United States v. City of Chicago, 385 F. Supp. 543 (N.D. Ill. 1974)
November 7, 1974
CITY OF CHICAGO et al., Defendants, Louis Arado et al., Intervenors-Defendants.
Renault ROBINSON and Afro-American Patrolmen's League, an Illinois not-for-profit corporation, Plaintiffs,
James B. CONLISK, Jr.,[*] et al., Defendants.
Tadeo Robert CAMACHO, et al., Plaintiffs,
CITY OF CHICAGO, et al., Defendants.
United States District Court, N. D. Illinois, E. D.
*544 *545 Richard J. Phelan, Chicago, Ill., Sp. Counsel, Isham, Lincoln & Beale, Chicago, Ill., William R. Quinlan, Daniel R. Pascale, Richard F. Friedman, Ann Acker, Jerome A. Siegan, Michael Small, Asst. Corp. Counsels, City of Chicago, Chicago, Ill., for defendants.
Philip B. Kurland, Norman J. Barry, Donald E. Egan, Alan L. Unikel, Rothschild, Barry & Myers, Chicago, Ill., for intervenors.
Ilana Diamond Rovner, Asst. U. S. Atty., Chicago, Ill., David W. Allen, Donald Pailen, David L. Rose, Eileen M. Stein, U. S. Dept. of Justice, Washington, D. C., for United States.
Thomas A. Gottschalk, Gary M. Elden, John W. Conniff, John P. Wilson, Jr., James A. Cherney, Kirkland & Ellis, Chicago, Ill., for Robinson, and another.
Michael L. Meyer, Barry S. Alberts, Schiff, Hardin & Waite, William J. McNally, Judith S. Bernstein, Chicago, Ill., for Camacho and others.
PRELIMINARY INJUNCTION ORDER
MARSHALL, District Judge.
This case having come before the court for a consolidated hearing on plaintiffs' motions for preliminary injunction, and the Court having heard the testimony of the witnesses of the parties for approximately 17 days and having heard oral arguments, and the court having considered the oral and documentary evidence presented, the briefs, and arguments of counsel; and having made findings of fact and conclusions of law, and being of the opinion that this preliminary injunction should be entered,
It is hereby ordered, adjudged and decreed:
1. The defendants City of Chicago, James M. Rochford, Superintendent, Chicago Police Department, William E. Cahill, Reginald Dubois, Quentin J. Goodwin, Commissioners, Chicago Civil Service Commission, and Charles A. Pounian, Secretary, Chicago Civil Service Commission, and their officials, agents, employees, and all persons and organizations in active concert or participation with them are enjoined until further order of court from engaging in any act or practice which has the purpose or effect of discriminating against any employee of, or any applicant or potential applicant for employment with the Chicago Police Department because of such individual's race, sex, color, or national origin, and specifically from:
(a) Failing or refusing to recruit, hire, assign and promote black and Spanish-surnamed persons on an equal basis with whites of non-Spanish origin,
(b) Failing or refusing to recruit, hire, assign and promote women on an equal basis with men,
(c) Failing or refusing to eliminate qualifications, tests, standards and procedures which are not job-related *546 and which disproportionately exclude blacks, Spanish-surnamed persons or women from employment opportunities in the police department as compared with white males of non-Spanish origin.
2. No further certifications for appointment to the rank of Patrolman for the Chicago Police Department will be made from the current Patrolman's eligibility list based on the 1971 Patrolman's Examination, until further order of court. Defendants shall not administer or utilize police entrance examinations of the type administered in 1971.
3. No further use shall be made of a background investigation or the results thereof as a standard of appointment to the rank of Patrolman for the Chicago Police Department unless objective criteria are established and validated as job related or shown to have no adverse racial impact.
4. No further certifications for promotion to the rank of Sergeant for the Chicago Police Department will be made from the current eligibility list based on the 1973 Sergeant's Examination, until further order of court. Defendants shall not administer or utilize sergeant promotion examinations of the type administered in 1973.
5. No further use shall be made of the Chicago Police Department efficiency ratings as a standard or factor for promotion within the Chicago Police Department unless objective criteria are established and validated as job related or shown to have no adverse racial impact.
6. The classification Patrolman will be discontinued and the classification Patrol Officer (or some similar term) will be substituted therefor, and women shall henceforth be recruited, hired and assigned to positions in the Chicago Police Department in accordance with the same standards and procedures and on an equal basis with men.
7. This preliminary injunction shall issue without bond and remain in effect until further order of this court.
These consolidated civil rights actions challenge the hiring and promotion practices of the Chicago Police Department (hereafter the "Department"), which, it is said, impermissibly discriminate against women, blacks and Spanish surname Americans (hereafter "Hispanics"). The actions are here under 42 U.S.C. §§ 1981 and 1983 and their jurisdictional counterparts, 28 U.S.C. §§ 1331, 1343(3) and 1343(4), and 42 U.S. C. §§ 2000e et seq., and 28 U.S.C. § 1345.
Other claims are and have been made by the various plaintiffs. In 70 C 2220, Renault Robinson and the Afro-American Patrolmen's League assert First Amendment violations arising out of alleged discriminatory and chilling disciplinary action taken by certain of the defendants against Robinson and members of the League for which they seek money damages as well as injunctive relief. In 73 C 1252 Tadeo Camacho and his fellow plaintiffs alleged (and settled by consent decree entered June 28, 1974) that certain of defendants' height, weight and medical requirements for patrolman recruits impermissibly discriminated against blacks and Hispanics. In 73 C 2080 the Government complains that defendants' police employment and promotion practices contravene not only 42 U.S.C. § 2000e-2, but also the regulations and guidelines of the Department of Justice and the Law Enforcement Assistance Administration which provide for equal employment opportunity in federally assisted programs and activities, 28 C.F.R. 42.201, et seq., and 42.301, et seq. But the claims which are peculiar to particular plaintiffs have not detracted from the common questions of law and fact which attend the allegations of discriminatory hiring and promotion practices. Accordingly, the cases have heretofore been consolidated pursuant to Rule 42(a) of the Federal Rules of Civil Procedure.
There are pending the motions of all plaintiffs for preliminary injunctive relief on the common questions of hiring *547 and promotion. An extensive evidentiary hearing was held pursuant to Rule 65(a) (2), F.R.Civ.P. The parties have submitted voluminous briefs and proposed findings and conclusions. The motions are ready for decision. In essence, plaintiffs seek to have defendants restrained pendente lite, from hiring any new patrolmen from the 1971 eligible roster and promoting any new sergeants from the 1973 eligible roster because, as plaintiffs aver, the former is the product of a discriminatory written examination (hereafter "the 1971 patrolman's exam") and a discriminatory background check, and the latter is a product of a discriminatory written examination (hereafter "the 1973 sergeants' exam") and discriminatory departmental performance/efficiency ratings (hereafter "efficiency ratings").
