Pennsylvania v. Rizzo

466 F. Supp. 1219, 20 Fair Empl. Prac. Cas. (BNA) 130, 1979 U.S. Dist. LEXIS 14428, 20 Empl. Prac. Dec. (CCH) 30,241

Opinion

MEMORANDUM OPINION

BECHTLE, District Judge.

Presently before the Court are the motions of the plaintiffs for an injunction pendente lite and a permanent injunction. After careful review and consideration of the testimony and exhibits presented at trial, the pretrial and post-trial briefs and arguments of counsel, and the proposed findings of fact and conclusions of law submitted by the parties, the Court makes the following narrative findings of fact and conclusions of law. The bracketed references to the record set forth the primary sources from which statements contained herein were drawn.

These motions are the most recent development in an extensive course of litigation involving alleged discrimination against blacks by the Philadelphia Fire Department. In January, 1974, plaintiffs filed a class action complaint against the City of Philadelphia and various officials in the City government alleging discrimination against minorities in the initial hiring of fire fighters and in the promotion of minorities at the higher ranks, in violation of Title VII of the Civil Rights Act of 1964, as amended, 42 U.S.C. § 2000e et seq. (“Title VII”). This Court ordered extensive relief in 1975 after finding discrimination at the entry level and in the promotion procedures at the ranks of Lieutenant and Captain and ordered that certain promotions be made and that new examinations be developed at those ranks. The Court noted at that time that it retained jurisdiction over all aspects of the controversy. Subsequent to the relief ordered in 1975, but not pursuant to any Order of the Court, ranking officials of the Fire Department and City Personnel Department concluded that it would be prudent, in light of the litigation stance of the City, to develop new promotional examinations for the ranks of Fire Battalion Chief (“Battalion Chief”), Fire Deputy Chief (“Deputy Chief”) and Fire Assistant Chief (“Assistant Chief”), the top uniformed ranks in the Philadelphia Fire Department (collectively, “the Chief ranks”). These examinations (the “1977 Fire Command series”) were developed over the course of a year and were administered on July 16, 1977. It is these examinations that the plaintiffs now allege are violative of Title VII [N.T. 3] and it is these examinations alone, viewed in the context in which they were given, which the Court will now examine [N.T. 4]. A hearing was held on the plaintiffs’ motions and the matter was taken under advisement. The motions were consolidated pursuant to Fed.R.Civ.P. 65(b) and the controversy is now ripe for this Court’s final adjudication.

FINDINGS OF FACT

I. Parties

The plaintiffs are the Commonwealth of Pennsylvania; Club Valiants, Inc., a nonprofit Pennsylvania corporation which is alleged to include in its membership 80% of the minority members of the Philadelphia Fire Department, though it is not limited to minority members; and, 20 individual minority members of the Fire Department, 3 of whom are candidates who unsuccessfully sat for the 1977 Fire Command series written knowledge examination. On January 5, 1975, the plaintiffs’ motion for certification as a class action pursuant to Fed.R.Civ.P. 23 was granted and the plaintiff class was defined to include, inter alia, all persons “who are presently or who will be, in the future, uniform members of the Philadelphia Fire Department.” [January 5, 1975, Order.] (The above-referenced parties will hereinafter be referred to collectively as “plaintiffs.”)

The defendants are the City of Philadelphia; Fire Commissioner Joseph R. Rizzo; Mayor Frank L. Rizzo; Hillel Levinson, Managing Director; Foster R. Roser, Personnel Director; Lewis Taylor, Personnel Director; Clarence M. Farmer, Chairman, Philadelphia Commission on Human Relations; and the Commissioners of the Philadelphia Civil Service Commission. (The above-enumerated defendants will hereinafter be referred to collectively as “the City.”)

II. Background

Plaintiffs’ complaint, filed on January 31, 1974, alleged discrimination by the City in the initial hiring of candidates for the position of fire fighter and in the promotional procedures in the supervisory ranks. Preliminary relief was granted, Commonwealth v. Rizzo, EPD ¶ 9681 (E.D.Pa.1974), and a final Order was entered on January 5, 1975. In that Order, the Court found that the City had not engaged in intentional discrimination but had, in fact, made reasonable efforts to eradicate the vestiges of discrimination and that the current administration fully intended to live up to its obligations under Title VII. [Findings of Fact and Conclusions of Law, 1975 Order, pp. 2, 4.] Nonetheless, due to the disparate impact of the hiring and promotional selection devices on minorities, the Court found that the plaintiffs had proven a prima facie case with regard to the initial hiring and promotion of minorities to the ranks of Lieutenant and Captain. The Court found that it was not possible to examine the statistical impact of the promotional procedures at the higher Chief ranks, however, because there were only six minority members at those ranks. [Findings of Fact and Conclusions of Law, supra, p. 9.] By way of relief, the City was ordered to hire and promote a certain percentage of minority candidates at each level at which discriminatory impact was found, and the City was also directed to promulgate new examinations at those same ranks. Even though no finding of discrimination was made as to the Chief ranks, pursuant to a stipulation of the parties, the City was directed to promote several minority candidates to the Deputy Chief and Battalion Chief ranks when the next promotions were made. However, the City was not directed to promulgate new examinations for promotions at the Chief ranks. Nonetheless, in view of the City’s litigation stance and the continuing jurisdiction of this Court, Commissioner Rizzo, in conjunction with other officials in the Fire Department, as well as the City Personnel Department [N.T. 481], determined that it would be prudent to promulgate new promotion examinations at the Chief ranks.

