Bridgeport Guardians v. Bridgeport Police Department

RULING ON PLAINTIFFS’ MOTION FOR PRELIMINARY INJUNCTION NEWMAN, District Judge. This is another lawsuit challenging the validity of a written employment examination, alleged to have a racially disparate impact. See, e. g., Washington v. Davis, 426 U.S. 229, 96 S.Ct. 2040, 48 L.Ed.2d 597 (1976); Albemarle Paper Co. v. Moody, 422 U.S. 405, 95 S.Ct. 2362, 45 L.Ed.2d 280 (1975); Jones v. New York City Human Resources Administration, 528 F.2d 696 (2d Cir. 1976), affirming 391 F.Supp. 1064 (S.D.N.Y.1975); Kirkland v. New York Dept. of Correctional Services, 520 F.2d 420 (2d Cir. 1975), affirming 374 F.Supp. 1361 (S.D.N.Y.1974); Vulcan Society v. Civil Service Commission, 490 F.2d 387 (2d Cir. 1975), affirming 360 F.Supp. 1265 (S.D.N.Y.1973) (“Vulcan”); Bridgeport Guardians, Inc. v. Bridgeport Civil Service Commission, 482 F.2d 1333 (2d Cir. 1973), affirming in part 354 F.Supp. 778 (D.Conn.1973) ("Bridgeport Guardians I”); Chance v. Board of Examiners, 458 F.2d 1167 (2d Cir. 1972), affirming 330 F.Supp. 203 (S.D.N.Y.1971). The suit challenges a promotion exam for the rank of detective in the Bridgeport Police Department. Plaintiffs are a non-stock corporation of Black police officers and an individual Black police officer who took and failed the challenged exam. Defendants are appropriate officials of the Police Department and Civil Service Commission with responsibilities for giving employment examinations and making appointments. Intervention was allowed to two groups, one composed of several patrolmen who took and passed the challenged exam and the other composed of other po *934 lice officers in the Bridgeport Police Department who assert an interest in maintaining hiring and promotion procedures based on objectively scored examinations. For purposes of considering the pending motion for a preliminary injunction to bar certification of a list of successful applicants and to prevent appointments from such list, the plaintiffs’ motion for class action certification need not be considered at this point. See Vulcan, supra, 360 F.Supp. at 1266 n. 1; Bridgeport Guardians I, supra, 354 F.Supp. at 783. The challenged examination was given on June 12, 1976. On March 2, 1977, this Court entered a temporary restraining order, barring use of the exam results as a basis for any permanent appointments to the rank of detective pending decision on the motion for a preliminary injunction, but permitting the exam results to be used for making temporary appointments on an acting basis. The suit is brought under Title VII of the Civil Rights Act of 1964, 42 U.S.C.A. § 2000e et seq. Defendants have sought dismissal for failure to exhaust statutory administrative remedies. Whether or not a district court has jurisdiction to consider a motion for preliminary injunction prior to a plaintiff’s receipt of a right-to-sue letter, compare Berg v. Richmond Unified School District, 528 F.2d 1208, 1211 (9th Cir. 1975), and Drew v. Liberty Mutual Insurance Co., 480 F.2d 69 (5th Cir. 1973), with Troy v. Shell Oil Co., 378 F.Supp. 1042 (E.D.Mich.1974), appeal dismissed as moot, 519 F.2d 403 (6th Cir. 1975), defendants’ objection has now been remedied by the issuance of a right-to-sue letter on February 16, 1977. On February 23, 1977, an amended complaint was filed to reflect this development. Defendants’ motion to dismiss is denied. The context in which this lawsuit arises has not changed since promotion procedures within the Bridgeport Police Department were challenged in 1973 in Bridgeport Guardians I, although some change in the procedures has occurred. It still remains true that there are no Black or Hispanic supervisory officers in the department, and only one Black detective. The comparison with other large Connecticut cities remains a cause for dismay. See 354 F.Supp. at 785. The procedure for promotion has been modified from what existed in 1973. See id. at 794. As a result of the earlier litigation, patrolmen are now eligible for promotion to detective after only one year in grade, a change that was made to afford the minority patrolmen appointed pursuant to Bridgeport Guardians I an early opportunity to compete for promotion. Eligible candidates take a written multiple-choice exam, scored on a scale from 0 to 100. The passing grade is 75%, still the result of a requirement of the rules of the Civil Service Commission, applicable to all written tests for any position within municipal employment. Those who pass the written exam are given an oral exam, and are also given a seniority rating based on length of service in the department up to a maximum of 20 years. A candidate’s total score is determined by weighting the various grades as follows: written exam, 70%; oral exam, 20%; seniority, 10%. Those who pass the written exam are ranked on an eligibility list in accordance with their total scores, and promotions are made strictly in order of rank on this list. ' Plaintiffs’ principal, if not only, attack is upon the written exam. Before that attack is considered, the appropriate test for a preliminary injunction in this context must be identified. Plaintiffs appear to invoke the traditional test of probable success on the merits and possible irreparable injury, and also the alternative test of sufficiently serious questions on the merits to make them a fair ground for litigation