COFFIN, Circuit Judge.
This case epitomizes the classic, clumsy and yet unavoidable attempt to rectify, through the courts, long standing though not consciously intended discriminatory selection policies in public employment in a northern community. The fact that the discrimination was not intended has raised somewhat novel issues relating to the standards for judging policies
The plaintiffs are six black and two Spanish-surnamed residents of Boston who claim violations of their civil rights under 42 U.S.C. § 1981 and § 1983 in the police recruiting and hiring practices of defendants, the members of the Massachusetts Civil Service Commission, the Director of Civil Service, and the Commissioner of the Boston Police Department. Plaintiffs brought the action on their own behalf and on behalf of black and Spanish-surnamed persons who have applied, might apply, or, but for the alleged discriminatory practices, might have applied for the job of policeman in Boston, in the other cities and towns of Massachusetts, or with the Metropolitan District Commission (M.D.C.), the Massachusetts Bay Transportation Authority (M.B.T.A.), or the Capitol Police. Their claims are that they and their class suffer discrimination in the dissemination of information about employment opportunities; in the imposition of an irrelevant and discriminatory educational requirement; and in the use of a non-job predictive and discriminatory written examination, a five foot seven inch minimum height, and a one hundred yard swimming test. Plaintiffs seek declaratory and injunctive relief barring recourse to existing eligibility lists or to any requirement which has not been found to be predictive of successful performance on the job, and affirmatively, ordering preferential hiring of qualified blacks and Spanish-surnamed applicants until the percentage of such officers on the Boston force equals the percentage of black and Spanish-surnamed people in the Boston population.
After a substantial trial, the court addressed itself in three separate rulings to plaintiffs' request that a class be certified. It first held that, all the evidence having been submitted, the request was not timely. It then ruled that plaintiffs, all of whom had taken one or more police examinations, were not in the position of those black and Spanish-surnamed individuals who had not known of the examinations or had been discouraged from taking them. Finally, and most importantly, it observed that if a comprehensive class sharing the same characteristics were to be recognized, it would be not all black and Spanish-surnamed actual or potential applicants but rather all who wished police employment, took the examination, and failed for cultural reasons. The court variously described the class as also including "non-mainstream whites" and, most comprehensively, "minorities which did not share the prevailing white culture: that is, . . . groups such as blacks, yellows, browns, American Indians, persons reared in lands where the preferred language is not English, and even whites from backwood areas."
In its opinion on the merits, 334 F.Supp. 930 (D.Mass.1971), the court first acknowledged that there was no claim or evidence that defendants had a discriminatory purpose or pursued any practice that was discriminatory on its face. It then approached each of plaintiffs' claims to inquire whether the challenged standard or requirement was significantly related to a policeman's job and, if not, whether objectively it had the consequence of discriminating against black and Spanish-surnamed police applicants. It found a discrepancy sufficient to put it on inquiry; namely, a 16.3 per cent black population in Boston in 1970, with blacks on the police force representing only 3.6 per cent.
As to the principal grievance, shared by all the named plaintiffs, that the written examination was both non-job predictive and discriminatory, the court had much more to say. Though the examinations after 1967 placed less stress on memory and general information and more on "general intelligence" than in earlier years, the court found that the test designers lacked training in either test design or police selection; that the passing grade of 70 was arbitrarily selected; that no attempt was made to relate the questions to an occupational analysis of a policeman's job; that stress was placed on verbal and academic skills having little relevance; and that the racial data collected by questionnaire for the September, 1970, examination revealed that while only 25 per cent of the black and 10 per cent of the Spanish-surnamed applicants passed, 65 per cent of all others did. The court held not only that the examinations given in 1968-1970 were not rationally related to the capacity to perform a policeman's job but that they were "discriminatory against minorities which did not share the prevailing white culture."
The court proceeded to fashion a remedy stemming from its findings. The action against the defendant Police Commissioner was dismissed, since his actions were based upon the examinations, the responsibility of the Civil Service defendants, and involved no further discrimination on his part. Plaintiffs were held entitled to a declaration that the 1968-1970 examinations were not significantly job-related and "gave a discriminatory advantage to white persons whose original primary language was English, and who had been trained and educated in the mainstream of American society"; and an injunction against issuance of further certificates of eligibility based on these examinations and against the use of any examination which is not significantly job related or which discriminates against any "racial, national, cultural, or other comparable group."