The planned format for the examinations consisted of a written knowledge examination for each rank, which candidates were able to sit for on the basis of certain seniority and time-in-rank criteria. On the basis of the score on the written examination, the Fire Department, in conjunction with the Personnel Department, would select a cutoff score (based on factors now in dispute here) and those candidates who scored at or above the cut-off point would be permitted to take an oral examination before three ranking fire fighters from the Philadelphia Fire Department, joined by ranking mem bers of fire departments in other cities. The traditional method used by the Fire Department in setting the cut-off score of promotional examinations had been to examine the predicted needs of the Fire Department, based upon estimations of attrition and retirement and draw the line at that point. Once again, in view of the litigation stance of the City, it was determined by Commissioner Rizzo and other officials that an effort should be made to increase minority representation in the oral examination process. Consequently, the cut-off scores for the written examinations were lowered and the numbers permitted to go on to the oral examination stage were increased, in the hope that the larger number of candidates would make it more likely that minority candidates would be represented. The written examination score, the oral examination score and a seniority rating were each given a specific weight in advance with regard to each rank, and a final eligibility list was developed based upon these factors. For each rank, promotions were to be offered to the highest ranking candidate first, and the eligibility list was to be in effect for two years. The developmental process for the written knowledge examination began in July of 1976 and involved roughly ten separate and distinct stages. [N.T. 341.] First, a comprehensive job analysis was done for each of the three ranks which was then reviewed by incumbents for verification, and the incumbents were invited to offer suggestions. These suggestions were considered by the Personnel Department, after which the finalized task list was reviewed once more by a large number of incumbents. The final task list was given to incumbents who rated each task in frequency of performance, importance and criticality. The Personnel Department then analyzed the data and assigned each task a relative importance weight. Each incumbent was then asked to generate the “Skills, Knowledges, Abilities, and Personal Characteristics” (“SKAPs”) which were required to perform these tasks. Then each SKAP was assigned a relative weight. This process formed the basis for the development of each written examination. On the basis of these relative weights, test items (“questions”) were then developed by the City Personnel Department under the direction of David Wagner of Management Scientists, Inc. [N.T. 353.] For the 1977 Fire Command series, written knowledge examinations were administered on July 16,1977, and the examinations were taken by the following number of candidates:

Assistant Chief Deputy Chief Battalion Chief

White candidates 25 45 99

Black candidates _2 _1 6

Total candidates 27 46 105

Upon completion of the grading process, but before any identities were known, the cumulative rank order scores were tabulated. [See Appendix A.] On August 30, 1977, a special meeting was convened at the office of Commissioner Rizzo for the purpose of establishing the cut-off score. Rather than following the historical practice of establishing the cut-off score only on the basis of predicted needs over the life of the eligibility list, it was decided that additional factors should be considered. Specifically, it was decided that it would be important to try to increase minority representation at the oral examination. [N.T. 498. et seq.] Consequently, at this meeting, it was established that the anticipated needs over the two-year life of each eligibility list would be as follows:

Assistant Chief Deputy Chief Battalion Chief

Anticipated needs (over the two-year life of the eligibility list) 3 8 26

Number of candidates yielded by a cut-off score based solely on departmental need 4 13 26

At the meeting, the parties agreed that the likelihood of minority representation would be enhanced if the cut-off points were lowered so that more examinees could be included on the ultimate list, but several factors were recognized as being detrimental to this approach. They were: (1) the difficulty of giving oral examinations to large numbers of people; (2) loss of morale due to unfulfilled expectations if members achieved a position on the eligibility list but had little chance of promotion; (3) the costs due to lost time spent in the oral examinations; and, (4) the difficulty in finding qualified ranking Chiefs from the Philadelphia Fire Department, as well as from fire departments in other cities, to give the oral examinations. Despite these drawbacks, and before any identities were known, the cut-off points were lowered to allow the following numbers of candidates to be eligible for inclusion on the list:

Assistant Chief Deputy Chief Battalion Chief

Anticipated needs 3 8 26

Number of candidates at or above the lower cut-off score 8 - 15 40

Several days after the establishment of the lower cut-off scores yielding a greater number of examinees, the racial identities of those appearing at or above the cut-off level were disclosed. None of the minority candidates had scored at or above the cutoff scores on any of the three examinations. Moreover, on both the Assistant Chief and Battalion Chief examinations, a black candidate scored one point below the cut-off scores. [See Appendix A.] The results of the three examinations tabulated on the lower passing score are as follows:

# Black Candidates Black Candidates Passing # White Candidates White Candidates Passing

Assistant Chief 2 0 25 8

Deputy Chief 1 0 45 15

Battalion Chief 6 0 99 40

The Court believes and, accordingly, finds that there was no knowledge of the racial or personal identities of any candidate at the time the cut-off scores were established. Indeed, there is no allegation made of such knowledge in the plaintiffs’ motions. Nonetheless, these results form the basis of the plaintiffs’ motions for an injunction pendente lite, permanent injunctive relief and the following relief asked for by the plaintiffs in the following terms:

1. That the City be directed to develop new examinations for the Chief ranks.

2. That the cut-off points be lowered.

3. That oral examinations be given to all candidates.

4. That no new promotions be made pending the outcome of this litigation.

5. That all minority candidates be promoted.

6. That an award of back pay be made.

7. That costs and attorneys’ fees be awarded.

III. The Hearing

A hearing on the plaintiffs’ motions was held beginning on October 12, 1978, and evidence was presented by the plaintiffs and the City for four days. The plaintiffs first presented Dr. Felix Lopez, Ph.D., who was qualified as an expert in test validation and construction and who testified with regard to two issues. First, Dr. Lopez testified, with some understandable equivocation, that from the point of view of test content and structure there was sufficient similarity in the three Fire Command series examinations to aggregate the results for the purpose of performing a statistical analysis to determine adverse impact. Secondly, he stated that in his professional opinion the test had not been validated because, inter alia, the validation strategy chosen by the City was not sufficiently rigorous, nor was the establishment of the cut-off point valid from a psychometric point of view. Dr. Jagbir Singh, Ph.D., qualified as an expert statistician, testified for the plaintiffs as to the statistical significance of the adverse impact on black candidates. He testified that a candidate who scored one point below the cut-off point would have a statistically significant chance of moving up in rank if he were permitted to take the oral examination. Richard B. McCord, Chief of Examination Administration for the City, testified as to the procedure used by the City in setting the cut-off score. Lastly, Captain Donald Brown, Captain Robert Dobson, Fire Deputy Chief Donald Patton and Fire Deputy Chief Ronald Lewis, all unsuccessful minority candidates who had taken the written examination, testified as to taking the test and identified their generally superior performance ratings over the last several years. At the conclusion of the plaintiffs’ evidence, the City moved for dismissal pursuant to Fed.R. Civ.P. 41(b) on the ground that the plaintiffs had failed to prove a prima facie case. [N.T. 276.] The motion was denied. [N.T. 283.]

The City then presented Allen L. Sockoloff, Ph.D., qualified as an expert in psychology, testing and statistics, who testified that there was no statistically significant difference in the mean scores of blacks and whites; that the establishment of the cutoff score did not have a statistically significant adverse impact on blacks; and, that the tests in the 1977 Fire Command series were not sufficiently similar to aggregate the raw scores for the purpose of analyzing adverse impact. William J. McNulty, a personnel officer in the Fire Department, testified as to the establishment of the cut-off score. Next, David J. Wagner, an expert in test construction, testified that the Fire Command series examinations were content valid, that the establishment of the cut-off scores was reasonable by professional standards and that the tests scores could not be aggregated to determine the statistical significance of the impact on blacks. Lastly, Commissioner Rizzo testified generally with regard to the promulgation of the examinations and the establishment of the cut-off score. No rebuttal was presented by the plaintiffs.

Counsel for all parties agreed that there would be no further evidence submitted, with the exception of certain responsive reports prepared by their respective experts which concerned the validation of the 1977 Fire Command series examinations. The Court then ordered the consolidation of the hearing on the merits with the application for the injunction pendente lite, pursuant to Fed.R.Civ.P. 65(a)(2), and the matter was taken under advisement.

IV. Applicable Legal Principles

In Teamsters v. United States, 431 U.S. 324, 97 S.Ct. 1843, 52 L.Ed.2d 396 (1977), the Supreme Court recognized that there are basically two types of discrimination in Title VII litigation. The first consists of allegations of “disparate treatment,” where the employer treats some people less favorably than others because of their race. Id., at 333, n.15, 97 S.Ct. 1843, 1854 n.15. The second type of claim is known as that having a “disparate impact,” and involves “employment practices that are facially neutral in their treatment of different groups but that in fact fall more harshly on one group than another and cannot be justified by business necessity.” Id. The plaintiffs have acknowledged that the issues at hand point to the “disparate impact” type. Simply stated, plaintiffs contend that, although the 1977 Fire Command series written examinations were facially neutral, the examinations in fact had a grossly disproportionate impact on minorities and that, in the context of historical discrimination by the Fire Department and the City, they have proven a prima facie case. In Dothard v. Rawlinson, 433 U.S. 321, 97 S.Ct. 2720, 53 L.Ed.2d 786 (1977), the Supreme Court outlined the respective burdens of proof in disparate impact cases. The Court noted that “Congress required ‘the removal of artificial, arbitrary, and unnecessary barriers to employment when barriers operate invidiously to discriminate on the basis of [race] . . . .” Id., at 328, 97 S.Ct. at 2726. After examining prior decisions, the Court stated:

Those cases make clear that to establish a prima facie case of discrimination, a plaintiff need only show that the facially neutral standards in question select applicants for hire in a significantly discriminatory pattern. Once it is thus shown that the employment standards are discriminatory in effect, the employer must meet “the burden of showing that any given requirement [has] ... a manifest relationship to the employment in question.” Griggs v. Duke Power Co. (401 U.S. 424, 91 S.Ct. 849, 28 L.Ed.2d 158). If the employer proves that the challenged requirements are job-related, the plaintiff may then show that other selection devices without a similar discriminatory effect would also “serve the employer’s legitimate interest in ‘efficient and trustworthy workmanship.’ ” Albemarle Paper Co. v. Moody (422 U.S. 405, 95 S.Ct. 2362, 45 L.Ed.2d 280), quoting McDonnell-Douglas Corp. v. Green (411 U.S. 792, 93 S.Ct. 1817, 36 L.Ed.2d 668).