and a balance of hardships tipping decidedly towards the party seeking preliminary relief. Sonesta International Hotels Corp. v. Wellington Associates, 483 F.2d 247, 250 (2d Cir. 1973). Naturally, the plaintiffs emphasize the latter test, urging that the balance of hardships does tip significantly in their direction. It is not entirely clear that the alternative formulation is appropriate in litigation of this sort. In this Circuit the principal *935 decision upholding a preliminary injunction that barred use of a promotion exam drew from both tests in accepting the trial court’s conclusions that the balance of hardships tipped decidedly toward the plaintiffs and that there was a strong likelihood the plaintiffs would prevail at trial. Chance v. Board of Examiners, supra, 458 F.2d at 1178. In any event, the alternative formulation does not dispense with consideration of plaintiffs’ probability of success. Our Circuit has observed that where the balance of hardships tips decidedly towards the plaintiff, the burden of showing probable success is “less” than obtains when only possibly irreparable injury is shown. Dino deLaurentiis Cinematográfica, S.p.A. v. D-150, Inc., 366 F.2d 373, 375 (2d Cir. 1966). This lesser burden has been expressed in terms of raising questions so serious, substantial, and difficult as to make them a fair ground for litigation. See, e. g., Checker Motors Corp. v. Chrysler Corp., 405 F.2d 319, 323 (2d Cir. 1969); Unicon Management Corp. v. Koppers Co., 366 F.2d 199, 205 (2d Cir. 1966). But, as Checker Motors makes clear, the significance of raising such substantial questions is to demonstrate the need for “more deliberate investigation.” 405 F.2d at 323. In this case, though the plaintiffs have not agreed to combine the hearing on the preliminary injunction with trial on the merits, Fed.R.Civ.P. 65(a), the injunction hearing extensively explored plaintiffs’ claim on the merits. Testimony was given by plaintiffs’ nationally known expert, the ubiquitous Dr. Richard Barrett, and the examination was considered in detail. Moreover, while plaintiffs have indicated that additional evidence would be presented at a trial on the merits, they have not identified the evidence to be presented nor the topics to be pursued. Thus, there has not been shown a need for more deliberate investigation than has already occurred. In such circumstances, the probability of success on the merits remains an important criterion for injunctive relief, even if plaintiffs’ burden on that issue is reduced by the relative balance of hardships. As to probability of success on the Title VII claim, plaintiffs have adequately established a discriminatory impact of the challenged exam, sufficient to shift to defendants the burden of demonstrating that the exam is job-related. See, e. g., Vulcan, Bridgeport Guardians I, Chance. The exam was taken by 172 candidates, of whom 15 were Black and 5 Hispanic. Of the 37 candidates who passed the exam, all were White. The passing rate for Whites was 24%, and for minority group members, 0%. Both the disparity in passing rates and the numbers of candidates from White and minority groups, cf. Bridgeport Guardians I, 354 F.Supp. at 795, are sufficient to establish a prima facie case of discriminatory impact. The probability of this disparity occurring by chance is substantially less than 1 in 100. Defendants seek to justify use of the exam by demonstrating that it has content validity. That approach raises an initial question as to whether content validity, rather than criterion-related validity, is an acceptable indicator of job-relatedness. The EEOC guidelines suggest that content validity may be appropriate where criterion-related validity is not feasible. 29 C.F.R. § 1607.5(a). In this Circuit, no particular validation approach seems to enjoy a preference. See Vulcan, 490 F.2d at 394. Defendants contend that the initial group of detectives to be appointed, numbering only 13, is too small for an adequate predictive criterion-related study, and that the number of present detectives, 43, is less than the 50 preferred by some psychologists as an adequate number for a concurrent validation study. Moreover, the accuracy of a criterion-related study depends on the accuracy of performance evaluations, which are then sought to be correlated with test scores. With an occupation such as detective work, opinions vary as to the proper measures of performance, and most measures inevitably involve subjective judgments, which are subject to at least as much infirmity as content validity studies. *936 See Vulcan, 490 F.2d at 395 n. 10. Under all the circumstances, there is no sound reason to reject the content validity approach. Test Preparation Determination of whether an exam has adequate content validity, we are advised, should be based primarily on the adequacy of test preparation and to a lesser extent on analysis of the test itself. Vulcan, 490 F.2d at 396. The detective exam was prepared by McCann Associates, a management consulting firm working exclusively with state and local governments on public personnel matters. Preparation of tests for hiring and promotion is a major activity of the firm. The firm was asked by Bridgeport personnel officials to prepare a detective’s exam in 1974, after the decision in Bridgeport Guardians I, which had invalidated use of a patrolman’s exam prepared by a different firm. McCann was specifically asked to prepare an exam free of any racial or cultural bias. McCann’s initial step was to conduct a job analysis at Bridgeport. Interviews were conducted with 9 of the 43 detectives, and the captains of the detective and youth bureaus. Those interviewed were asked to describe the duties of a detective and to identify those aspects of job performance that distinguished the best and the worst detectives from the average. McCann personnel reviewed statistics of the detective bureau, forms used by Bridgeport detectives, and examples of their investigative reports. Based on these steps, a job description was prepared, identifying 20 categories of tasks, within each of which 10 to 20 specific tasks are detailed. 