On the principal claim in the case—that the entrance examinations were discriminatory—the court, after finding both discrimination and a lack of significant job relatedness, adopted the proposition that the racially discriminatory effect
The result of the affirmative relief granted is that police selection practices in Massachusetts will start with a clean slate—though "slate" is hardly the word to connote something so overwhelming white. For the dynamics are such as to relegate to the remote future the achievement of significant representation of black and Spanish-surnamed persons on metropolitan or Commonwealth police forces. Boston's police force, the central focus of interest, consisted of 2805 officers in 1970, of whom 2.3 per cent, or 65, were black or Spanish-surnamed. The yearly rate of appointment from these minority groups from 1960 through 1970 has, with the exception of 1968 when a special program was pursued, ranged from zero to five. The relief ordered by the district court is unlikely to increase significantly this level of representation. It can be calculated from the stipulations that approximately 5 per cent of the applicants to the Boston force who took the September, 1970, examination and completed the questionnaire were black or Spanish-surnamed. Assuming that black and Spanish-surnamed persons continue to apply to the force in the same ratio to other persons, that they pass as frequently as do others a perfectly nondiscriminatory examination, and that 100 appointments are made per year (1960-1970 average: 87), five black or Spanish-surnamed applicants will be appointed each year. On these assumptions, the percentage of black and Spanish-surnamed policemen on the Boston force will increase at a rate of less than .002 per cent per year.
This projected profile would be acceptable if the court was correct in declining to treat the case as one involving a claim validly limited to racial discrimination. The question whether a claim can be so limited appears to be a novel one. In the conventional discrimination case, review is triggered by the use of a racial classification or by a deliberate purpose to exclude blacks or another sharply identifiable racial group from rights or privileges. When, however, review is triggered by a racially discriminatory impact, the claim cannot take its coloration from the explicit racially discriminatory intent of the discriminators. The issue, then, is whether the racial characterization can be permitted as being justified on other grounds.
In this case, for example, we agree intuitively with the district court that the vice of the police entrance examinations was that they favored "mainstream whites" and discriminated against all others. If statistics had been gathered to show such factors as school years completed, family size and income, occupation of parents, city or town where an applicant had lived his youthful years, etc., there might well be correlations establishing, prima facie, a discrimination against members of low income families generally, against ghetto residents generally, or against members of families which had a primary language other than
Apart from the conceptual problems, the fact is that in this case review was triggered and the trial went to conclusion on the basis of hard statistical data showing that one racial group and one cultural-language group fared demonstrably worse than others. And there was no evidence as to other racial or cultural groupings. This is the same posture as that which was presented in such cases as Griggs v. Duke Power Co., 401 U.S. 424, 91 S.Ct. 849, 28 L.Ed.2d 158 (1971); Carter v. Gallagher, 452 F.2d 315 (8th Cir. 1972); Chance Board of Examiners, 458 F.2d 1167 (2d Cir., 1972); Penn v. Stumpf, 308 F.Supp. 1238 (N.D. Calif.1970); Arrington v. Massachusetts Bay Transp. Auth., 306 F.Supp. 1355 (D.Mass.1969). In all these cases specific minority groups challenged hiring practices as being racially discriminatory in effect. In all of them it could have been as accurately stated as here that the examinations and practices had a discriminatory effect broader than racial. While the issue does not seem to have been confronted directly, none of the courts restated the claims in such all-embracing terms. This seems to us implicit recognition that a statistical case showing a racial discrimination is not to be deflected by applying the observation that discrete racial and cultural groups have no monopoly of the burdens imposed by the malfunctioning of our society. Were this not so, as is evident from this case, racial groups, though possessed of the strongest evidence of discrimination against them, must either take on other people's burdens, with the attendant practical problems, or forego the vindication of their own rights. We therefore hold that the court erred in refusing to recognize this as a case involving racial discrimination.