Id., at 329, 97 S.Ct. at 2726 (citations omitted). This manner of proving a ease which permits a plaintiff to prove a prima facie case upon the showing of a grossly disparate impact on minorities by use of an employment criterion without proving an intent to discriminate, Griggs v. Duke Power Co., 401 U.S. 424, 91 S.Ct. 849, 28 L.Ed.2d 158 (1970), and then requires the employer to validate that criterion, Dothard v. Rawlinson, supra, is not a rule of evidence or procedure, but rather is construed by the Court to be an analytical framework within which diverse factual patterns can be examined.

V. The Grierers Standard, As Applied to Municipal Employers

The City concedes that a plaintiff need not allege and prove discriminatory intent in disparate impact cases brought against private employers, but argues that the same standard may not be applied in a case alleging discrimination against a municipal employer. Specifically, the City argues that, in light of the decisions in Washington v. Davis, 426 U.S. 229, 96 S.Ct. 2040, 48 L.Ed.2d 597 (1976), and Village of Arlington Heights v. Metropolitan Housing Development Corp., 429 U.S. 252, 97 S.Ct. 555, 50 L.Ed.2d 450 (1977), Congress did not have the constitutional authority to mandate the application of the Griggs standard to municipal employers in Title VII litigation, pursuant to its legislative authority under the Enforcement Clause of the Fourteenth Amendment, U.S.Const. Amend. XIV, Section 5 (“Section 5”). Secondly, the City argues that the use of the Griggs standard in litigation against municipal employers cannot be justified as an exercise of legislative authority pursuant to the Commerce Clause, in light of the recent decision of the Supreme Court in National League of Cities v. Usery, 426 U.S. 833, 96 S.Ct. 2465, 49 L.Ed.2d 245 (1976). The plaintiffs respond by arguing that the application of the Griggs standard to municipalities is a valid exercise of legislative authority pursuant to Section 5 of the Fourteenth Amendment. We agree with the plaintiffs’ position that the Griggs standard is applicable to municipalities and hold that they need not prove discriminatory intent and that a showing of grossly disparate impact on minorities is sufficient to prove a prima facie case of discrimination pursuant to Title VII.

As originally enacted, Title VII excluded state and local governments from its definition of “employer.” Pub.L. No. 88-852, 42 U.S.C. § 2000e(h) (1970). However, in light of evidence of widespread discrimination by state and local employers, see United States Commission on Civil Rights, A Report on Equal Opportunity in State and Local Government (1969), Congress passed the Equal Employment Opportunity Act of 1972, Pub.L. No. 92-261, 86 Stat. 103 (1972), codified in various sections of Title VII (“1972 amendment”), which brought municipal and state employers within the definition of “employer.” 42 U.S.C. § 2000e(h) (Supp. Ill, 1973). It is clear that Congress was aware of the Griggs standard at the time of the amendment, S.Rep.No.92-415, 92nd Cong., 1st Sess., pp. 10, 14, and we must conclude, therefore, that Congress approved of it. Furthermore, it is equally clear that Congress intended to extend to government employees “the same benefits and protections in equal employment as the employees in the private sectors of the economy.” S.Rep.No.92-415, 92nd Cong., 1st Sess., p. 9. We must, therefore, conclude that Congress intended to mandate the application of the Griggs standard to employment practices in disparate impact cases brought against municipal employers. We must next examine whether Congress had constitutional authority to mandate such application.

While Title VII as a whole is based on various constitutional premises of legislative authority, Congress intended the 1972 amendments to be primarily exercises of power under the Commerce Clause and under Section 5. S.Rep.No.92-415, 92nd Cong., 1st Sess., pp. 10-11. Moreover, in Fitzpatrick v. Bitzer, 427 U.S. 445, 96 S.Ct. 2666, 49 L.Ed.2d 614 (1976), the Supreme Court stressed that the 1972 amendment was primarily based on Section 5. Id., at 453, n.9, 96 S.Ct. 2666.

It is the position of the City that Washington v. Davis, supra, which held that intent to discriminate must be proven in a discrimination action based on constitutional equal protection, bars Congress from legislatively prohibiting, pursuant to Section 5, any practices which are not prohibited by the strictures of the first four sections of the Fourteenth Amendment. In support of its position, the City cites several lower court decisions which have held that intent must be proven against a municipal employer. For example, in Friend v. Leidinger, 446 F.Supp. 361 (E.D.Va.1977), the court reasoned that Congress’ power to enforce the Fourteenth Amendment under Section could be no broader than the constitutional provision which Congress was seeking to vindicate and that, in light of Washington v. Davis, supra, intent must be proven. Similarly, the court in Scott v. City of Anniston, 430 F.Supp. 508 (N.D.Ala.1977), stated that “it is simple logic that a statute can be no broader than its Constitutional base.” Id., at 515. We decline to follow these decisions because we believe they misconstrue the grant of legislative authority inherent in Section 5.

