at all unusual" for a public employer such as the Board to use a score of seventy as the pass-fail cut-off for civil-service examinations. Tr. at 72.
"Tr." refers to the transcript of the November 21st hearing.
In Dr. Siskin's second analysis, he compared the percentage of Hispanic and white test-takers, again regardless of whether they met the other qualifications, that had achieved a score of 85 or higher; Dr. Siskin referred to this as the "effective" passing score because "[a]nyone that scored below eighty-five could not have been hired, no matter what. He would have essentially been knocked out by the test." Id. at 14. Dr. Siskin found that 11.9% (5 out of 42) of Hispanic test-takers achieved the effective passing score, as compared to 31.7% (101 out of 319) of white test-takers; thus, Hispanics effectively passed the exam at only 37.54% the rate of whites. Dr. Siskin testified that this disparity was also statistically significant, representing 2.46 units of standard deviation.
Dr. Siskin also compared the nominal and effective pass rates of qualified test-takers (i.e., those who took the exam and met the other qualifications) by subtracting from his analyses unqualified test-takers (i.e., those who took the exam but lacked one or more of the other qualifications); to determine the number of unqualified test-takers, he added the number of unqualified test- passers to the number of unqualified test- failers. For unqualified test-passers, Dr. Siskin used the results of the Board's normal hiring process, which included a post-exam qualifications review of test-passers. Because, as Dr. Siskin conceded, excluding only unqualified test-passers could be taken to mean that all test-failers were qualified, see Tr. at 24 ("The problem with that, of course, is [that] it doesn't remove the failures who were not qualified."), he estimated the number of unqualified test-failers based on three alternative assumptions: (1) that unqualified test-failers occurred in the same proportion as unqualified test-passers, (2) that the proportion of unqualified test-failers was double the proportion of unqualified test-passers, and (3) that the proportion of unqualified test-failers was triple the proportion of unqualified test-passers.
Dr. Siskin rejected the alternative methodology of calculating the number of unqualified test-failers used by Dr. Phillip Bobko, the expert retained by the Board when it was contesting the United States' claims. Dr. Bobko derived the number of unqualified test-failers from data generated from the Board's after-the-fact review of test failers' qualifications, which was conducted solely for the purpose of litigation; using that methodology, he opined that qualified Hispanics passed the exam at 89.3% the rate of qualified whites. Dr. Siskin's reason for rejecting Dr. Bobko's methodology for determining unqualified test-failers was that a significant number (according to Dr. Siskin, 20-25%) of test-passers deemed unqualified succeeded in getting that determination overturned through an administrative appeals process, an option that was not available to test-failers.
Dr. Siskin also testified that a study conducted for litigation raises the risk that the evaluators intentionally or unintentionally made their determinations with an eye towards reaching a particular result.
The results of Dr. Siskin's two analyses are summarized in the following two tables:
These tables are recapitulations of the tables presented in Dr. Siskin's 1997 Report and at the November 21st hearing, with the exception of the fractions in parentheses, which the Court has added to provide a more concrete context for Dr. Siskin's percentages.