The Request for Class Certification
As we have noted, the court's view of the nature of the discrimination inherent in the police entrance examination played a critical part in its refusal to certify the action as a class action. True, the court also deemed the formal written request, made after conclusion of the evidentiary proceedings, to be untimely. But the complaint had described the action as a class action; plaintiffs had, on the second day of trial, orally moved for certification; defendant Police Commissioner McNamara had admitted the class allegations in his responsive pleading; the evidence was throughout consistent with plaintiffs' aspirations to represent a class; and no objections or countersuggestions had been made. Having in mind the court's responsibility under Rule 23(c) (1) F.R. Civ.P. to determine, as soon as practicable after a class action is brought, whether it is to be maintained as such,
This does not mean that it was an abuse of discretion not to have granted plaintiffs' request as made. As will later appear, we find that as to some of the subsidiary claims, the plaintiffs have failed to make a prima facie case of racially discriminatory impact. The district court was within its discretion in treating those as individual rather than class claims. We refer additionally to the court's ruling that plaintiffs, all of whom had been aware of the police employment opportunities and none of whom had been deterred from taking one or more examinations, could not satisfy the requirement of Rule 23(a) (2) that failure adequately to apprise black and Spanish-surnamed persons of employment opportunities was a question of law or fact common to the class; or of 23(a) (3) that the claim of the representative parties as to being prejudiced by discrimination in recruitment information —which did not exist in their cases— was typical of the claims of the class. While discrimination in recruitment has been successfully raised in some class actions, we are not disposed to disturb the court's ruling on this issue.
Burden of Proof
In the general course, a court faced with a claimed denial of equal protection must first ascertain whether the plaintiff has made such a threshold showing as to require a justification, must then identify the classification employed, and must finally determine whether the classification has been justified under governing standards. In pursuing an inquiry as to the relationship between each requirement and successful performance on the job, a court might adopt a relaxed standard of review whereby it would be a sufficient justification that under some reasonable version of the facts the classification is rationally related to a permissible goal, here the selection of qualified persons for public employment. Cf. McGowan v. Maryland, 366 U.S. 420, 426, 81 S.Ct. 1101, 6 L.Ed.2d 393 (1960). Such a justification is clearly adequate where the initial showing is simply that a plaintiff or class of plaintiffs has been excluded from employment by the classification.
As to classifications which have been shown to have a racially discriminatory impact, more is required by way of justification. The public employer must, we think, in order to justify the use of a means of selection shown to have a racially disproportionate impact, demonstrate that the means is in fact substantially related to job performance. It may not, to state the matter another way, rely on any reasonable version of the facts, but must come forward with convincing facts establishing a fit between the qualification and the job. In so concluding, we rely in part on the Supreme Court's opinion in Griggs v. Duke Power Co., supra. Faced with a showing of racially discriminatory impact without intent, the Court invalidated a requirement in the alternative which it found to be a barrier to transfer among the company's departments, stating that
Although differing from the present case in the respects that it was a decision under Title VII of the Civil Rights Act of 1964 and that it concerned intra-company assignment of employees by a private employer, Griggs construed broad statutory language proscribing classification of employees
With substitution of "applicants" for "employees", the quoted statutory language not unfairly describes what we hold to be the constitutional constraints on a public employer. We cannot conceive that the words of the Fourteenth Amendment, as it has been applied in racial cases, demand anything less.