The relationship between Section 5 and the substantive provisions of the Fourteenth Amendment was examined in Katzenbach v. Morgan, 384 U.S. 641, 86 S.Ct. 1717, 16 L.Ed.2d 828 (1966), where the Supreme Court concluded that Congress was empowered to legislatively prohibit a voting test pursuant to Section 5 which was not violative of the equal protection clause of the Fourteenth Amendment. In Lassiter v. Northampton Elections Board, 360 U.S. 45, 79 S.Ct. 985, 3 L.Ed.2d 1072 (1959), the Supreme Court had upheld a literacy test challenged on the ground that it violated the Fourteenth Amendment. Subsequently, Congress enacted the Voting Rights Act of 1965, 42 U.S.C. §§ 1973, et seq., Section 4 of which prohibited tests such as the one upheld in Lassiter. Thus, the Court in Katzenbach was squarely faced with the issue of whether Congress may legislatively prohibit a voting device pursuant to Section 5 when the device was not violative of the substantive strictures of the Amendment. The Court began its analysis by noting that by “including § 5 the draftsmen sought to grant to Congress by a specific provision applicable to the Fourteenth Amendment, the same broad powers expressed in the Necessary and Proper Clause, Art. I, § 8, cl. 18.” Id., 384 U.S. at 650, 86 S.Ct. at 1723. The Court pointed out that the test that is to be used in examining an exercise of Section 5 power is the test formulated in McCulloch v. Maryland, 4 Wheat. 316, 4 L.Ed. 579 (1819), with regard to the Necessary and Proper Clause. Id., 384 U.S. at 651, 86 S.Ct. 1731. That test is phrased in terms of whether the legislative action in question constitutes “appropriate legislation” under Section 5, keeping in mind that Section 5 “is a positive grant of legislative power authorizing Congress to exercise its discretion in determining whether and what legislation is needed to secure the guarantees of the Fourteenth Amendment.” Katzenbach v. Morgan, supra, 384 U.S. at 651, 86 S.Ct. at 1723, 1724. The Court stressed that it was for Congress to “assess and weigh the various conflicting considerations,” Id., at 653, 86 S.Ct. at 1725, and that the Court could only determine if there was a basis for Congress to resolve the conflict as it did. Id. Applying these standards, the Court found that the Voting Rights Act was “appropriate legislation” and was “plainly adapted” to further the aims of the Equal Protection Clause and, thus, was a valid exercise of Section 5 authority.

Applying these standards to the 1972 amendment to Title VII, we hold that it is “appropriate legislation” and that it is “plainly adapted” to the end of enforcing the Equal Protection Clause. There can be no doubt that the amendment was the result of congressional weighing and evaluation of the problem of discrimination by municipal employers, see S.Rep.No.92 — 415, 92nd Cong., 1st Sess., pp. 10-11 (1971); 118 Cong.Rec. 1816 (1972) (remarks of Senator Williams, floor manager of the bill); H.Rep. No.92-238, 92nd Cong., 1st Sess., p. 19, et seq. (1971); U.S.Code Cong. & Admin.News 1972, p. 2137. Further, it is clear that the application of the Griggs standard is “plainly adapted” to the end of eliminating intentional discrimination, as well as to the end of eliminating the racially disparate consequences of facially neutral employment screening devices in the municipal employment context. The City argues, however, that, even if the 1972 amendment can be premised on Section 5, National League of Cities v. Usery prohibits such an incursion into the sovereign affairs of the municipality. We do not agree. The Usery Court made it clear that it was expressing no view as to whether a different result would obtain if it were examining an exercise of Section 5 power. National League of Cities v. Usery, supra, 426 U.S., at 852, 96 S.Ct. 2465. In conclusion, therefore, we hold that the Griggs standard has been mandated by Congress in actions against municipal employers under Title VII; that Congress was validly exercising its power pursuant to Section 5 of the Fourteenth Amendment; and, that, therefore, the plaintiffs need not prove that the City intended to discriminate against minorities in the development and administration of the 1977 Fire Command series written examination. See Firefighters Institute v. City of St. Louis, 549 F.2d 506 (8th Cir. 1977); Harrington v. VandaliaButler Board of Education, 418 F.Supp. 603 (S.D.Ohio 1976).

VI. Statistical Evidence

[5,6] It is well established that statistical evidence alone, where persuasive, may be sufficient to prove that a testing device has a grossly disproportionate impact. Teamsters v. United States, supra, 431 U.S., at 339, 97 S.Ct. 1843; Robinson v. City of Dallas, 514 F.2d 1271, 1273 (5th Cir. 1975). However, the Teamsters Court made it clear that no hard and fast rules can be applied with regard to statistical evidence and that a court should not abdicate its responsibility as a finder of fact in deciding a case simply on the basis of statistics because “statistics are not irrefutable; they come in infinite variety and, like any other kind of evidence, they may be rebutted. In short, their usefulness depends on all of the surrounding facts and circumstances.” Id., 431 U.S. at 340, 97 S.Ct. at 1856-1857. Therefore, as a starting point from which to examine the impact of a testing device or an employment criterion, courts have looked at the relative success rates for each racial group or, in the alternative, have examined the rate at which an employment criterion disqualifies racial groups. See, e. g., Griggs v. Duke Power Co., supra (high school diploma requirement was met by 34% of white males, but was met by only 12% of black males). Secondly, the court must ascertain whether the number of candidates is so small that the statistical results do not reflect the “reality of the employment situation,” Dendy v. Washington Hospital Center, 431 F.Supp. 873, 876 (D.D.C.1977), or that the statistical universe is so small that the results achieved are due merely to chance or random distribution. Id. See also Perham v. Ladd, 436 F.Supp. 1101 (N.D.Ill.1977); Stewart v. General Motors Corporation, 542 F.2d 445 (7th Cir. 1976); Ochoa v. Monsanto Corp., 473 F.2d 318 (5th Cir. 1973). Thirdly, statistical evidence tending to show a disparate impact, though based on a relatively small number of employment decisions, may be augmented by evidence of significant acts of discrimination, Kaplan v. Internat’l Alliance of Theatrical, 525 F.2d 1354 (9th Cir. 1975), which must be viewed in the context of the past employment practices of the defendant. See Teamsters v. United States, supra.