1 McCann personnel, both those who had done the work at Bridgeport and senior officials of the firm, then completed a subject matter check list, regularly used by the firm for constructing police promotion exams. This check list contains subject-matter areas in such general fields as criminal investigation, police and related techniques, and legal knowledge. Of the 52 areas, 14 were selected as being most pertinent to the duties of Bridgeport detectives, and percentages (5, 10, 15, or 20) were assigned to each of the selected areas to reflect the percent of exam questions to be devoted to each area. The appropriate percentages were determined by the job analysts sent to Bridgeport and McCann’s senior personnel, based both on their assessment of the relative significance of the tasks performed by Bridgeport detectives and their knowledge of police practices and tests throughout the country. They considered both time spent at the several tasks and their importance. Having thus determined the appropriate topics and the number of questions for each topic, the McCann personnel then selected multiple-choice questions from their bank of 4,000 police examination questions. Questions were selected that seemed appropriate to a community of Bridgeport’s size, and that spanned a range of difficulty as evidenced by results of prior use in other communities. A sufficient number of difficult questions was included to assure that the exam would be a useful device for distinguishing among the better candidates. Selection of questions also took into account Bridgeport’s requirement of a 75% passing score, a matter discussed in detail under “Passing Score,” infra. As a final step, the McCann firm applied to both the exam and the exam instructions a reading index developed by a recognized language expert to determine the degree of reading difficulty. This disclosed that the exam and the instructions were at the eighth and ninth grade levels of reading difficulty. *937 The steps outlined in preparation of the exam disclose an adequate level of skill, objectivity, and concern with achieving content validity. Plaintiff’s expert, Dr. Barrett, criticized these steps for lack of complete conformity with the procedures and criteria recommended by the American Psychological Association. While those standards are relevant to an assessment of an exam challenged in a Title VII suit, not every departure establishes a lack of job-relatedness sufficient to show a Title VII violation. The test preparation may not have been ideal. Pew things are. But it was adequate to support the “inference, rebuttable to be sure,” Vulcan, 490 F.2d at 396, that the end product achieved sufficient content validity. Plainly the preparation exceeded what has been found inadequate by district courts in prior cases. See Jones, Kirkland, and Vulcan. Exam Content As to the exam itself and its results, the McCann firm analyzed each candidate’s answers to all questions. A reliability coefficient was computed to measure the consistency with which each candidate correctly answered the questions. The test is divided in half, and a candidate’s scores on each half are compared. This is repeatedly done by computer for virtually all possible combinations whereby the exam can be divided in half. On a scale from .000 to 1.000 the reliability coefficient for all candidates was .748. Candidates’ responses to each question were also analyzed to identify questions answered incorrectly by candidates whose total score was in the first or second quartile of all candidates taking the test. Such questions were carefully reviewed for ambiguities in the wording of the question or possible answers that might have led a generally good candidate to select a wrong answer. As a result of this review, the McCann firm decided to accept as correct one additional answer to nine questions. Dr. Barrett leveled several criticisms at the exam itself. He considered the reliability coefficient below the .8 or .9 level he thought appropriate for tests of this sort. More fundamentally, he questioned whether the test truly distinguishes between those qualified to perform a detective’s work and those who are not, and whether, among those who passed, it distinguishes between those who can perform better than others. He conceded the difficulty of providing any assurance on these matters without a criterion-related study. (Tr. 78-81). Using specific questions as examples, he expressed two major criticisms. Some questions, he pointed out, measured whether a candidate knew the reason for taking a correct course of action. His criticism of such questions is