Since under Illinois law defendants can appoint patrolmen and promote sergeants only from the 1971 and 1973 rosters respectively, Ill.Rev.Stat.1973, ch. 24, ¶ 10-1-1 et seq., to grant plaintiffs' motions will be to restrain defendants from making any patrol appointments or sergeant promotions pending decision of the case on the merits. In recognition of these consequences as well as for the purpose, as they say, of preliminarily remedying past discrimination, plaintiffs urge that the preliminary injunction permit defendants to appoint and promote from the challenged rosters pursuant to a suggested formula which will result in the disproportionate appointment and promotion of blacks and Hispanics. Most outspoken in their opposition to and criticism of the formula or "quota" are the intervening defendants who are a group of patrolmen in the Department who took the 1973 sergeants' exam and who hold positions on the 1973 sergeants' roster. They assert that any decree which orders or approves promotion of a patrolman because of his race will effectively deprive them of their property without due process of law and deny them equal protection of the laws in violation of the Fifth and Fourteenth Amendments.
A motion for a preliminary injunction is always addressed to the court's discretion. Plaintiffs must establish the likelihood of ultimate success on the merits. They must show that they or those whom they represent, will suffer irreparable harm if preliminary relief is not granted. They must also show that the consequences which will flow to the defendants if preliminary relief is granted, do not outweigh the harm to the plaintiffs if it is denied.
Plaintiffs have clearly met the first two of the standards. For the reasons hereafter stated, they have established a strong likelihood of ultimate success. And it cannot any longer be questioned that practices which deny members of minority groups an equal opportunity to compete for employment inflict a terrible and irreparable injury upon them.
However, the potential consequences of the requested preliminary restraint to defendants and those whom they represent are great. The Department, in deference to this action brought by the Government on August 13, 1973, has refrained from appointing patrolmen or promoting sergeants for over a year during which more than 700 patrolman and 50-100 sergeant vacancies have accrued. The vacancies will be frozen by a preliminary decree and they will undoubtedly increase through normal attrition during that same period. Eventually, the shortage of patrol officers will become critical in a city the size and complexity of Chicago.
Thus, were the issues which form the crux of the pending motions a long time from final resolution, the scales might tip against granting preliminary relief. The parties have, however, indicated that they have pretty well exhausted their proof on the question of the "validity" of the 1971 patrolman's exam and the 1973 sergeants' exam. Rule 65(a) (2), F.R.Civ.P. If that be so, the findings and conclusions rendered here on those questions of "validity" will become the final findings and conclusions on those issues. In that posture of the case, the injunctive relief plaintiffs now *548 seek would be required as the only effective remedy pending development by defendants of valid testing procedures.
Furthermore, the evidence is that defendants are already well along in the development of a new patrol officers' examination which, hopefully, will meet the requirements of law. Thus the further hiatus in patrol hiring occasioned by the grant of preliminary relief should not be lengthy.
Finally, while this memorandum and the accompanying preliminary injunction do not embrace plaintiffs' quota or formula solution to the vacancy problem, it has been deferred, not rejected. An application by the parties upon a showing of necessity for interim relief will be entertained. Compare, Vulcan Society, etc. v. Commission, 360 F. Supp. 1265, 1278 (S.D.N.Y.1973).
Therefore, for the following reasons, which will stand as written findings and conclusions under Rule 52(a), F.R.Civ. P., a preliminary injunction will issue restraining defendants from making patrolman appointments to the Chicago Police Department from the 1971 patrolman eligible list, from continued use of the present method of background investigation regarding patrolman appointments and from making any further promotions to the rank of sergeant from the 1973 sergeants' eligible list.
PLAINTIFFS' CASE IN CHIEF
Plaintiffs' case in chief developed along lines which have become commonplace in litigation of this type. See, Griggs v. Duke Power Co., 401 U.S. 424, 91 S. Ct. 849, 28 L. Ed. 2d 158 (1971); Vulcan Society, etc. v. Commission, supra, aff'd 490 F.2d 387 (2d Cir. 1973); NAACP v. Beecher, 504 F.2d 1017 (1st Cir. No. 74-1067, Sept. 18, 1974); Shield Club et al. v. City of Cleveland, 370 F. Supp. 251 (N.D.Ohio, 1972). The following was proved.
As of August, 1973, there were approximately 115 women employed in sworn (or officer) positions by the Department out of 13,500 sworn personnel. Thus women, while comprising 50% or more of the persons eligible by residence for employment with the Department, constituted only .85% of the Department's officer complement. At the time of trial, the number had risen to approximately 160 or 1%.
It is undisputed that prior to the commencement of the Government's action, women in the Department were employed, assigned and promoted on a segregated basis. They were not permitted to compete with men on entrance examinations. Their eligibility was limited to a relatively few positions as "police women" and "police matrons." They were given limited responsibility, primarily for processing, searching and care and custody of women prisoners, and a limited amount of youth work. They were not used for patrol work.
At the time of the Government's action, there was one sergeant of police women, and one lieutenant of police women. Women were allowed to compete with men for promotions purposes for the first time on August 18, 1973, after the Government's action had been brought.
Defendants are presently in the process of developing what they have denominated as a "unisex" entrance examination under which women will be examined and compete with men for patrol appointments.
2. 1971 Patrolmen Hiring
Patrolmen must reside in the City of Chicago, the population of which is approximately 60% white, 33% black and 7% Hispanic. At the present, the Department's sworn (or officer) personnel are approximately 83% white, 16% black and 1% Hispanic. In the mid-1960's when the black population of the City was approximately 25%, the black sworn personnel of the Department fluctuated between 22 and 26%. In the late 60's it began to decline, although the black population of the City was increasing in gross as well as percentage. Today the Department falls far short of *549 mirroring the community which it represents and from which its membership must be recruited.
(a) The 1971 Patrolman's Exam
On December 4, 1971, the 1971 patrolman's exam was administered by the Civil Service Commission (hereafter "the Commission") for the purpose of selecting patrolmen for the Department. The Commission prepared the exam, its answers and scored it. The exam consisted of 120 multiple choice, machine graded questions to be answered in two hours. There was one "correct" answer to each question. Essentially it was an aptitude test which covered vocabulary, verbal analogies, grammar, number series, arithmetic problems and reasoning, and word usage. While there were (and are) no educational requirements for appointment as a patrolman (i. e., high school diploma), the exam was such that the better an applicant's education, the better he should perform on the exam.
Those who answered correctly 75 out of the 120 questions (or 62.5%) "passed" and proceeded to the next screening step a physical and medical examination. Those who passed the physical and medical examination were posted on the eligible list in the order of their scores on the written exam.