Table 1: "Nominal" Passing Score (70 or Above)
Comparator Group Hispanic White Hispanic Pass No. of Pass Rate Pass Rate Rate ÷ White Standard Pass Rate Deviations passers takers takers All test-takers 69.0% 85.6% 80.61% 2.51 (29/42) (273/319) Excluding unqualified 64.9% 83.5% 77.72% 2.51 (24/37) (233/279) Excluding unqualified 70.6% 85.3% 82.77% 1.95 (Assumes proportion of unqualified (24/34) (233/273) failers = proportion of unqualified passers) Excluding unqualified 77.4% 87.6% 88.36% 1.29 (Assumes proportion of unqualified (24/31) (233/266) failers = 2x proportion of unqualified passers) Excluding unqualified takers 85.7% 90.0% 95.22% 0.37 (Assumes proportion of unqualified (24/28) (233/259) failers = 3x proportion of unqualified passers)Table 2: "Effective" Passing Score (85 or Above)
Comparator Group Hispanic White Hispanic Pass No. of Pass Rate Pass Rate Rate ÷ White Standard Pass Rate Deviations passers takers takers takers All test-takers 11.9% 31.7% 37.54% 2.46 (5/42) (101/319) Excluding unqualified 5.4% 32.3% 16.72% 3.19 (2/37) (90/279) Excluding unqualified 5.9% 33.0% 17.88% 3.05 (Assumes proportion of unqualified (2/34) (90/273) failers = proportion of unqualified passers) Excluding unqualified 6.5% 33.8% 19.23% 2.92 (Assumes proportion of unqualified (2/31) (90/266) failers = 2x proportion of unqualified passers) Excluding unqualified 7.1% 34.7% 20.46% 2.76 (Assumes proportion of unqualified (2/28) (90/259) failers = 3x proportion of unqualified passers)As Table 1 shows, the nominal pass rate of Hispanics fell below the EEOC's "80 percent" rule of thumb only if one (a) excluded unqualified test-passers, and (b) unreasonably assumed that there were no unqualified test-failers. If one assumed that failers were unqualified in at least the same proportion as passers, Hispanics achieved a nominal passing score at between 82.77% and 95.22% the rate of whites; Dr. Siskin conceded that those disparities (which equate to standard deviations of between 1.95 and 0.37) are not statistically significant. See Tr. at 21 ("If you make this assumption [regarding the number of unqualified test-failers], it [is] not statistically significant."); see also id. at 25 (acknowledging that a standard deviation of 1.96 is a "bright line test" for statistical significance).
By contrast, Table 2 shows that the disparities in effective scores fell well below 80% under any of Dr. Siskin's three assumptions about the number of unqualified test-failers. In addition, Dr. Siskin testified that these disparities were all statistically significant; he testified, for example, that a fair test would yield a standard deviation of 3.05 (the standard deviation for effective pass rates assuming the proportion of unqualified test-failers was equal to the proportion of test-passers) "roughly three in a thousand times." Tr. at 27. Even in the extreme case — a proportion of unqualified test-failers equal to three times the proportion of unqualified test-passers — qualified Hispanics achieved an effective passing score at only 20.46% the rate of qualified whites, yielding a statistically significant standard deviation of 2.76.
On cross-examination, the Brennan Intervenors asked Dr. Siskin about one of the other tables in his 1997 report; according to that table, which was not discussed on direct examination, no Hispanics were hired from the eligibility list for the exam. Relying on a list provided by the Board during discovery, the Brennan Intervenors posited that Frank Zapata, Jr. and Jorge Trifun were hired based on the exam, and that both those individuals were classified as Hispanic males. Dr. Siskin conceded that if Zapata and Trifun are Hispanic, they were possibly "misclassified [as white] in every single study [he] did of the 8206 test." Tr. at 44.
B. The Brennan Intervenors' Evidence
Dr. Sharf testified that comparisons of pass rates should be restricted to qualified test-takers, agreeing with Dr. Siskin that to "assume[] that all test failers were `otherwise qualified'" was "completely unrealistic . . . given the numbers of test passers who were contemporaneously deemed not otherwise qualified." Brennan Intervenor's Ex. 2 (Aff. of James C. Sharf (Oct. 18, 2004)) ¶ 11. Dr. Sharf also "agreed with [Dr. Siskin's] assumptions when he assumed that the qualification rate for those passing was comparable to those not passing and then a double multiple and triple multiples." Tr. at 96.
Dr. Sharf testified that the statements in his October 2004 affidavit "reflect [his] current views about the adverse impact of the 8206 test on Hispanics." Tr. at 79.
Dr. Sharf testified, however, that only nominal pass rates should be considered in assessing whether the exam had a disparate impact on Hispanics, rejecting Dr. Siskin's effective-score analysis on the ground that it "tells you more about the labor market than it does about the test," Tr. at 71, in that the score will vary depending on the number of applicants and the length of time that the resulting eligibility list remains in effect, a variable that may not be known for several years after the test is administered. He further testified (1) that Dr. Siskin used the effective-score analysis only in his 1997 report, and not in his rebuttals to the other experts, and (2) that the small sample sizes involved made the analysis unreliable, particularly in light of the possible misclassification of Zapata and Trifun. Dr. Sharf did not, however, independently run the data to determine the effect of the possible misclassification.