that some may know the reason and not act accordingly, and others may act correctly without knowing the reason. He also objected to questions that required “abstract knowledge” not related to either the correct performance of a task or the correct reason for such action. For example., one question asks for the most accurate among several possible nationwide statistics concerning the commission of forgery. It does not appear that knowing the correct answer will be of any help in catching a forger. Dr. Barrett’s other criticisms seem relatively inconsequential. He objected to a question that used “X” to identify the criminal and “V” to identify the victim. While such symbols may tax the most rudimentary levels of comprehension, they do not seem unduly sophisticated for those aspiring to identify and apprehend criminals. He also criticized a question that asked, in the context of an arson investigation, what substance gave off an odor similar to wet match heads. His point was that a written description of the odor is a less satisfactory way of testing identification ability than exposure to the actual odor. Though mindful that the “burden of judicial examination-reading,” Vulcan, 490 F.2d at 396, need not inevitably be assumed, I considered it appropriate in this case for two reasons. First, it was important to come to some conclusion as to the frequency of questions encountering Dr. Barrett’s major criticisms. Second, the subject matter *938 seemed not too foreign from judicial competence. The questions test for knowledge of many matters that good detectives ought to know. These include pertinent facts concerning various crimes and criminal behavior; techniques for investigating, interrogating, apprehending, and testifying; and legal matters, such as the elements of offenses and the lawfulness of arrest and search procedures. In addition a few questions attempt to assess reasoning powers by requiring a logically correct inference to be made from a complicated set of facts. I found very few questions concerned with abstract knowledge. There were some instances where the question asked for the right reason to do or refrain from doing something. I share Dr. Barrett’s concern that correctly answering such questions is not an assurance of proper conduct, but with the art of predicting individual human behavior at a current level only slightly above sheer guesswork, I cannot say that testing for knowledge of why certain conduct is correct is an unlawful way of distinguishing between those who are more or less likely to take correct action. Nothing in Title VII or the EEOC guidelines permits a court to reject the test preparer’s premise that knowing the reason for correct action is some indication of acting accordingly. After considering each question, I found eight questions for which I thought an additional answer was as arguably correct as the one specified by McCann. By that I mean that either the alternative was correct in fact or, in my opinion, there was no basis for thinking that a person who selected the McCann “correct” answer would in any way more likely be a better detective than the person who selected the alternative. In addition, I found ten questions unsuitable, either because the knowledge tested for seemed totally unrelated to performance of a detective’s duties, or, in one or two instances concerning legal matters, the “correct” answer was wrong or the question is currently.subject to such judicial dispute as to preclude a “correct” answer. Identification of these ten unsuitable questions and the eight questions for which an additional answer was appropriate presents three choices. First, it might be possible to find the exam adequately job-related notwithstanding these dubious questions. The existence of a small number of inappropriate questions need not inevitably defeat a finding of job-relatedness. After all, every candidate who took this exam had the right to an administrative appeal before the local Civil Service Commission with review thereafter in the state courts, and in such proceedings he could litigate the legitimacy of any particular question or the correctness of any answer, and secure a recomputation of his score to the extent that he prevailed. See Lukachik v. Jankura, 27 Conn.Sup. 1, 228 A.2d 523 (1966). The second choice is to consider the number of unsatisfactory questions sufficiently high to defeat a finding of job-relatedness. That might be a permissible conclusion, but surely not compelled at the level of dubious questions encountered here, unless there is some reason to believe that the number of unsatisfactory questions may have contributed to the racially disparate test scores. That possibility suggests another approach. The third choice is to rescore the exam, eliminating the ten unsuitable questions, and allowing credit for an additional “correct” answer to eight questions. This possibility seems to have received little consideration in the reported cases. It was suggested as appropriate relief in Vulcan to remedy the inclusion in a fireman’s test of 20 civics questions, 490 F.2d at 393-94, and apparently rejected because of infirmities in the balance of the exam. Id. at 397. Inquiry as to rescoring seems appropriate here for three reasons. The number of unsuitable questions is not clearly so low as to be ignored nor so high as to be fatal. Moreover, the candidates