VII. The Prima Facie Case

Plaintiffs seek to prove a prima facie case on the basis of the statistical results of the 1977 Fire Command series viewed in the historical and institutional context in which the examinations were given. [N.T. 3-4.] Specifically, they argue that: (1) the pass rates of blacks and whites is grossly disparate and clearly violates the Griggs standard; (2) the number of candidates can be statistically analyzed; (3) it can be demonstrated that the proportions of passing candidates was not simply due to chance; and, (4) the statistics are bolstered by evidence of a history of discrimination by the City.

In response, the City argues that: (1) although the passing rates do demonstrate a grossly disparate impact, in light of the small numbers of black candidates, the results are due to chance alone; (2) there is no statistically significant difference in the mean scores for blacks and whites on the examination; and, (3) the statistics are not sufficiently augmented by evidence of a history of discrimination or evidence of specific acts of current discrimination.

The results of the examinations are presented below in tabular form and show that, by any standard, the impact on black candidates was grossly disproportionate:

Assistant Chief Deputy Chief Battalion Chief

Passing rate — black candidates 0% 0% 0%

Passing rate — white candidates 32% 33% 40%

A comparison of the passing rates by race shows that no black candidates passed any of the examinations, while 32% to 40% of the white candidates passed. Such a disparity in passing rates far exceeds that of the employment criterion struck down in Dothard v. Rawlinson, supra. In Dothard, the Supreme Court found that a prima facie case had been proven with regard to a height requirement which disqualified 33% of the women candidates, but only disqualified 1% of the men candidates. Dothard v. Rawlinson, supra, 433 U.S., at 329, 97 S.Ct. 2720. See also Jackson v. Nassau Co. Civ. Ser. Comm., 424 F.Supp. 1162 (E.D.N.Y. 1976). Therefore, we find that the examinations had a grossly disproportionate impact on minority candidates.

A more difficult question presented is whether these figures are based upon a statistical universe sufficiently large to be probative, and whether or not these statistics reflect the “reality of the employment situation.” Dendy v. Washington Hospital Center, supra. We must find that the results were not merely due to chance or random distribution due to the small number of employment decisions and that the determination of adverse impact is statistically significant, see Commonwealth v. O’Neill, 473 F.2d 1029 (3d Cir. 1973), in order to find that a prima facie case has been proven. To this end, the plaintiffs presented the testimony of Dr. Jagbir Singh, a statistician who addressed the question of how likely it was that these results were merely due to chance. In response, the City presented the testimony of Dr. Allen Sockoloff. Dr. Singh testified that he examined the data and performed a “Fisher exact probability test” (“Fisher test”), by which he calculated the exact probability of observing the results which actually occurred under the hypothesis that blacks and whites had an equal opportunity to pass. Statistically, he examined the question, “Does a smaller proportion of blacks score at or above the cut-off point?” Dr. Singh performed the test on three sets of data. First, he examined the results of the Battalion Chief examination and concluded that the probability of observing 105 candidates, 99 of whom were white and 6 of whom were black, and having all of the 40 passing candidates be white was .0513 or 5.13%. [N.T. 169.] Second, he aggregated the results of the three examinations and calculated that the probability of having 154 candidates for three different tests— 145 white candidates and 9 black candidates — and having all of the 58 passing candidates be white was .014 or 1.4% [N.T. 170.] Lastly, because some of the candidates took more than one examination, he calculated that, out of the total of 178 examinations taken (169 by white candidates and 9 by black candidates), the probability of having only 62 whites pass was .0189 or 1.89%. [N.T. 196.] The last two aggregated calculations would be very persuasive if premised upon valid assumptions. However, the City argues that it has not been shown that the tests are similar enough from the psychometric point pf view to be aggregated for purposes of determining adverse impact. Further, the City argues that the results of the tests performed on that data received from the Battalion Chief examination was not significant at the 5% level and that, therefore, the plaintiffs have not proven a prima facie case with regard to that examination. The plaintiffs argue in response that it was valid to aggregate the scores and that the results of the Battalion Chief analysis are statistically significant.

The Court notes that Dr. Singh agreed on cross-examination that, if it were not proper from a psychometric point of view to aggregate the results of the three examinations, his statistics would not be valid. [N.T. 182.] The plaintiffs attempted to prove the propriety of aggregation through the testimony of Dr. Felix Lopez, an expert in test construction, who testified as follows:

Q. . [D]o you have an opinion as to whether for purposes of analyzing the statistical impact of the cutting score it’s appropriate as a matter of psychometrics [to] aggregate the three examinations and determine whether there is an adverse impact taken collectively?