Drawing upon cases where an explicit or implicit racial classification was challenged, plaintiffs would have us require demonstration of a "compelling interest" or of "compelling necessity". To the extent that those phrases advert simply to the importance of the state's goal or to the substantial congruence of employment requirements to job performance, this asks no more than what we have already conceded. To the extent, on the other hand, that the phrases connote a lack of alternative means, we do not think them appropriate in their full rigor to a racial impact case such as is now before us. In effect, rigorous application of an alternative means test would, irrespective of history or purpose to discriminate, make it a denial of equal protection for a public employer not to utilize a racial quota, and accordingly, a denial of equal protection for the employer not to establish compensatory programs to remedy disparities of preparation. Such a result would be to translate what is a discretionary power of courts in giving relief into a mandatory standing obligation for employers. No employment case cited to us even arguably goes this far. A public employer may constitutionally take its applicants as it finds them. And while alternative means of selection not involving compensatory training are surely relevant, a court would go astray in this technical field to strike down an otherwise properly justified means of selection in favor of a purely speculative alternative. Despite the variances in language endemic to equal protection analysis, other constitutional decisions concerning public employer hiring practices appear to be in basic accord. See, e. g., Carter v. Gallagher, supra; Penn v. Stumpf, supra; Arrington v. Massachusetts Bay Transp. Auth., supra, 306 F.Supp. at 1358; cf. Chance v. Board of Examiners, supra, at 1176-1178. If, on the other hand, experimental evidence or expert testimony showed two methods to be adequate screening devices, but one to be discriminatory and the other not, a hiring authority would be bound to use the non-discriminatory one.
The Height Requirement
The current status of height requirements under Massachusetts law is a matter of some doubt. When this action was commenced in the district court, a Massachusetts statute provided that
On June 3, 1971, the proviso was amended to read
The statute is silent as to the continuing validity of such five foot seven inch requirements as may have been enacted by city councils prior to its passage, ambiguous as to the authority of a city council now to impose without the director's approval a minimum height requirement of precisely five feet seven inches, and unclear even as to the survival or not of a statutory five foot seven inch minimum height. Nor do the briefs draw our attention to any evidence in the record pinpointing city council action as a source of the height minimum allegedly enforced by Boston or by any other city either prior to or after the statute was amended nor to approval by the director of a height requirement for the police force of any city or town. Moreover, the Boston Commissioner of Police has disavowed mechanical adherence to any height minimum, and the Civil Service Commission has evidenced a similar attitude.
Even were an insistence upon a minimum height of five feet seven inches made clear, however, the record would be infirm in another respect. Quite simply, the plaintiffs have failed to demonstrate that a minimum height requirement has a disproportionate impact on Spanish-surnamed persons. Nor can a court employ a rigorous standard of review on the basis of a mere supposition that a classification has such an impact. While one Spanish-surnamed plaintiff had standing to raise the claim, having shown that he was less than five feet seven inches tall, no data was presented concerning the average of Spanish-surnamed males as compared with other males, either for any city, for Massachusetts, or even for the nation. Superintendent Sullivan testified that from his personal observation the average height of Spanish-surnamed males in Boston was, indeed, five feet seven inches, while another bit of testimony was that Spanish-surnamed males seemed smaller than stateside Americans. The only other evidence was an equally inconsequential summary of a police conference report indicating that one person present was of the opinion that Puerto Ricans could not meet the requirement and a two-line excerpt from an unreported opinion of the Equal Employment Opportunities Commission regarding the average height of Puerto Ricans. This simply does not suffice. In the absence of a showing of prima facie discriminatory impact, the standard of review is, as we have indicated, a relaxed one, which a minimum height requirement for policemen clearly meets.
The Swim Test Requirement
Since 1968 applicants for all police forces except the M.B.T.A. and Capitol forces have been required, apparently by the Civil Service Commission, to pass a one hundred yard swim test. Two individual plaintiffs who are non-swimmers have standing to raise the claim, but again plaintiffs have offered no comparative racial statistics. And, again, the Civil Service defendants evidenced a sensible willingness to remove the swim test as a mandatory requirement. While we would impose on the state a heavy burden of justification had a disproportionate racial impact been shown, in the absence of a demonstrated impact the swim test was a not irrational means of testing one aspect of physical fitness.
The Educational Requirement
The educational requirement can be met by graduation from high school, by a certificate of equivalency, or by an honorable discharge after three years of military service. M.G.L.A. c. 31, § 6A. While the evidence shows a significant disparity between blacks and Spanish-surnamed persons and others in respect to the percentage who have a high school education, we have no evidence showing to what extent blacks and Spanish-surnamed persons meet one of these alternative tests, nor do we have a composite picture.