do have the opportunity to challenge the impact of any particular questions upon their scores. Finally, consideration must be given to the limits of relief in the event of a finding that the exam is not job-related. This exam, unlike the one in Vulcan, is a promotional *939 exam. Our Circuit has ruled that while quota hiring is a permissible remedy for discrimination at entry levels, it is not appropriate for promotional ranks. Bridgeport Guardians I, 482 F.2d at 1341. Though the Court of Appeals noted the absence of a finding that the promotional exam there challenged was not job-related, the principal ground of the decision as to remedy appears to have been the unfairness of quota promotional hiring upon those already in the department. [T]he imposition of quotas will obviously discriminate against those Whites who have embarked on a police career with the expectation of advancement only to be now thwarted because of their color alone. The impact of the quota upon these men would be harsh and can only exacerbate rather than diminish racial attitudes. Ibid. If a quota is not an available remedy for use of an inadequate exam with discriminatory impact at a promotional level, the principal appropriate remedy appears to be a new exam, one free of the defects of the challenged exam. Rescoring provides the equivalent of that remedy. More pertinently, it- affords an opportunity to determine whether the unsuitable questions have contributed to the racially disparate results. At the Court’s request, the McCann firm rescored the exam. Accepting an additional “correct” answer to eight questions raised most scores one or two percentage points, but this change, and the elimination of the ten unsuitable questions, produced very slight variation in the ranking of the candidates. Most significant to this lawsuit, the reseoring did not produce a passing grade for any Black candidate and did so for only one Hispanic applicant. 2 The White passing rate increased from 24% to 29% and the minority passing rate (by virtue of the one passing candidate) increased from 0% to 5%. Plainly the unsuitable questions do not contribute, to the racially disparate scores. Thus, a determination that these questions render the exam not job-related with a consequent remedy of a new exam or a required use of the rescored results of this exam would not remedy nor significantly impact the racial grievance that gave rise to this lawsuit. At most it would simply substitute this Court’s judgment as to the suitability of some questions and the correctness of some answers for the judgments that are available, upon proper challenge, to be made by the Civil Service Commission and the state courts. That is not the function of a Title YII suit. Passing Score A final issue, also raised by Dr. Barrett’s criticisms, concerns the appropriateness of the 75% passing score that Bridgeport requires for all municipal employment tests. When the exam was constructed, the McCann firm reckoned with this specified cut-off point. In the opinion of McCann’s president, Forbes McCann, a police promotion exam should be so structured that successful applicants score at least seven percentage points better than the average score achieved by candidates taking similar exams throughout the country. In his opinion, this seven-point margin above the national average is needed to provide “a reasonable likelihood” of success on the job. (Tr. 140). Knowing that a passing score was required to be 75%, the McCann firm selected questions on which the average correct-answer rate achieved by detective candidates throughout the country for all 100 questions was 68%. Thus, at least in theory, a Bridgeport detective candidate who scored at least 75 and thereby passed would have scored at least seven points better than the national average of detective candidates tested on these same questions. Dr. Barrett found this approach inadequate. He expressed the criticism that the *940 passing grade was not shown to distinguish between those who are qualified for the job and those who are not. Cut-off points have also been criticized in prior cases. In Kirkland, the point was selected at the highest level permitted by law, a point likely to produce enough successful candidates to meet the expected need. 374 F.Supp. at 1377. Similarly, in Western Addition Community Organization v. Alioto, 360 F.Supp. 733, 738 (N.D.Cal.1973), the cut-off point was selected, after the test results were computed, at a level to meet known needs. In Officers for Justice v. Civil Service Commission, 371 F.Supp. 1328, 1338 (N.D.Cal.1973), the cut-off point was selected arbitrarily. While the original selection of the cut-off point here was arbitrary, see Bridgeport Guardians I, 482 F.2d at 1338, the relationship between this test and that cut-off point is not arbitrary. Those who constructed the test set out to distinguish between those who would score above and those who would score below a point somewhat above the average achieved by the many detective candidates who had previously answered the same questions. It makes no difference whether a passing grade is selected as appropriate for an exam, or, as in this case, an exam is selected as appropriate for a specified passing grade. The issue is whether the relationship between exam and passing score is based on a sufficiently sound approach. In spme fields of employment, it may be realistic to insist lipón a cut-off score that distinguishes between those who are. qualified and those who are not. For some tasks, the principal and perhaps only legitimate inquiry of an employer may be whether a person can do the job or not. But for many tasks, and surely for detective work, it is not realistic to restrict an employer to that choice. Within limits of reason, a city must be free to decide that it wants more than minimal qualification, that it wants a suitable standard of competence. In this case, when city officials retained the McCann firm, they left to that organization the judgment as to how much competence was appropriate for promotion to the rank of detective. 