C. The Parties' Contentions 1. Nominal-Score Analysis
The United States and the Arroyo Intervenors argue that Dr. Siskin's comparison of the nominal scores of all test-takers (summarized in the first row of Table 1) is sufficient to establish a prima facie case that Exam 8206 had a disparate impact on Hispanics. In response, the Brennan Intervenors argue that since custodial positions required qualifications other than a passing exam score, the comparison of all test-takers, both qualified and unqualified, is insufficient to establish a prima facie case of past testing discrimination warranting corrective affirmative action.
Although the parties frame the issue in terms of evidence sufficient to make out a prima facie case, it bears repeating that the actual standard is whether there was a strong basis in evidence — that is, evidence approaching a prima facie case — that the exam had a disparate impact on Hispanics.
The Court agrees with the Brennan Intervenors. As noted in the September 11th M O, Johnson instructs that "[w]here a job requires special training . . . the comparison should be with those in the labor force who possess the relevant qualifications." New York City Bd. of Educ., 448 F. Supp. 2d at 424 (quoting Johnson, 480 U.S. at 632). Two years after Johnson, the Supreme Court applied this instruction in Wards Cove Packing Co. v. Atonio, 490 U.S. 642 (1989), holding that a statistical comparison between the racial compositions of an unskilled work force and a skilled work force did not make out a prima facie case of disparate impact because "the [unskilled] work force in no way reflected the pool of qualified job applicants or the qualified population in the labor force." Id. at 651, 109 S.Ct. 2115 (emphasis in original; internal quotation marks omitted). As the Court explained, "[i]f the absence of minorities holding . . . skilled positions is due to a dearth of qualified nonwhite applicants (for reasons that are not [the employer's] fault), [the] selection methods or employment practices cannot be said to have had a `disparate impact' on nonwhites," id. at 651-52, 109 S.Ct. 2115; rather, "[i]t is . . . a comparison [] between the racial composition of the qualified persons in the labor market and the persons holding at-issue jobs [] that generally forms the proper basis for the initial inquiry in a disparate-impact case." Id. at 650-51, 109 S.Ct. 2115. Thus, the relevant comparison here must be between qualified test-takers.
Drs. Siskin and Sharf agree that excluding the number of unqualified test- passers, without excluding any test- failers (i.e., the second row of Table 1), does not yield a reliable measure of qualified test-takers. Both experts also agree that Dr. Siskin's other assumptions are appropriate estimates of the number of unqualified test-failers; however, under those assumptions, Dr. Siskin's nominal-score analysis reveals that qualified Hispanics passed the exam at more than 80% the rate of qualified whites, and that the disparities that do exist are not statistically significant. While based on the first assumption, the Hispanic pass rate was 82.77% that of whites, with a standard deviation of 1.95, the Court is not inclined to conclude that either the pass rate or the standard deviation approaches a prima facie case given the lack of judicial guidance as to where or how one draws the line marking the nuanced difference between a prima facie case and evidence approaching a prima facie case.
While Dr. Bobko offered an alternative method for calculating unqualified test-failers, based on the Board's after-the-fact review of experience papers, the Court credits Dr. Siskin's opinion, which was not challenged by Dr. Sharf, that the reliability of that review was undermined by the fact that a significant number of test-passers succeeded in getting an adverse qualifications determination overturned.
2. Effective-Score Analysis
Having concluded that the relevant comparison must be restricted to qualified test-takers, the Court confines its consideration of Dr. Siskin's effective-score analysis to rows 3 through 5 of Table 2, which reflect what both Dr. Siskin and Dr. Sharf agreed were appropriate assumptions about the number of unqualified test-takers. The Brennan Intervenors do not dispute that the effective-score analysis revealed that the effective pass rate of qualified Hispanics was less than 80% the rate of qualified whites; however, they raise several challenges to the methodology of the analysis and the underlying data.
a. Methodology
With respect to methodology, the Brennan Intervenors argue that the analysis does not isolate the disparate impact of the exam as a screening device since it does not control for other variables in the hiring process, such as the size of the applicant pool, the length of time the eligibility list for the exam remained open, and the vagaries of the interview process. They also argue that it relies on sample sizes too small to allow for meaningful statistical analysis. Finally, they argue that it is a belated resurrection of a theory of disparate impact that Dr. Siskin did not rely on in his rebuttal reports criticizing the analyses of Drs. Bobko and Sharf.