8,136 applicants wrote the 1971 exam. Of these, 66% were white, 29+% were black, and 4+% were Hispanic, in comparison to a City mix of 60-33-7.
Black applicants failed (i. e., scored less than 62.5%) the 1971 exam at a rate of 67%. Hispanic applicants failed at a rate of 68%. White applicants failed at a rate of 33%. The black and Hispanic applicants who passed tended to score lower than the white applicants who passed. Thus, of the 1478 patrolmen who have been appointed in the order of their exam scores from the 1971 roster, only 147 (or 10%) are black and 23 (1.5%) are Hispanic.
(b) The Background Investigation
As the Department needs patrolmen, the Commission furnishes names of applicants from the top of the eligible list. The Department then conducts an investigation to determine whether the applicant's "background" disqualifies him from employment.
The background investigation is conducted without regulations, standards or guidelines other than to determine whether the applicant is of "bad character, dissolute habits or [guilty of] immoral conduct," or whether he has made any false statements to the Commission, violated any Commission rule or been dismissed for cause from public employment. The investigation includes arrests, convictions, traffic citations, whether any family member has ever been arrested for a serious crime, education, employment, military service, driving history and financial condition. Based upon the investigation, the Recruit Processing Section of the Department makes a determination whether the applicant should be disqualified for employment. No objective system of evaluation is used in making this determination. If the Section recommends disqualification, its recommendation is approved by the Department in 90% of the cases. A disqualified applicant is notified in writing of his disqualification, but not the reasons therefor. The notice directs him to appear before the Commission and show cause why he should not be disqualified. If he appears, he is then advised of the reasons for his disqualification. State court judicial review of a contested adverse decision by the Commission is available.
Across the board, since 1968, 25.7% of the black as compared to 15.2% of the white applicants who have undergone background investigations have been disqualified. For certain reasons, the rate of black to white disqualifications has been as high as 2 to 1 (arrest record) *550 and 3 to 1 (negative employment record).
(c) Summary of 1971 Patrolmen Hiring
In summary then, the plaintiffs' evidence showed that the racial composition of the Department does not reflect that of the population of the City (83% white, 16% black and 1% Hispanic compared to 60% white, 33% black and 7% Hispanic). The defendants' hiring practices have operated to exclude greater percentages of blacks and Hispanics than whites. The 1971 patrolman's exam disqualified only 1/3 of the white applicants while 2/3 of the black and Hispanic applicants were disqualified. Of those who passed the 1971 exam, blacks and Hispanics scored lower than whites. When impact of the exam was coupled with the physical and medical requirements (which have been settled) and the background investigation, the results have been patrol appointments from the 1971 eligible roster of 88.5% white, 10% black and 1.5% Hispanic as compared to an original applicant pool of 66% white, 29+% black and 4+% Hispanic.
In the language of the cases, defendants' hiring practices have been shown to have a "racially disproportionate impact," Castro v. Beecher, 459 F.2d 725, 732 (1st Cir. 1973); Vulcan Society, etc. v. Commission, 490 F.2d 387, 393 (2d Cir. 1973); NAACP v. Beecher, 504 F.2d 1017 (1st Cir. No. 74-1067), Sept. 18, 1974), the consequences of which are discussed hereafter.
3. The 1973 Sergeants' Roster
Blacks and Hispanics constitute 17% of all uniformed personnel in the Department. They represent, however, only 9% of all persons in the ranks above patrolman and only 10% of all sergeants.
In 1968, an examination for the rank of sergeant was given to 5429 candidates. Of that total, 4393 (80.9%) were white and 1036 (19.1%) were black. Of the first 298 people who were promoted to sergeant from the 1968 list, 293 (90%) were white and 35 (10%) were black.
(a) The 1973 Sergeants' Exam
On August 18, 1973, the 1973 sergeants' exam was administered to 6555 candidates, of whom 1185 (18%) were black, 113 (1.7%) were Hispanic and 5257 (80%) were white. The exam consisted of 100 multiple choice single "correct" answer questions which were machine graded. It was an achievement type of examination which, on its face, appears to be based upon police training and experience and calls for responses to hypothetical situations which might confront a sergeant.
The examination was scored and an eligibility list published based upon the exam scores (60%), efficiency ratings for the six-month period January-June, 1973 (30%), and seniority (10%). Candidates making a composite score of 70 or above were considered eligible for promotion and their names were placed on the eligibility list in the order of their composite scores. Of the 6555 patrolmen who wrote the exam, 1918 obtained the required composite score of 70. They were 87% white, 11% black and 2% Hispanic.
The percentage of white candidates who took the 1973 sergeants' exam and obtained a composite score of 70 or higher was 32%; the percentage of black candidates similarly achieving *551 was 17%. Thus the success rates of whites was almost double that of the blacks.
In recent years the Department has annually promoted between 50 and 100 patrolmen to the rank of sergeant and has used the sergeant promotion list for no more than four years. Accordingly, the number of candidates on the eligible list who are likely to be promoted does not exceed 400. Among the first 400 names on the eligible list, 21 (5.3%) are black, 8 (1.9%) are Hispanic and the balance of 371 (92.8%) are white, compared with an initial candidate population of 18% black, 1.7% Hispanic and 80% white. There are only five minority candidates in the first 100 on the list.
The 400th person on the eligibility list had a composite score of 73.90. That score represents a practical cut-off point between those likely to be promoted and those unlikely to be promoted. 7.07% of the white candidates obtained a composite score of 73.90 while 2.23% of the blacks and Hispanics obtained that score.
(b) The Efficiency Ratings
Supervisory rank officers, normally sergeants, rate patrolmen every six months. The period January-June 1973 was used in compiling the 1973 sergeants' roster. The rating card lists five traits (quantity of work, quality of work, personal relations, dependability and attendance and promptness). Each trait is to be rated on a scale of 0-100 and then an average of those 5 scores is struck for the patrolman's overall rating. It is a Department tradition to average the overall efficiency ratings of units or groups of patrolmen at 85. As a consequence, 92% of all patrolmen receive overall ratings between 80-95. During the period January-June 1973, the mean efficiency rating for blacks was 84.3 and for whites 85.2. Under the 30% for efficiency ratings method used to compile the sergeants' roster, this difference translates into a mean difference of approximately .3 points on the roster.
Of the candidates taking the 1973 sergeants' exam, 18.1% were black. Of the candidates obtaining the top exam scores, approximately 10% were black. Of the candidates obtaining the top composite scores (i. e., after efficiency ratings were added to the written scores), only 5.3% were black.
355 candidates had efficiency ratings of 95 and above. Of those, 314 (89%) were white, 35 (10%) were black and 6 (1%) were Hispanic. Among the highest 514 efficiency marks given to candidates, 446 (86.7%) were white, 60 (11.6%) were black, and 8 (1.5%) were Hispanic.