In any event, however, we agree with the district court that the requirement is clearly valid. Importantly, it is supported by a "meaningful study of [its] relationship to job-performance ability". Griggs v. Duke Power Co., supra, 401 U.S. at 431, 91 S.Ct. at 853. We refer to the 1967 report of the President's Commission on Law Enforcement and Administration of Justice, "The Challenge of Crime in a Free Society", pp. 106-10,
At the most recent administration of the civil service examination, the only one for which racial statistics were compiled, only 25 per cent of the eighty black applicants and 10 per cent of the Spanish-surnamed applicants passed, while 65 per cent of all others did. This prima facie showing of racially discriminatory impact was acknowledged by the Civil Service defendants, both in the district court and in their brief on appeal. In the absence of a satisfactory justification, the district court found that the examinations given from 1968 through 1970 were not "significantly related to the capacity of applicants to be trained for or to perform a policeman's job."
The court's opinion seems, however, to suggest that in addition the examinations were manifestly irrational hurdles. For reasons which will appear when we discuss relief, this is of some significance. We do not agree that the examinations cannot survive even the accommodating McGowan standard. When pressed, plaintiffs' own expert witness would not make "the categorical statement that the content of this exam or certain parts of it would, categorically, not be related." Though testing of verbal ability and reading comprehension and retention might well not be broad indicators of success on the job, the Civil Service could rationally believe such to be related to some of the policeman's functions.
The relief decreed by the district court reflected its characterization of the 1968-1970 examinations as discriminating broadly against all non-mainstream whites. This meant, first, that all those presently on the lists of eligibles were in such a position at the expense of many other groups. And, secondly, the breadth of the discrimination counselled against any preferential, compensatory treatment of blacks and Spanish-surnamed persons. Our contrary view of the discrimination leads us to remedial measures differently affecting both those now on eligible lists and the class represented by plaintiffs.
As to those now eligible for appointment, the class represented by intervenors, we see no compulsion to strip them of all advantage. The only proven vice of the examinations was that they discriminated against the racial and cultural groups represented by plaintiffs. To the extent that eligibles profited at the expense of these groups, they should forfeit advantage—but not to any greater extent. The fact that others unknown, unrepresented, and uncomplaining may have been disadvantaged by the examinations does not rise to the level of constitutional defect voiding the examinations for all purposes. Subject to the following comments, therefore, we hold that the current lists of eligibles remain valid.
As to the class now before us, those black and Spanish-surnamed applicants who failed one or more of the examinations given during the period from 1968 to 1970, we feel that some form of compensatory relief is mandated. See, e. g., Carter v. Gallagher, supra, 452 F.2d at 328-31.
Our suggested range of ratios may appear to be unduly strong medicine. We do not think so. Our view of the probable size of the priority pool of those who will have met all non-discriminatory requirements is that it will not be large. If it is, this will enter into the court's judgment in fixing a smaller ratio. If on the other hand, the residue in the pool is exceedingly small, a larger ratio could well be adopted. We believe that preferential status for the priority pool will yield a significant increment of black and Spanish-surnamed police officers in the near term. We have therefore refrained from couching our guidelines in terms of percentage representation, which would have entailed resort to assumptions as to relevant areas and populations, and within these, the numbers and proportions of black and Spanish-surnamed persons disposed to seek police work.
Since, according to a stipulation, the Director of Civil Service certifies seven names for every multiple of five vacancies, the remedial approach we have outlined cannot be fully effective without the cooperation of the appointing authorities in the various police forces. Of these, only the Boston Commissioner of Police was named a defendant, and the record amply demonstrates his willingness to assist in all efforts to achieve a wider racial and cultural representation on his force. As a matter purely of legal nicety, we think it appropriate that he remain a party. By the same token, the absence of other appointing authorities does not at this stage compel further modification of the certification procedure, there being no evidence which would lead us to expect that they will be any less cooperative.
One final point. The district court did not require that the new examination be submitted to court and counsel, but rather set forth guidelines which, if followed, would ensure prima facie validation. Some have protested that nothing less than adherence to the guidelines governing examinations under the Equal Employment Opportunity Act, 32 C.F.R. § 1607, would suffice. We do not see fit to interfere with the exercise of the court's discretion on such a point. We have interposed enough. We add only
Affirmed in part; reversed in part; remanded for further proceedings in accordance with this opinion.