The Brennan Intervenors' first objection to the effective-score analysis must fail under Waisome v. Port Authority of New York and New Jersey, 948 F.2d 1370 (2d Cir. 1991). At issue in Waisome was a promotional examination that served two functions:
First, it served as a pass-fail mechanism that required each candidate to obtain a passing score before moving on to the next step of the examination. Second, candidates' scores on the written test were factored into the composite scores that were then used to compute a candidate's rank on the Eligibility List.Id. at 1377. The Second Circuit held that "[w]here a test serves [such] dual functions . . ., evidence that the scores of members of a protected group were clustered at the low end of the grading scale — though such group members may have passed the examination in sufficient numbers — provides support for a finding that the test had a disparate impact on that group, assuming the clustering could not have occurred by chance." Id. Thus, the circuit court held that it was error for the district court to reject the argument that "the rate at which black candidates achieved the minimum score on the written test necessary for promotion was significantly less than the rate at which white candidates obtained the minimum score." Id.
The 8206 Exam served the same functions as the test at issue in Waisome: It was used first to eliminate applicants scoring below 70 from further consideration, and then to rank those scoring above 70 on an eligibility list. While the Court agrees with Dr. Sharf that an effective passing score will depend on such variables as the size of the applicant pool and the length of time the eligibility list remains open, a fair test should yield roughly proportional pass rates regardless of the score. Dr. Siskin's effective-score analysis, by contrast, reveals disproportionate clustering: Of the 24 qualified Hispanic test-passers, 22 (or 91.7%) received scores of 70-84, while only 2 (or 8.3%) received scores of 85 or above; of the 233 qualified white test-passers, 148 (or 61.4%) received scores of 70-84, while 90 (or 38.6%) received scores of 85 or above.
Regarding the Brennan Intervenors' second challenge, the Court concludes that the samples (42 Hispanic test-takers and 319 white test-takers) were sufficiently large to make meaningful comparisons. Compare Connecticut v. Teal, 457 U.S. 440, 443 n. 4 (1982) (comparison of 48 black test-takers and 259 white test-takers sufficient), with Pollis v. New School, 132 F.3d 115, 121-122 (2d Cir. 1997) (sample size of eight insufficient).
Finally, regarding the claim that the effective-score analysis was an eleventh-hour invention, the Brennan Intervenors do not dispute that Dr. Siskin included the analysis in his 1997 report; since Drs. Bobko and Sharf did not address that analysis in their critiques of Dr. Siskin's report, his failure to include it in his rebuttals is of no consequence.
Thus, notwithstanding the Brennan Intervenors' challenges to Dr. Siskin's methodology, his effective-score analysis provides the requisite strong basis in evidence that the 8206 Exam had a disparate impact on Hispanics, unless, as the Brennan Intervenors contend, data errors undermine the results of the analysis.
b. Data Errors
The Brennan Intervenors challenge the accuracy of the data because it identified only two qualified Hispanics — Wilfredo Cruz and Kevin LaFaye — who scored 85 or higher on the exam; they argue that their cross-examination of Dr. Siskin established that Zapata and Trifun were erroneously excluded from that total. In addition, they argue in their post-hearing submissions that three more — Gil Perez, Isaac Alboher and John Fernandez — were erroneously excluded from the number of qualified Hispanics scoring 85 or higher because they appear on a hiring list for the 8206 Exam provided by the Board during discovery.
In an affidavit submitted after the November 21st hearing, Dr. Siskin avers that the information he received from the Board classified Trifun and Zapata as white. While he does not dispute that Perez, Alboher and Fernandez are Hispanic, he attests that the information he received from the Board classified them as unqualified; while Trifun was also classified as unqualified, Dr. Siskin acknowledges that he was subsequently hired. Finally, Dr. Siskin avers that all of the experts relied on the data provided by the Board. See Decl. of Bernard R. Siskin, Ph.D. (Dec. 14, 2006).