The effect of efficiency ratings on position on the sergeants' roster is disclosed by a comparison of written exam scores with the composite eligibility scores which include efficiency ratings. There are 24 (8.9%) blacks included in the top 267 test scores, which consist of all candidates with a score of 64 and above. The top 267 on the sergeants' roster, however, include only 14 (5.2%) blacks. Among the top 383 test scores which include all candidates with a score of 63 and above, there are 36 (9.4%) blacks. The first 383 candidates on the sergeants' roster include only 20 (5.2%) blacks. Among the top 514 written test scores which include all candidates with a score of 62 and above, there were 48 (or 9.3%) blacks. The first 514 on the sergeants' roster include only 31 (or 6%) blacks.
(c) Summary re the 1973 Sergeants' Roster
In 1973, there were 11,431 patrolmen in the Chicago Police Department of which 17% were black and Hispanic and 83% were white. Over half the patrolmen competed on the 1973 sergeants' examination. Of the 6555 patrolmen taking it, 20% were black and Hispanic and 80% were white. To be placed on the eligibility list, a composite score of 70 was required. A total of 1,918 patrolmen made the roster. Of these, 13% were black and Hispanic and over 87% were white. The success rate for white *552 patrolmen was nearly twice that of blacks: 32% to 17%.
A maximum of 400 persons can expect promotion from the 1973 sergeants' roster. The first 400 persons on the roster include only 21 blacks (or 5.3%) and 8 Hispanics (or 1.9%). Whites comprise 92.8% of those patrolmen with a reasonable expectation of promotion. Thus, the "practical" success rates of whites and blacks and Hispanics was 371 out of 5257 or 7.07% for whites and 29 out of 1298 or 2.23% for blacks and Hispanics. Among those likely to be promoted, the success rate of whites was over three times as great as that of blacks and Hispanics.
Again, in the language of the cases, defendants' promotion practices have been shown to have a "racially disproportionate impact," the consequences of which follow.
THE BURDEN OF PERSUASION
The foregoing computations which defendants do not challenge in themselves, establish a prima facie case of de facto racial discrimination in employment and promotion. Vulcan Society, etc. v. Commission, 360 F. Supp. 1265, 1269 (S.D.N.Y.1973), aff'd 490 F.2d 387 (2d Cir. 1973); Carter v. Gallagher, 452 F.2d 315, 323 (8th Cir. 1971), cert. denied, 406 U.S. 950, 92 S. Ct. 2045, 32 L. Ed. 2d 338 (1972); NAACP v. Beecher, 504 F.2d 1017 (1st Cir. No. 74-1067, Sept. 18, 1974); Chance v. Board of Examiners, 458 F.2d 1167 (2d Cir. 1972). Defendants dispute the prima facie consequences of the statistical showing of disproportionate passing and acceptance rates. They assert that statistical imbalance alone has never constituted a sufficient showing of denial of equal protection of the laws. Yet, a review of the authorities previously cited, which in the main take their lead from Griggs v. Duke Power Co., 401 U.S. 424, 91 S. Ct. 849, 28 L. Ed. 2d 158 (1971), make clear that a showing of imbalance, the likes of which has here been demonstrated, does constitute a prima facie case of discrimination in hiring and promotion.
Defendants also deny that a similar strength should be accorded the disproportionate results of the patrolman's background investigation and the departmental efficiency ratings. The disparity in disqualification under the former (25% black and 15% white) and the impact of the latter when coupled with the 1973 sergeants' exam scores, lead to the conclusion that they are deserving of inquiry. Bridgeport Guardians, Inc. v. Commission, 482 F.2d 1333, 1335 (2d Cir. 1973). This is particularly so where, as here, the background investigation and efficiency rating practices are inextricably entwined with written testing practices which demand scrutiny.
The conclusion that a prima facie case has been established by plaintiffs is, however, only one of the consequences of plaintiffs' proof. Plaintiffs assert that their case is so strong as to shift to defendants the burden of persuading the court that the challenged practices are job-related and hence valid. Defendants demur arguing that this case is no different from any other civil case and that the burden of persuasion rests with the plaintiffs if defendants adduce any evidence which tends to show the job relatedness and hence the validity of their practices. Careful consideration of the authorities leads inescapably to the conclusion that plaintiffs' case is of sufficient strength to shift to the defendants the burden of persuasion.
Again, Griggs, supra, is the beginning point. There the Court was faced with testing practices which had a disproportionate impact comparable to those shown here. In holding the tests violated Title VII of the Civil Rights Act of 1964, the Court said (401 U.S. at 431, 432, 91 S.Ct. at 853):
". . . The Act proscribes not only overt discrimination but also practices that are fair in form, but discriminatory in operation. The touchstone is business necessity. If *553 an employment practice which operates to exclude Negroes cannot be shown to be related to job performance, the practice is prohibited.
"On the record before us . . . the general intelligence test is [not] shown to bear a demonstrable relationship to successful performance of the jobs for which it was used. . . ."
". . . [G]ood intent or absence of discriminatory intent does not redeem employment procedures or testing mechanisms that operate as `built-in headwinds' for minority groups and are unrelated to measuring job capability.
". . . Congress directed the thrust of the Act to the consequences of employment practices, not simply motivation. More than that, Congress has placed on the employer the burden of showing that any given requirement must have a manifest relationship to the employment in question."
Courts of appeals construing and applying Griggs to both private and state and local public employment practices, have consistently concluded that upon a showing of disproportionate impact, the likes of which has been made here, the burden of persuasion is placed upon the employer with respect to the validity of the employment practice. The employer "must come forward with convincing facts establishing a fit between the qualifications and the job." Castro v. Beecher, 459 F.2d 725, 732 (1st Cir. 1972.) "[A] showing of a racially disproportionate impact puts on the municipal or state defendants not simply a burden of going forward but a burden of persuasion." Vulcan Society, etc. v. Commission, 490 F.2d 387, 393 (2d Cir. 1973).
In United States v. United Brotherhood of Carpenters and Joiners of America, Local 169, 457 F.2d 210, 214 (7th Cir. 1972), cert. denied, 409 U.S. 851, 93 S. Ct. 63, 34 L. Ed. 2d 94 (1972), the Court of Appeals for this circuit held that once a prima facie case is established in a Title VII action, "the burden of going forward and the burden of persuasion [are] shifted to the accused . . . ." And in United States v. Hayes International Corp., 456 F.2d 112, 120 (5th Cir. 1972), the Court of Appeals for the Fifth Circuit stated that once a prima facie case is established, "the onus of going forward with the evidence and the burden of persuasion is [on the defendant]."