3 McCann thought successful candidates should be those who scored seven points above the national average of all patrolmen answering the same questions. That is a reasonable judgment as to required test performance. The EEOC guidelines specify only that the cutoff score “will be reasonable and consistent with normal expectations of proficiency within the work force. . . ” 29 C.F.R. § 1607.6. What is less certain is whether scoring seven points above the national average is a sufficiently reliable indicator of job performance at the level of competence sought by use of this test. This is the issue that casts the greatest doubt on use of content validation. For, as Dr. Barrett noted, a test that measures knowledge is not a certain indicator of performance and, as he conceded, assurance that test results correlate closely with performance can be provided only by some form of criterion-related validation. But these problems are inherent in content validation. That approach has been accepted by the EEOC and approved in this Circuit in Vulcan. Given the availability of that approach, T am satisfied that McCann’s “7% solution” is not a ground for rejecting this test. Conclusion Considering the preparation of the test, the content of the test, the relationship between the test and the passing score, and the opposing opinions of Dr. Barrett and Forbes McCann, I am satisfied that the test is sufficiently job-related on the basis of content to withstand a challenge under Title VII, notwithstanding its racially disparate impact. While plaintiffs have contended that multiple-choice exams are available for police promotion selection that do not have a racially discriminatory impact, they presented no evidence of such *941 tests or their results. Though the burden of proof is upon the defendants to show adequate job-relatedness once a sufficient racial disparity in test scores is evident, the defendants’ evidence, if generally satisfactory, is entitled to be considered persuasive in the absence of any evidence that other tests would achieve more racially neutral results. See Albemarle Paper Co. v. Moody, supra, 422 U.S. at 425, 95 S.Ct. 2362. The plaintiffs have no burden of producing such evidence, but they cannot complain if, in the absence of such evidence, a court finds satisfactory a defendant’s showing of content validity. Of course, the rejection of a Title VII challenge does not mean that the defendants’ use of this and similar tests is beyond criticism. The distressing absence of minority group members from the supervisory ranks of the Bridgeport Police Department should be a cause for continuing concern by responsible officials of that city. They need not be content with a test about which the most favorable statement that can be made is that it is not unlawful. Their preference for objectively scored tests to minimize claims of improper influences in selection procedures is understandable. But they need not leave to the plaintiffs the task of searching for tests that hold some promise of achieving more racially neutral results. The efforts they make in this direction will be relevant to any future challenges to tests that produce racially disparate results. On the basis of all the evidence presented, since there is not a probability of success on the merits, the motion for a preliminary injunction is denied. The Temporary Restraining Order previously entered is terminated. Footnotes [1] The data from the job analysis were kept as notes of those who did the work at Bridgeport because city officials instructed McCann to halt further work sometime in 1974 as a result of litigation. When the officials were satisfied in 1976 that the way was clear to proceed with a promotion exam for detectives, they asked McCann to complete its assignment as quickly as possible. McCann’s subsequent work proceeded from the notes made in 1974; the notes were not reduced to a written job analysis until just prior to the court hearing. That approach is not preferable, but not a sufficient infirmity to undermine content validity. [2] The McCann firm at no time identified candidates by race. It computed the original and revised scores solely from results identified only by candidates’ coded numbers. Identity of . a candidate’s number is known only by the Civil Service personnel. The Court was furnished with a list of numbers identifying the race (but not the name) of minority candidates so that the racial significance of rescoring could be determined. [3] While it would surely be sounder municipal management for such decisions to be made by appropriate local officials, such delegation does , not create a vulnerability under Title VII.