The Brennan Intervenors have moved to strike Dr. Siskin's affidavit or, in the alternative, to reopen the hearing. The Court notes that, despite having a full and fair opportunity to contest Dr. Siskin's data, the Brennan Intervenors did not adduce any evidence regarding Perez, Alboher and Fernandez at the November 21st hearing. Nevertheless, the Court will assume, arguendo, that the Brennan Intervenors are correct that Perez, Alboher and Fernandez should be reclassified as qualified, Zapata as Hispanic, and Trifun as qualified and Hispanic; those reclassifications yield the following results:
Table 3: "Effective" Passing Score (85 or Above) with Perez, Alboher, Fernandez, Zapata and Trifun Reclassified
Table 3 is based on the Court's own calculations. The numerators for the Hispanic and white pass rates were derived by adding five (Perez, Alboher, Fernandez, Trifun and Zapata) to Dr. Siskin's number of qualified Hispanic effective passers and subtracting one (Zapata) from his number of qualified white effective passers. Calculation of the denominators was more complicated because the reclassifications change the proportions of unqualified effective test-passers from 5/29 to 2/31 for Hispanics and from 40/273 to 39/271 for whites; applying the new proportions to Dr. Siskin's three assumptions yields the denominators given above.
Comparator Group Hispanic White Hispanic Pass Pass Rate Pass Rate Rate ÷ White Pass Rate
Excluding unqualified takers 17.1% 32.8% 52.13% (Assumes proportion of unqualified failers = (7/41) (89/271) proportion of unqualified passers) Excluding unqualified takers 17.5% 33.6% 52.08% (Assumes proportion of unqualified failers = (7/40) (89/265) 2x proportion of unqualified passers) Excluding unqualified takers 17.9% 34.5% 51.88% (Assumes proportion of unqualified failers = (7/39) (89/258) 3x proportion of unqualified passers) Thus, giving the Brennan Intervenors the benefit of all doubts about Perez, Alboher, Fernandez, Zapata and Trifun results in smaller disparities, but does not change the conclusion that the effective pass rate of qualified Hispanics was less than 80% that of qualified whites; there is still a dramatic difference between the effective pass rates of Hispanics and whites.Reclassifying Perez, Alboher, Fernandez, Zapata and Trifun undoubtedly changes the standard deviations presented in Table 2, and neither Dr. Siskin nor Dr. Sharf has recalculated them. Since, however, statistical significance is an alternative test for disparities that do not run afoul of the "80 percent" rule of thumb, see 29 C.F.R. § 1607.4(D) ("Smaller differences in selection rate may nevertheless constitute adverse impact, where they are significant in both statistical and practical terms[.]"); Xerox Corp., 196 F.3d at 365 (describing statistical significance as "an alternative measure of differences between groups"), recalculation of the standard deviations is unnecessary.
D. Conclusion
In sum, notwithstanding the Brennan Intervenors' challenges, Dr. Siskin's analysis demonstrates that there was a large disparity between the effective pass rates of qualified Hispanics and qualified whites taking the 8206 Exam. Accordingly, the Court finds that there was a strong basis in evidence — that is, evidence approaching a prima facie case — that the exam had a disparate impact on Hispanics.
THE CALDERO INTERVENORS' MOTIONS
A. For Reconsideration
The September 11th M O assumed that the Agreement was intended to remedy both testing and recruiting discrimination, and that the beneficiaries were, depending on whether they had taken and failed a challenged exam, accordingly divided into testing-claim beneficiaries and recruiting-claim beneficiaries. These assumptions were based on the Relief Chart prepared by the United States, but the Court recognized that the Relief Chart was not necessarily sacrosanct. See, e.g., New York City Bd. of Educ., 448 F. Supp. 2d at 446 (" If the United States' Relief Chart is proven to be accurate, 28 [blacks and Hispanics] were afforded full retroactive seniority [under the testing claim]." (emphasis added)).