In short, the disproportionate impact raises more than an inference of impermissible discrimination which a fact finder is permitted to draw. The conclusion of discrimination is in the nature of a presumption and is required unless the employer persuades the fact finder that the practices used fit the job, predict performance on the job, and are necessary for the on-going effectiveness and efficiency of the particular employing agency.
Defendants urge that it is unfair and unprecedented to impose such a burden upon them. The precedent has already been discussed. As for the fairness, who is in a better position to establish the validity and necessity of the examination than the employer who relies upon it? It is the employer who knows the demands of the job. It is the employer who has had experience with multiple applicants and performers. It is the employer who either developed and administers the challenged practices or who commissioned their development and administration.
Defendants make one final argument in respect to plaintiffs' prima facie case. They assert that the Government's action must fail because it has failed to establish a "pattern or practice" within the meaning of 42 U.S.C. § 2000e-6. Thus, they maintain that the authority of the Attorney General to file a civil action *554 is limited to those situations in which the defendant has engaged "in a pattern or a practice of resistance to the full enjoyment of any rights secured by" the Act. Pointing to the Seventh Circuit's decision in United States v. International Association of B. S. & O. I. W., Local No. 1, 438 F.2d 679 (1971), and the Fourth Circuit's decision in United States v. Hunter, 459 F.2d 205 (1972), they urge that at best an isolated or accidental instance of conduct violative of the Act has been shown.
It would be a waste of resources and a gross misuse of power were the Attorney General to champion vicariously the cause of every individual allegedly discriminated against in his or her employment. And that is the import of Section 2000e-6. But that is not the situation here. In the case at bar, thousands of applicants have been screened through the results of the 1971 patrolman's exam and background investigation and thousands more have been screened through the 1973 sergeants' exam and departmental efficiency ratings. This is no isolated instance of discrimination. If defendants' employment and promotion practices are discriminatory, certainly they constitute a "pattern or practice" within the meaning of the Act.
THE EVIDENCE OF JOB RELATEDNESS
Plaintiffs having proved significant statistical disparities in both the 1971 patrolman's roster and the 1973 sergeants' roster, the burden rests with the defendants to persuade the court that their selection practices are job related. A job related selection practice is one which accurately appraises the ability and fitness of the candidate to perform the job he seeks or at least is designed to make that appraisal. Therefore, if defendants' selection practices are job related, it follows that black and Hispanic candidates have less ability and are less fit to perform the jobs of patrolman sergeant than their white counterparts. Vulcan Society, etc. v. Commission, 490 F.2d 387, 392 (2d Cir. 1973).
Defendants do not, however, urge that conclusion. Quite the contrary: at every opportunity they have stated their conviction, based upon experience in the field, that blacks and Hispanics are as qualified as whites to perform the duties of patrolman and sergeant. If the conviction is firm, one would expect defendants themselves to question the job relatedness of their selection practices. And this is particularly so in light of their recruitment practices of the mid-1960's which did not produce the white-minority disparities they are presently experiencing.
However, rather than joining in a search for the cause of the disparities to the end that they might be remedied, defendants have chosen to lead the court "deep into the jargon of psychological testing." Vulcan Society, etc., supra, 490 F.2d at 394. The result has been a virtual morass of competing theories advanced by professional testers of tests in which the debate has centered on predictive, concurrent, criterion and construct *555 validation and the court has been left with the unwelcomed task of testing the testers. It is not amiss to observe that plaintiffs have not shunned the debate.
No effort has been made to validate the discriminatory treatment of women in hiring and promotion. Defendants suggest that the issue is moot because they have extended patrol opportunities to women on a volunteer basis and they plan to offer a "unisex" exam to the next group of patrol officer candidates. The issues are not moot and the discriminatory treatment is clearly unlawful and should be enjoined. United States v. Local 38, IBEW, 428 F.2d 144, 151 (6th Cir. 1970), cert. denied, 400 U.S. 943, 91 S. Ct. 245, 27 L. Ed. 2d 248; Cypress v. Newport News General and Nonsectarian Hospital Association, 375 F.2d 648 (4th Cir. 1967).
2. 1971 Patrolmen Hiring
As previously noted, the 1971 patrolmen hiring is the product of the 1971 exam, physical and medical requirements which have been settled and the background investigation. Each operated as a "built-in headwind" for blacks and Hispanics. The first was the written exam.
(a) The 1971 Patrolman's Exam
At the outset the exam must be recognized for what it is: a poor aptitude examination seemingly designed to test vocabulary, verbal analogies, grammar, number series, arithmetic problems and reasoning and word usage. The 120 questions included a number of items which were ambiguous, had more than one possible correct answer, no correct answer or for which credit was given for an answer which was arguably incorrect.
The 1971 patrolman's exam was virtually the same as the 1970 exam which black and Hispanic candidates had failed at a rate of 77% and 70% respectively while whites failed at a rate of only 42%. Nevertheless, prior to the preparation and administration of the 1971 exam no analysis of the patrol function was made. The test was drawn on the basis of a job description which was not produced at the hearing. According to the testimony of defendants' witnesses, one of the 1971 exam purposes was to ascertain the candidate's ability to express himself in writing because, as the same witnesses testified, one patrolman's lot is one of report writing.
On the eve of trial, defendants engaged Prof. Keith J. Edwards to conduct a validity study of the exam. The study consumed 15 days. It included a six-hour "ride-along" with a couple of patrolmen as well as some discussions with officers. This was the extent of his review of the patrol function. As a consequence, he was unable to express an opinion as to the exam's construct validity.
The balance of Prof. Edwards' time was spent on an attempt at criterion validation in which he sought statistically significant correlations between the 1971 exam scores and efficiency ratings, departmental awards, performance on the 1973 sergeants' exam and disciplinary actions.
Insofar as the efficiency rating correlation is concerned, Prof. Edwards admitted it was not statistically significant. Furthermore, as discussed hereafter, the efficiency ratings themselves have not been shown to be an accurate measure of a patrolman's job performance.
Prof. Edwards' effort to show a correlation between exam scores and departmental awards was equally fruitless. In the first place, while there was a measurable correlation for whites, he found none for blacks. Furthermore, there was no definition of awards, how they are made or their reliability as a criterion of overall performance.
The correlations with the 1973 sergeants' exam showed that there is a relationship between success on two written examinations. But the sergeants' exam was not shown to be a measure of *556 patrolman performance and "promotability" is not a proper criterion when so few patrolmen (400 out of 12,000) are promoted.
Prof. Edwards analysis regarding disciplinary actions was not disclosed on direct examination presumably because, as he testified on cross, he found no correlation with the exam scores.
Finally, defendants sought to show a correlation between the results of the 1971 patrolman's exam (and earlier exams) and performance in the police academy. Again one would expect a correlation between performance on written exams. But significantly, defendants established no corresponding correlation between academy performance and patrol job performance. It is the latter that the 1971 exam must be related to. The absence of evidence of a correlation between academy and the job would suggest that there is none.