A. ... I certainly feel that it is perfectly legitimate to do that in this case.

[N.T. 63.] On the other hand, the City’s expert on test construction, Mr. David Wagner, testified that it was not valid to aggregate the scores [N.T. 425], and this opinion was echoed by Dr. Sockoloff in his testimony. [N.T. 305.] While it would appear that we must choose between diametrically opposed positions held by equally impressive and knowledgeable experts, when Dr. Lopez was provided with an opportunity to explain his opinion, this conflict disappears. On cross-examination, he testified that it was his practice to examine the test content for the three jobs, then to look at the common elements (here, 61 out of 117, or 53% of the questions were common to all three examinations) [N.T. 149], and that he then “would combine the 61 common items . . . and look what that is doing to the total score ... I would do some kind of analysis like that.” [N.T. 154.] On cross-examination, he further explained his position:

Q. Dr. Lopez, would you combine the three distributions ... to determine disparate impact?

A. I would do it on the 61 common items.

Q. Would you do it on all the items contained in the examinations?

A. No, I don’t think I would.

[N.T. 155.] On redirect examination, when asked the same question, he said:

I think that it is appropriate to do several studies combining these [scores], you see, take the first — all three, and then taking the 61 [common items] and then you go to the next step and you take the Deputy Chief and the Battalion Chief and . you do another study and see what that is doing. .

[N.T. 157.] On recross, when asked how he would combine the data, he once more described how he would analyze the scores to determine the correlation between common items and overall test performance. [N.T. 159.] In light of this testimony, we find that Dr. Lopez intended to assert that a study would have to be done in order to justify aggregation of the data, which would correlate the scoring on the common elements with the scoring on all of the elements. Since no evidence was presented as to any such study, we, therefore, find that the data from the aggregated scores is not a valid statistic because it is based upon an invalid assumption.

We are thus left with the raw scores of the three examinations and the results of the Fisher test performed on the Battalion Chief data. We find that the sample sizes are far too small to be probative for the Assistant Chief and Deputy Chief examinations. We note that, if one black candidate had passed either examination, these examinations would have had a grossly disproportionate impact against white candidates. We are presented with a more difficult issue to decide with regard to the Battalion Chief examination. The essence of Dr. Singh’s testimony was that the probability of observing the results which occurred was .0513 or 5.13% or, in different terms, the proportion of blacks scoring at or above the cut-off point was less than that of whites and that this result was statistically significant at the .0513 level. The City argues that we must compare the statistical significance of this result with the 5% level ordinarily employed in employment discrimination litigation and that we should consider the result of a similar test performed by Dr. Sockoloff, which was clearly not significant at the 5% level.

While it is true that the 5% level of significance has been traditionally used in the testing profession and has been approved by several courts, Harless v. Duck, 14 F.E.P. 1616, 1623 (N.D.Ohio 1977), and by the EEOC guidelines for test validation, 29 CFR § 1607.5(c)(1) (1976), as well as various other contexts, we will not reject the results of plaintiffs’ study simply by mechanically lining it up with the 5% level. Rather, we think the City’s argument misconstrues the meaning and import of a “level of statistical significance.” To define a level of statistical significance is to determine how certain a trier of fact must be in making a particular determination. We will examine this borderline figure in light of other statistical and nonstatistical evidence. This other evidence consists primarily of the Fisher test result which Dr. Soekoloff arrived at and his examination of mean scores of blacks and whites.

Dr. Sockoloff statistically examined the question, “Are there differences in the proportions of black and white candidates scoring at or above the cut-off point?” [N.T. 299.] He described this as a two-tail test. He distinguished Dr. Singh’s test as a one-tail test in which Dr. Singh examined the question, “Does a smaller proportion of blacks score at or above the cut-off?” Dr. Sockoloff was unable to say that one test was the appropriate one and that the other was not, and we agree with Dr. Sockoloff that both tests are to be examined in arriving at our decision. Dr. Sockoloff further testified that he compared the mean scores of blacks and whites and concluded that there was no statistically significant difference between those means. [N.T. 297.]

In light of this evidence, we conclude that the plaintiffs have not shown by a preponderance of the evidence to a statistically significant level of certainty that the results of the Battalion Chief examination were not simply due to chance. In reaching this conclusion, we consider several factors to be of primary importance. First, at the 5% level of significance which has been approved by various judicial and administrative authorities, the plaintiffs’ test results are of borderline significance. [N.T. 170.] Second, in light of Dr. Sockoloff’s two-tail test in which he was unable to conclude that different proportions of white and black candidates score at or above the cut-off point, Dr. Singh’s borderline result is weakened further. [N.T. 299.] Third, there does not appear to be a difference in means for black and white candidates (see Appendix A), and this observation is supported by Dr. Sockoloff’s statistical analysis. Finally, and most importantly, the fact that only six black candidates took the examination is convincing evidence that, although none achieved the cut-off score, this disparity in pass rates is due to random distribution because a small number of employment decisions could have changed the situation radically. For example, if two of the six black candidates had scored at or above the cut-off point, the black pass rate would have been 33%, a figure which would make it difficult to find a disparate impact.