The Caldero Intervenors now ask the Court to reconsider its reliance on the Relief Chart. In that regard, they note that the Agreement does not categorize the beneficiaries, and that the United States prepared the Relief Chart based on the assumption that the Agreement could only provide relief to actual victims of discrimination, an assumption rejected by the Court in the September 11th M O.
The Caldero Intervenors' request for reconsideration presents, in essence, an issue regarding the United States' and the Board's intent in drafting the Agreement and, in particular, in creating the list of Offerees. It is possible, as the Caldero Intervenors contend, that the parties developed the list without distinguishing between the testing claim and the recruiting claim. It is also possible, as the Relief Chart suggests, that the parties intended to offer relief based on the testing claims to blacks and Hispanics who had taken and failed one of the challenged exams, and relief based on the recruiting claim to blacks and Hispanics (as well as women and Asians) who had not.
Whatever the parties intended, their intent was surely based on the assumption that both the testing claim and the recruiting claim warranted an affirmative-action remedy. The Court has rejected the latter half of that assumption; the Agreement is silent, and therefore ambiguous, as to how the parties would have addressed this change in circumstances had they foreseen it. Accordingly, an evidentiary hearing is needed to determine the parties' intent as a matter of fact. See, e.g., Spencer, White Prentis, Inc. v. Pfizer, Inc., 498 F.2d 358, 363 (2d Cir. 1974) ("[T]he meaning of an ambiguous [contract] presents a question of fact on which resort may be had to extrinsic aids of construction throwing light upon the intent of the parties.").
Determination of the parties' intent has potential implications on the Court's legal analysis. If the Court finds that the intent of the Agreement was to give relief to blacks and Hispanics regardless of whether they failed a challenged exam, then the Court must reassess whether the Agreement, so construed, "unnecessarily trammels" the rights of non-minority employees under Title VII and is narrowly tailored under the Fourteenth Amendment.
The September 11th M O sets forth the factors relevant to the Court's "unnecessary trammeling" analysis:
[T]he award of retroactive seniority to non-victims insofar as it affects transfers and TCAs passes muster under Title VII. In the first place, it has a limited effect: While being passed over for a transfer or a TCA is not inconsequential, it is less harsh than losing one's job or being required to accept a reduction in pay; in any event, there is no absolute entitlement by any employee to either of these benefits.
Moreover, neither race, national origin nor gender are "absolute bars" to advancement by the non-minorities; there are no quotas or set-asides for minorities and women on the transfer and TCA lists. With respect to transfers, seniority is only one factor; with respect to TCAs, seniority only affects the order in which TCAs are awarded, not their frequency. And even when race, national origin or gender proves to be the deciding factor, the unsuccessful applicant remains eligible for future transfers and TCAs. Finally, the award is temporary — it comprises a one-time benefit to current employees; future hirees will attain and accrue seniority without regard to race, national origin or gender.New York City Bd. of Educ., 448 F. Supp. 2d at 431. While the addition of a large number of beneficiaries might eventually compel the conclusion that an affirmative-action plan unnecessarily trammels the rights of others, the relatively modest addition of 11 beneficiaries here does not change the fact that employees have no absolute entitlement to a particular transfer or TCA, that there are no quotas or other absolute bars to obtaining these benefits, or that the Agreement provides for a one-time adjustment of seniority. Thus, the Court concludes that construing the Agreement to provide benefits to 11 blacks and Hispanics who did not fail a challenged exam will not alter the Court's conclusion that, for purposes of transfers and TCAs, it comports with Title VII.
As explained in the September 11th M O, the narrow tailoring inquiry requires consideration of "the necessity for the relief and the efficacy of alternative remedies; the flexibility and duration of the relief, including the availability of waiver provisions; the relationship of the numerical goals to the relevant labor market; and the impact of the relief on the rights of third parties." New York City Bd. of Educ., 448 F. Supp. 2d at 438 (quoting United States v. Paradise, 480 U.S. 149, 171 (1987)). Some of those factors may be affected by the number of beneficiaries. For example, awarding benefits to individuals who did not fail a challenged exam may prove to be unnecessary if providing benefits only to those who did fail an exam would be sufficient to remedy the disparate impact of the exams on the Board's workforce. If reassessment of the narrowing tailoring analysis proves necessary, the Court will invite submissions as to whether offering relief to black and Hispanic males who did not fail a challenged exam was necessary to remedy testing discrimination.