In conclusion, defendants' efforts to show the job relatedness of the 1971 patrolman's examination have fallen far short of the mark. The exam's disproportionate impact on blacks and Hispanics has not been justified as either an accurate or necessary predictor of patrol job performance. It follows that the hiring of patrolmen from the 1971 roster constructed from the results of the exam must be enjoined. Vulcan Society, etc. v. Commission, supra; Shield Club et al. v. City of Cleveland, supra.
(b) The Background Investigation
The details and procedures of the background investigation are not disputed by defendants. They consist of inquiry by the Recruit Processing Section of the Department into the social status, arrest and conviction records, education, employment, military background, driving history and finances of the candidate.
The standards used by the Recruit Processing Section include criteria such as "bad character, dissolute habits, and immoral conduct." Lt. Chausse, supervisor of the Recruit Processing Section, testified that there were no other standards or regulations defining those criteria. Indeed, he admitted that he had no idea what "dissolute habits" meant until shortly before his appearance at the hearing.
If the Recruit Processing Section determines that an applicant should be disqualified for employment as a patrolman, it forwards its recommendation to the Deputy Superintendent of the Department, who, in 90% of the cases, approves the recommendation and forwards it to the Civil Service Commission. The Commission then writes the applicant, directing him to appear before it and show cause why his certification to the position of patrolman should not be disapproved. Prior to his appearance, the applicant is not given a copy of the report of the Department, is not informed that he has the right to review the report, and is not even informed of the reason for his disqualification.
The practical inability of applicants to present any challenge to the Department's disqualification decision is fundamentally unfair. It prevents applicants from determining prior to the hearing whether the basis for their disqualification was illegal. Thus, the background investigation includes consideration of an applicant's record of arrests, convictions and other police contacts. In Gregory v. Litton Systems, Inc., 316 F. Supp. 401 (C.D.Cal.1970), aff'd as modified, 472 F.2d 731 (9th Cir. 1972), the district court declared that "the policy of excluding from employment persons who have suffered a number of arrests without any convictions, is unlawful under Title VII." The court cited "overwhelming and utterly convincing" national statistics which establish that blacks are more frequently arrested than whites in proportion to their numbers in the general population. Although the Department no longer asks applicants directly whether they have ever been arrested, the defendants nonetheless obtain that information from police agencies and from neighbors, and applicants are *557 asked about traffic citations, pending warrants and whether any member of their immediate family has ever been arrested for a serious crime.
In the background investigation, defendants also consider an applicant's socio-economic status. The utilization of such criteria has been held to constitute discrimination since blacks are more likely than whites to possess negative socio-economic attributes. For example, several EEOC decisions have held that an employer may not inquire into an applicant's financial status (such as credit information and disclosure of past-due loans) because census data indicates that a disproportionate percentage of non-whites are below the poverty level and are therefore more likely to suffer financial difficulties. EEOC Decision No. 74-27, 2 CCH, Employment Practices Guide, ¶ 6396 (Sept. 17, 1973); EEOC Decision No. 74-02, 2 CCH Employment Practices Guide, ¶ 6386 (July 10, 1973); EEOC Decision No. 72-1176, 2 CCH Employment Practices Guide, ¶ 6359 (Feb. 28, 1972). Similarly, in Johnson v. Pike Corporation of America, 332 F. Supp. 490, 494-495 (C.D.Cal.1971), the district court invalidated a company policy of discharging employees whose wages had been garnished, saying:
"The fact that blacks and other racial minorities are so often subject to garnishment action is related to the fact that they are to a disproportionate extent from the lower social and economic segments of our society. . . . The Supreme Court, in Griggs, repeatedly stressed that Congress' intention in Title VII was to invalidate all employment practices which in their final effect or consequence discriminate against racial minorities. A policy of dismissing employees whose wages are attached has this impermissible effect."
See also, Wallace v. Debron Corp., 494 F.2d 674 (8th Cir. 1974).
The defendants also inquire into an applicant's social status, family history and military background. Consideration of these factors has been held to violate Title VII. In EEOC Decision No. 74-25, 2 CCH Employment Practices Guide, § 6400 (Sept. 10, 1973), the Commission invalidated respondent's honorable military discharge requirement, since the foreseeable impact, based on national data, would be that a substantially disproportionate percentage of those persons rejected for lack of an honorable discharge would be black.
Of course, the Department must protect itself from those who would undermine it or work at cross purposes with it. Recent events make that abundantly clear. But it is equally clear that a hiring practice such as this virtually undefined background investigation with its disproportionate impact on minority groups will not pass muster without a persuasive showing that it serves the purpose for which it is intended. That showing has not been made. Accordingly, the utilization of defendants' current background investigation in the hiring of patrol officers must be enjoined.
3. The 1973 Sergeants' Roster
As previously noted, this roster results from a formula in which the score on the 1973 written sergeants' exam comprises 60%, departmental efficiency ratings 30% and seniority 10%. While no challenge has been made against the seniority factor, the disproportionate results of the written exam and the efficiency ratings, separately and combined, are such as to require their validation by defendants.
(a) The 1973 Sergeants' Exam
Defendant undertook to conduct a concurrent validity study of the 1973 sergeants exam under the direction of defendant, Dr. Charles A. Pounian. The sample used in the study was a matched control group of 176 incumbent sergeants, 88 of whom were black (out of 136 black sergeants) and 88 of whom were white (out of 1234 white sergeants). The 88 white sergeants were matched with the 88 black sergeants on the basis of age and efficiency ratings. The 1973 sergeants' exam was given to *558 this sample group. The purpose of the study was to determine if there was a correlation between the participants' performances on the exam and their performances on the job.
The 88 white sergeants selected for the study were not shown to be representative of white sergeants or white patrolman-sergeant candidates. The 88 black sergeants, while representative of that group, were not representative of black patrolman-sergeant candidates. Since the vast majority of white and black patrolmen fail to obtain promotion to sergeant, incumbent sergeants cannot be deemed to be representative of patrolmen who aspire to the rank of sergeant.
The criteria used in the study were the various efficiency ratings received by the sample group of incumbent sergeants during the rating period of January to June, 1973, although defendants' own expert, Dr. Guion, testified to a "great deal of mistrust in the use of administrative rating systems for criterion measures." No reliability studies were performed on the performance ratings despite the fact that in a concurrent validity study the criteria used must represent a fair measure of performance on the job.
The range of test scores for the total sergeant group was 30 to 71, on a scale of 100. The range for the black sergeants was 30 to 71, and for the whites 32 to 70. The average test score for the blacks was 53.72 with a standard deviation of 7.63 and the average for the whites was 53.91 with a standard deviation of 7.44.