Our finding that the statistical universe is too small to be probative is supported by examination of the various cases addressing this issue in similar contexts. In Officers for Justice v. Civil Ser. Commission, City and County of San Francisco, 371 F.Supp. 1328 (N.D.Cal.1973), the court held that the plaintiff had not made out a statistical prima facie case involving a test given to 205 candidates, 11 of whom were minority candidates. Similarly, in Bridgeport Guardians, Inc. v. Members of Bridgeport Civil Service Commission, 354 F.Supp. 778 (D.Conn.1973) (“Bridgeport Guardians I”), the court scrutinized a test given for promotions to the rank of Sergeant, taken by 201 candidates, 10 of whom were minority candidates. The court concluded that the statistical universe was too small, particularly in light of the fact that the probability of the result which occurred was only 8%. Id., at 795. This compares with an 8% level of certainty arrived at by Dr. Sockoloff in his examination of the Battalion Chief examination. In Dendy v. Washington Hospital Center, supra, 35 candidates took an examination, 26 white and 9 black, and all whites passed while only 4 blacks passed. After examining these figures, the court noted that:

If, on the one hand, the courts were to ignore broadly based statistical data, that would be manifestly unfair to Title VII complainants. But if, on the other hand, the courts were to rely heavily on statistics drawn from narrow samples, that would inevitably upset legitimate employment practices for reasons of appearance rather than substance. The courts must be astute to safeguard both of these conflicting interests.

******

With so meager a sample, if just a handful of test results had turned out differently, the comparative percentages of black . . . and white success on the exam would have been correspondingly, and substantially, different. In the court’s estimation, “Such small numbers are insufficient to support any conclusion as to whether the rule has a discriminatory effect.” Robinson v. City of Dallas, 514 F.2d at 1273.

Id., at 876. See also Harper v. TWA, Inc., 525 F.2d 409 (8th Cir. 1975); Ochoa v. Monsanto, supra (of 684 candidates, 9 were Mexican, only 1 of whom was hired). The plaintiffs attempt to distinguish the tests examined in the above cases from the examinations before this Court by arguing that here no black candidate passed. The Court’s response is two-fold. First, the Fisher tests performed by the respective statisticians of necessity take into consideration the fact that no black passed. Second, the number of blacks was, in any event, too small to permit a determination by the Court that the results were not simply due to chance. On the other hand, courts have examined testing devices taken by relatively small numbers of minority candidates and have found the results to be probative when the statistics are bolstered by corroborating evidence of intentional discrimination and/or significantly larger number of candidates than those in the case at bar. For example, this Court held that a prima facie case had been proven with regard to the Philadelphia Fire Department’s 1972 Captain’s examination which was taken by 157 candidates, 13 of whom were black. In Bridgeport Guardians v. Bridgeport Police Dept., 431 F.Supp. 931 (D.Conn. 1977) (“Bridgeport Guardians II”), an examination was taken by 172 candidates — 15 black candidates and 5 hispanic candidates. Of 37 candidates passing, all were white. The court noted that the probability of this result occurring by chance was substantially less than 1%. Id., at 935. This case is distinguishable in that it was taken by 66 more total candidates and more than three times as many minority candidates as the tests in the case at bar. Furthermore, in Bridgeport Guardians II, the probability of the result being due to chance was less than 1%, whereas in the case at bar the probability is between 5% and 8%. In League of United Latin American Citizens v. City of Santa Ana, 410 F.Supp. 873 (C.D.Cal.1976), an examination was taken by 96 candidates, 86 of whom were white and 10 of whom were members of minority groups. 36 white candidates failed and 7 minority candidates failed. While the court conceded that the sample was small, it held that the evidence was corroborated by expert testimony as to inherent cultural bias in the examinations themselves. Id., at 902. This is distinguishable from the case at bar in that we have found no act of discrimination with regard to the test itself. The plaintiffs attempt to bolster their statistical case in three ways. First, they argue that, if the sample sizes for the 1977 Fire Command series examination are too small to be statistically significant, it is only because past discrimination by the City in the lower ranks has limited the number of blacks eligible for the examinations. Therefore, the plaintiffs argue, the City’s claim that the numbers are too small “proves too much.” We would agree that in some circumstances the discriminatory effect of a lower-level promotional examination may magnify the effect of a higher-level examination, Chance v. Board of Examiners, etc., of City of New York, 330 F.Supp. 203 (D.C.1971), affd. 458 F.2d 1167 (2d Cir. 1972), but we do not believe this argument per se may negate our finding that the data base is too small. First, such a position would require, in general, that the Court “review the validity of tests and practices which in no realistic sense present a justiciable controversy.” Bridgeport Guardians I, supra, at 796. Second, in its Order of January 5, 1975, this Court ordered extensive relief which, inter alia, made numerous minority members of the Fire Department eligible for promotional examinations at higher levels than they would have been eligible for if the previous discrimination had been unremedied. Secondly, in a related argument, the plaintiffs contend that the statistics are corroborated by evidence of a history of discrimination. We agree that the historical context must be considered and may bolster the statistical case. However, the probative value of such evidence in the ease at bar is reduced by virtue of two facts: (1) we made no finding of fact as to discrimination in the Chief ranks in any of the previous litigation, and the Court does not now find that the City was guilty of intentional discrimination at any rank and, indeed, had exercised good faith in attempting to increase minority representation; and, (2) in view of the Court’s finding that the City exercised good faith in promulgating the 1977 Fire Command series examinations and that there was no intent to discriminate in the administration or grading of the examinations, we find that the City has made significant progress in purging itself of the effects of former discrimination. Thus, while these factors are not directly determinative in the finding of a prima facie case, they do serve to minimize the probative value of the evidence of historical discrimination.

In conclusion, upon consideration of the raw data derived from the test results, the evidence submitted by the experts and the facts surrounding the tests in question, we conclude that the plaintiffs have not proven a prima facie case.

This Memorandum Opinion supports the Court’s Order dated January 23, 1979, denying plaintiffs’ motions for an injunction pendente lite and permanent injunctive relief.

APPENDIX A