If the Court finds that the intent of the Agreement was to distinguish between testing-claim and recruiting-claim beneficiaries, the Caldero Intervenors ask that the Court hold a hearing to determine whether any of the black or Hispanic recruiting-claim beneficiaries were actual victims of discrimination. In this regard, they proffer evidence that certain beneficiaries were not aware of the exams, and would have taken one of them had they been informed.
Although the September 11th M O did not "rule out the possibility that some beneficiaries were actual victims of recruiting discrimination," the Court expressed its doubt that such a claim could be proven "given the lack of evidence of a causal connection between the number of women and minority test-takers and the Board's recruiting practices." New York City Bd. of Educ., 448 F. Supp. 2d at 446 n. 57. The Caldero Intervenors' proffered evidence confirms this doubt: That some beneficiaries did not learn of an exam that they otherwise would have taken does not make out a prima facie disparate-impact claim in the absence of evidence connecting the resulting disparity to the Board's recruiting practices, a causal link not established by the statistical evidence presented in support of the recruiting claim.
While individual beneficiaries may be able to offer evidence in addition to that relied on by the United States and the Board in developing the Agreement, and thereby establish a disparate-impact (or even disparate-treatment) claim entitling them to a court-ordered remedy in separate litigation, the issue in this case is whether the Agreement, as a voluntary affirmative-action plan, was supported by sufficient evidence of discriminatory recruiting practices. Thus, if the Court finds that the Agreement intended to distinguish between testing-claim beneficiaries and recruiting-claim beneficiaries, it will not hold a hearing on whether any of the latter were actual victims of recruiting discrimination.
B. For Clarification
The September 11th M O focused on the validity of the Agreement's awards of retroactive seniority. The Caldero Intervenors ask the Court to clarify that the M O did not declare unlawful the Agreement's award of permanent-employee status to those Offerees who were still provisional employees at the time of its execution. Having previously conceded that the beneficiaries should not lose their jobs, see New York City Bd. of Educ., 448 F. Supp. 2d at 412, the Brennan Intervenors do not oppose clarification; however, they argue that the seniority benefits that flow from the permanent appointments is conceptually identical to retroactive seniority and, therefore, that the Court's holdings with respect to the latter should apply with equal force to the former. The Caldero Intevenors and the United States respond that the awards of appointment-date seniority did not adversely affect the Brennan Intervenors because they were all already permanent employees when the Agreement took effect.
In addition, the United States, relying on the Brennan Intervenors' statements at the fairness hearing and before this Court, argues that they have waived any challenge to appointment-date seniority. A review of those statements demonstrates that the Brennan Intervenors preserved their objections to all seniority benefits. See Tr. of Fairness Hearing at 21 ("[MAGISTRATE JUDGE LEVY:] And if it were changed not to affect the seniority rights of people in your client's position, would that take care of their objections? [COUNSEL FOR THE BRENNAN INTERVENORS:] Absolutely."); Tr. of Oct. 6, 2005 Oral Argument at 13 ("MR. ROSMAN: The retroactive seniority agreements were not the only provisions that affected our seniority, our relative seniority.").
As explained in the September 11th M O, seniority plays differing roles for purposes of transfers, TCAs and layoffs. With respect to layoffs, permanent Custodial Engineers retain the seniority they accrued as permanent Custodians. See New York City Bd. of Educ., 448 F. Supp. 2d at 412-13 (citing N.Y. Civ. Serv. L. § 80(1)). With respect to transfers and TCAs, permanent employees are eligible to bid for those benefits once they complete a one-year probationary period; the Board maintains separate lists for Custodians and Custodial Engineers. See id. at 411 n. 20 (transfers), 411-12 (TCAs). Thus, unlike layoff protection, Custodians and Custodial Engineers compete separately for transfers and TCAs.
Seniority also carries certain non-competitive benefits such as pension rights. The Brennan Intervenors do not challenge the Agreement's effect on entitlement to those benefits.