Test scores of the black sergeants, the white sergeants and the total group were first plotted against each sergeant's overall efficiency rating. Dr. Pounian reported statistically significant correlations, which were found, however, to be erroneous. The initial erroneous correlation coefficient between the total group's test scores and their efficiency ratings was .247 (significant at the one per cent level). The correlation coefficient between the black sergeants' test scores and their efficiency ratings was .205 (significant at the 5% level). For the white sergeants the correlation coefficient was .319 (significant at the one per cent level). Dr. Pounian also correlated the test scores of the same group with their "quality of work" subrating taken from the overall efficiency ratings used above and reported a correlation coefficient of .507 (significant at the one per cent level). However, this correlation coefficient was also erroneous.
Dr. Pounian then re-analyzed the data and recomputed the correlation coefficients. The correlation coefficient reported between test scores and the overall efficiency ratings was .293 for the total group; and between test scores in the subrating "quality of work" was .514. However, these correlation coefficients, like their forerunners, were found to be erroneous.
A third set of corrected correlation coefficients was submitted by Dr. Pounian. These showed comparisons of test scores with the participants' overall efficiency ratings and, in addition, comparisons with all of the subratings "quality of work", "quantity of work", "dependability", "personal relationships" and "attendance and promptness." These correlation coefficients were substantially lower. Thus, of the six correlations finally *559 computed by defendants for the entire sample of 176 sergeants who took the exam, the highest two correlations were .25 and .26. None of the third set of correlations reaches a level of practical significance in the judgment of either of defendants' expert witnesses, Dr. Pounian or Dr. Ebel.
As might be expected, the testimony of plaintiffs' expert, Dr. Barrett, was that defendants' concurrent validity study had failed to show that the sergeants' examination was a valid or useful test.
The Supreme Court has recognized that the EEOC Guidelines deserve thoughtful consideration in cases of this type and they "demand that employers using tests have available `data demonstrating that the test is predictive of or significantly correlated with important elements of work behavior . . ..'" Griggs v. Duke Power Co., 401 U.S. at 433, 91 S. Ct. at 855, (fn. 9). The data which defendants offered regarding the 1973 sergeants' exam was the above described concurrent validity study.
The concurrent validity study failed to satisfy the minimum requirements set forth in the EEOC guidelines in three key respects. First, the criteria used as measurements of job performance were the sergeants' efficiency ratings and there was no showing that the ratings had been developed through a proper job analysis or that they otherwise possessed validity. They were used simply because they were available.
Second, the sample of incumbent sergeants was not representative of patrolmen applicants. Incumbent sergeants had already demonstrated the ability to overcome the obstacles to promotion posed by a written test. In that respect, they were nonrepresentative of the applicant population and particularly the minority applicant population. See 29 C.F.R. § 1607.5(b) (1).
Third, the study failed in its own objective. Because of reporting errors, levels of statistical and practical significance in some correlations between test scores and efficiency ratings, which the defendants' two experts initially observed, were later corrected and reduced to levels of marginal or no practical significance. See 29 C.F.R. § 1607.5(b) (5).
Thus, the concurrent validity study offered by defendants constitutes an attempt, motivated by anticipated litigation and not in conformity with the EEOC Guidelines, to show a substantial relationship between the written sergeants' exam alone and later success on the job of sergeant. The study, in fact, did the opposite. Once the study's errors were corrected, an insubstantial relationship between success on the written test and success as a sergeant was revealed.
Defendants also attempted to validate the exam by eliciting the view of one expert, in the field primarily of educational testing, that the test was "content valid". Content validity is a form of "rational", rather than "empirical" validity. The analysis depends appreciably on opinions of psychologists. See, "Application of the EEOC Guidelines to Employment Test Validation: A Uniform Standard for Both Public and Private Employers," 41 George Washington L. Rev. 505, 518 (1973).
Here defendants' own concurrent validity study shows that incumbent sergeants, who were rated well by their supervisors, did not know enough of the *560 content on the written test to score appreciably better than the patrolmen candidates. Thus, while patrolmen applicants knew 51% of the test items on the average, incumbent sergeants knew only 53% of the test items on the average.
For a test to be content valid, the knowledge, skills and aptitudes required for successful examination performance must be the knowledge, skills and aptitudes required for successful job performance. Vulcan Society, etc. v. Commission, 490 F.2d 387, 395 (2d Cir. 1973). Any test on which successfully performing incumbent employees do not know 47% of the answers cannot reasonably be considered content valid. The written test does not begin to satisfy the EEOC Guidelines' requirement that a content-valid test consists of "suitable samples of essential knowledge, skills and behavior composing the job in question." 29 C.F.R. § 1607.5(a).
From all the evidence including defendants' own studies, the 1973 sergeants' exam has not been shown to have a demonstrable relationship to successful performance as a sergeant. Use of the 1973 sergeants' eligibility list derived in principal part from that discriminatory test must, therefore, be enjoined.
(b) The Efficiency Ratings
The place of the efficiency ratings in defendants' employment and promotion practices must not be underestimated. Obviously, they play a 30% role in the makeup of the sergeants' promotion roster. Additionally, they have been used in this litigation to "validate" the 1971 patrolman's exam and the 1973 sergeants' exam.
No evidence was offered of any attempt to validate the efficiency ratings as predictors of job success as a sergeant. This is not surprising, for efficiency ratings appear to be highly subjective. They are subject to influences which have little to do with the individual's performance. Several witnesses recognized that racial attitudes might enter into efficiency ratings, but the Department has never made a study to see if race does influence efficiency ratings despite a strong showing by defendants' own analysis of the evidence that blacks receive more ratings beneath 85 than whites do and whites receive more ratings above 85 than blacks do. The difference in means of 85.2 for whites and 84.3 for blacks may appear de minimis. But when 92% of all patrolmen are scored between 80 and 95, the one point difference in means is significant.
The efficiency ratings which count towards a patrolman's composite score in structuring the sergeants' eligibility list are the ratings received for the six-month period immediately prior to the written exam. Since promotional examinations are announced well in advance, supervisory raters know before giving ratings that they will or will not count for the purposes of promotion.
Frequently patrolmen have not been under a rating supervisor's command long enough for an adequate evaluation. When patrolmen are transferred during rating periods and assigned to a new rater, some new supervisors check back for an evaluation but some do not. Desk sergeants who have no knowledge of what a given patrolman has done in the field are asked to rate the patrolman. There is a recurrent suggestion that ratings for a group or unit average 85 (despite the fact that the purported rating scale runs from 0 to 100 points) and that they coincide generally with prior efficiency grades. The policy in favor of an 85 average causes raters to decide the average first and then manipulate the calculations to arrive at the average grade given.
Persons of higher authority are authorized to and do make changes on individual rating cards even though they have no direct knowledge of the performance of the men whose cards they may change. Changes can be made without consulting the person who originally did the rating and at any superior level without informing the last rater.