Under this analysis, at least one of the Brennan Intervenors — Dennis Mortenson — was adversely affected by the Agreement's awards of appointment-date seniority to provisional employees. Though appointed as a permanent Custodian in 1990, Mortenson was not appointed as a permanent Custodial Engineer until December 20, 2002, after the awards of permanent employment took effect in February 2000, and after the beneficiaries completed their probationary periods one year later. Thus, Mortenson has less seniority, relative to the beneficiaries who received permanent appointments as Custodial Engineers, than he would otherwise have but for the Agreement.
The Court agrees with the Brennan Intervenors that its prior holdings apply to both retroactive seniority and appointment-date seniority. Since the September 11th M O held that there was insufficient evidence of past recruiting discrimination to satisfy strict scrutiny, it follows that the recruiting claim cannot be the basis for race-based alterations of relative seniority for purposes of transfers, TCAs or layoffs.
THE BRENNAN INTERVENORS' MOTION
Having concluded that the evidence of the 8206 Exam's disparate impact on Hispanics was sufficient to justify affirmative action, the Court need not address the Brennan Intervenors' contention that other Hispanic Offerees in addition to Luis Torres received relief based on that exam. The Court therefore turns to their request for reconsideration regarding class certification.
In the September 11th M O, the Court certified a class "comprising all custodial employees whose layoff-protection rights have been adversely affected by the grant of seniority benefits to beneficiaries who are non-victims of discrimination." The class was so defined to encompass those custodial employees whose Title VII and Fourteenth Amendment rights were violated by the Agreement.
As the Brennan Intervenors point out, however, the class, as currently comprised, does not include custodial employees whose seniority rights for transfers, TCAs and layoff protection might have been displaced based on the recruiting claim. In addition, the Brennan Intervenors contend that the class should include all custodial employees whose seniority rights for transfers, TCAs and layoff protection were adversely affected by the Agreement; in that regard, they point out (1) that the issue of class certification is antecedent to the merits of the claims the putative class seeks to pursue, and (2) that appellate review may result in different conclusions regarding the Agreement's validity.
The Brennan Intervenors' points are well-taken. In addition, the Court notes that because the class has been certified under Federal Rule of Civil Procedure 23(b)(2), broadening the class has the practical advantage of binding class members to the extent the Court has upheld the Agreement's validity. See, e.g., Daniels v. City of New York, 198 F.R.D. 409, 415 (S.D.N.Y. 2001) ("When a class action is certified under Rule 23(b)(2) . . ., all persons comprising the class become mandatory members. In other words, all those who come within the description in the certification become, and must remain, members of the class because no opt-out provision exists."). Indeed, a class action under Rule 23(b)(2) has been described as "a uniquely appropriate procedure in civil-rights cases, which generally involve an allegation of discrimination against a group as well as the violation of rights of particular individuals." 7AA Charles Alan Wright, Arthur R. Miller Mary Kay Kane, Federal Practice and Procedure § 1776 (3d ed. 2005); see also id. § 1776.1 ("Rule 23(b)(2) has been utilized to protect a variety of constitutional rights.").
For these reasons, the class to be certified is redefined as all custodial employees whose seniority for purposes of transfers, TCAs and layoff protection has been adversely affected by the grant of seniority benefits to the Offerees.
CONCLUSIONS
1. The Court finds that the United States established a strong basis in evidence that the 8206 Exam had a disparate impact on Hispanics; therefore, to the extent the Agreement made awards based on the exam, except for purposes of layoffs, it comports with the Fourteenth Amendment and, a fortiori, Title VII.
2. The Court clarifies that the holdings of the September 11th M O apply to both retroactive seniority and seniority based on permanent appointment.
3. For purposes of injunctive and declaratory relief, The Court certifies a class comprising all custodial employees whose seniority for purposes of transfers, TCAs and layoff protection has been adversely affected by the grant of seniority benefits to the Offerees.
4. An evidentiary hearing is required to determine (a) whether the Board and the United States intended to categorize Offerees as testing-claim and recruiting-claim beneficiaries and, if so, (b) which Offerees fall into each category.