*561 Patrolmen given certain assignments receive higher ratings. The more an assignment places a patrolman in close and continued contact with a rating supervisor, the higher the supervisory rating of that patrolman is likely to be. Supervisors have been instructed that they should give beat patrolmen an average of 84 to 85, and those in the IAD special unit an average grade of 88.
Defendants' witness, Dr. Guion, and plaintiffs' witness, Dr. Barrett, agreed that supervisory ratings are not a fair measurement of an employee's suitability for promotion. Their view is shared by the courts. Rowe v. General Motors Corp., 457 F.2d 348 (5th Cir. 1972); Wade v. Mississippi Cooperative Extension Service, 372 F. Supp. 126 (N. D.Miss.1974).
On the record here, defendants did not validate the Department's patrolman efficiency ratings as a predictor of sergeant job performance. Accordingly, the use of the 1973 sergeants' eligibility list, which is based to a considerable degree on unvalidated supervisory ratings, must, therefore, be enjoined.
There remains the question of relief.
Plaintiffs seek mandatory relief in the form of minority preferences and hiring and promotion quotas. Defendants seek permission to continue to use the current eligible lists until new examinations are fashioned. Intervening defendants seek to preserve their positions on the sergeants' roster.
Since defendants' practices regarding the hiring of patrolmen and the promotion of sergeants have been found constitutionally and statutorily wanting, their continued use will be preliminarily enjoined.
In the present context of the case, however, preliminary mandatory quota relief is not indicated.
The evidence shows that defendants are presently developing a new patrolman's examination which, hopefully, will cure the deficiencies found in the 1971 exam. It is far better in cases of this nature that the remedy come from the parties rather than the court.
Insofar as promotion quotas are concerned, serious policy and constitutional questions must be considered and ventilated. Bridgeport Guardians, Inc. v. Members of the Bridgeport Civil Service Commission, 482 F.2d 1333, 1341 (2d Cir. 1973). As the court observed in Vulcan Society, supra, 360 F. Supp. 1265 at 1277-1278:
". . . No doubt there are exigent circumstances in which compensatory adjustments are appropriate and even necessary to remedy constitutional violations. However, there are considerations which suggest that the courts should be hesitant about imposing them. Adjustments based upon racial classification, however well-intentioned, contain within themselves the seed of further divisiveness regardless of their benevolent purpose. Attempts to make fair adjustments may be counterproductive and tend to generate resentments which serve to exacerbate rather than to diminish racial attitudes. Those on an eligible list who have high rankings, with the prospect of imminent appointment, may not understand why they should be bypassed in favor of minority applicants with lower rankings. This court is of the view that an examination list which is constitutionally impermissible should be stricken down in its entirety and the authorities directed to conduct a new examination, free from constitutional taint that the remedy is not to continue to base appointments on the condemned list but to require executive or administrative officials to live up to their responsibilities and to prepare and conduct an examination consonant with the Fourteenth Amendment. Only in this way can those authorities and the public understand the courts' insistence upon the sanctity of that Amendment."
For these reasons, the question of mandatory quota relief will be deferred until *562 the entry of the final decree or a request for interim relief is presented.
Accordingly, an order granting preliminary injunctive relief will enter, without bond, in accord with the views herein expressed, without prejudice to an application by the parties, upon a showing of necessity, for interim relief which would permit appointments from the current eligible lists upon an equitable basis pending the development of new hiring and promotion examinations and practices.NOTES
[*] JAMES M. ROCHFORD (substituted 3/14/74 as new Superintendent).
 Certain of the physical (height and weight) and medical (e. g., cardiac) requirements which had a disproportionate impact on blacks and Hispanics have been removed by the consent decree of June 28, 1974.
 Thus a score of 65 on the written exam produced 39 eligibility points.
 Thus an efficiency rating of 90 produced 27 eligibility points.
 The seniority points are not challenged. Virtually all promotion candidates had 10 seniority eligibility points.
 Candidates who are eligible can elect to use 3.5 points as veterans' preference points to advance to a higher position on the list. The veteran preference points are not challenged here. But see, Branch et al. v. DuBois et al., 73 C 2167 in this court.
 Title VII of the Civil Rights Act of 1964 and the Griggs interpretation of it became applicable to state and local employers by the Equal Employment Opportunity Act of 1972 (Pub.L. 92-261, March 24, 1972).
 It is unnecessary to reach the "most difficult question" of whether defendants must meet the compelling state interest equal protection standard in justifying their practices. Compare, Chance v. Board of Examiners, 458 F.2d 1167, 1177 (2d Cir. 1972).
 Nor do defendants argue that the statistical disparity is a reflection of educational and cultural deprivations of minority groups. Were they to do so it would be unavailing. Castro v. Beecher, 459 F.2d 725, 730-731 (1st Cir. 1972); Bridgeport Guardians, Inc. v. Bridgeport Civil Service Commission, 354 F. Supp. 778, 785-786 (D.Conn.1973), aff'd 482 F.2d 1333 (2d Cir. 1973); Vulcan Society, etc. v. Commission, 360 F. Supp. 1265, 1272 (S.D.N.Y.1973), aff'd 490 F.2d 387 (2d Cir. 1973). Intervening defendants do argue somewhat the converse in respect to the 1973 sergeants' roster: that employment opportunities for minorities have so improved in recent years that the highly qualified have left the Department leaving behind only the less qualified to compete with the regular sample of whites. This theory is based upon the testimony of a single member of the Afro-American Patrolmen's League who left the Department for different employment.
 The "average" sergeant participant in the study would not have made the lowest cutoff (i. e. composite score of 70) on the 1973 sergeants' roster unless he had an efficiency rating of 92 or above for the period January-June, 1973. The "average" participant in the study would not make the realistic cut-off of 73.9 regardless of his efficiency rating. I. e., .60(53.91) + .30(100) + 10 = 72.35. It follows that no incumbent sergeant participant who scored below the average would be promoted. To the layman, at least, these circumstances tend to show that either the test is not related to the sergeants' job or half of the incumbent sergeants are not related to their jobs.
 After the proofs were closed on the motions for preliminary relief and plaintiffs' briefs had been filed, defendants ascertained that two answers on the 1973 exam had been "miskeyed," i. e., the Commission's "correct" answers were incorrect. This provoked still a fourth set of correlations which were submitted by mail under date of August 23, 1974. That, in turn, provoked plaintiffs to demand additional discovery. To avoid delay, the court announced that it would not consider the "final" set of correlations.
 Dr. Ebel testified between the second (erroneous) and third set of correlations. His role was essentially that of approving Dr. Pounian's methodology. He did not approve the errors.
 On the average, they did not overcome the obstacle of this test. See note 9, supra, and text.