This case presents a challenge to civil service examinations on the familiar ground that they are racially discriminatory.1 In a thorough opinion, 391 F.Supp. 1064, the United States District for the Southern District of New York, Morris E. Lasker, J., held that five examinations given by the New York City Human Resources Administration unconstitutionally discriminated against black and Hispanic applicants. We affirm.
I
In October 1972, the Human Resources Administration (HRA), which administers various city social services programs, gave a series of nine examinations for certain positions.2 Before us are challenges to five of them, in two consolidated class-action lawsuits.3 After a non-jury trial, Judge Lasker found that the tests had a disproportionate impact on minority applicants and that defendants had not carried their burden of showing that performance on the examination reflected qualification for the job. Accordingly, in a final judgment and order dated May 23, 1975, the judge declared the five tests unconstitutional, enjoined defendants from making appointments *698based on the results of the discredited examinations, and ordered defendants expeditiously to develop lawful and nondiscriminatory selection procedures in accordance with the EEOC Guidelines on Employment Selection Procedures, 29 CFR § 1607.1. He also denied plaintiffs’ request for attorneys’ fees. 391 F.Supp. at 1086-87.
Both sides have appealed. Defendants concede that Judge Lasker correctly stated the legal principles that govern the case:
The ground rules established in [the decisions of this court listed in note 1 supra] require plaintiffs to make a prima facie showing that the examinations have a “racially disproportionate impact,” Vulcan, 490 F.2d at 391, Chance, 458 F.2d at 1175-76 .... Upon such a showing the burden shifts to the defendants to establish that the challenged examinations are job-related, Vulcan, 490 F.2d at 391. The burden on defendants is “a heavy one,” Chance, 458 F.2d at 1176, Guardians, 482 F.2d at 1337, but is discharged if they “come forward with convincing facts establishing a fit between the qualification and the job.” Vulcan, 490 F.2d at 393 .
391 F.Supp. at 1067. They argue, however, that his factual findings of disproportionate racial impact and non-job-relatedness are clearly erroneous. Plaintiffs appeal from the denial of attorneys’ fees.
II
A. Disproportionate Racial Impact
The statistical tables set out in Judge Lasker’s opinion, 391 F.Supp. at 1068-69, show that the passing rates for whites on the challenged examinations were 54%, 54%, 88%, 65% and 51%; for blacks the corresponding percentages were 17, 16, 18, 26 and 31; and for Hispanics, 19, 15, 37, 27 and 19. As the trial court pointed out,
the existing figures for all five examinations clearly indicate a disparity between the passing rates of white and minority candidates in excess of the 1.5 to 1 ratio which Chance held sufficient to establish a prima facie case.
391 F.Supp. at 1069.
Defendants’ primary objection to this analysis is that the statistics on which it is based are incomplete, and therefore form an inadequate basis for Judge Lasker’s conclusion. The problem is that the only individuals taking the test whose race is known are those who were already employed by HRA; no records were kept of the race of other applicants. Thus, for the promotional examinations the race of substantially all applicants is known, but for the three challenged open competitive examinations the differential passing rates cited above were based only on those who were already HRA employees: 51%, 54% and 60% of the total number of candidates.
We agree with Judge Lasker that the incompleteness of the data is not fatal to his findings that plaintiffs had made out a prima facie case of disproportionate impact. The inference that the available data accurately represented the results of the test for all candidates is a reasonable one. In the absence of any reason to believe otherwise, it seems highly unrealistic to believe that minority applicants who were not HRA employees would so far outperform their white counterparts as to wipe out the substantial disparity between the white and minority HRA employees who took the same test, particularly in the light of expert testimony that such a result was unlikely.4
Defendants also argue that on the particular facts of this case, even if the results of the five challenged examinations were racially discriminatory, plaintiffs still did not make out a prima facie case because the results of the other tests in the same series were not ra*699dally disproportionate, and the material covered on the challenged and the unchallenged tests, which were constructed by the same process, was largely identical. This argument does tend to rebut the inference plaintiffs seek to draw from the data, but it is an argument for the trier of fact. In light of Judge Lasker’s careful consideration of the argument and persuasive analysis of the statistics in the record concerning the unchallenged tests, 391 F.Supp. at 1073-75, we cannot find his rejection of defendants’ factual argument clearly erroneous.
B. Job-Relatedness
Defendants’ attack on Judge Lasker’s finding that the challenged examinations were not sufficiently job-related to overcome plaintiffs’ prima facie statistical case has two aspects. Judge Lasker, following the procedure approved by this court in Vulcan Society, supra note 1, 490 F.2d at 395-96, and Kirkland, supra note 1, 520 F.2d at 425 — 26, concentrated his attention on the method of test construction used by the defendants. Defendants argue that (1) the district court’s finding that their method of test construction was “inadequate,” 391 F.Supp. at 1083, was clearly erroneous, and (2) even if that finding was correct, the further finding that the test was not job-related is erroneous because the trial court insufficiently analyzed the content of the test.
Both points are dealt with thoroughly in Judge Lasker’s opinion. Rather than repeat the details of the test-construction method, we refer the reader to the lengthy discussion there, 391 F.Supp. at 1077-84. Defendants stress the expert testimony they presented to support the professionalism of their construction of the challenged tests. At the very least, it must be conceded that unlike the defendants in the cases cited in note 1 supra, HRA had made a good faith effort to prepare adequate job analyses and to construct a test which measured qualities, demanded by the jobs in question. But plaintiffs’ expert testified flatly that the job analysis “does not even remotely meet professional standards,” that “the written test was inadequate as a measure of performance in the job,” and that “there was no evidence for content validity or any other validity of this test.”5 After a careful analysis of the evidence, Judge Lasker chose to believe plaintiffs’ expert testimony rather than that of defendants. We see no reason to disturb his finding.
Defendants’ argument that Judge Lasker insufficiently analyzed the content of the examination is without merit. Pointing to the broad range of functions that might be performed by someone in one of the job categories for which the tests were given, defendants argue that this requires a test that covers not specific knowledge but mastery of certain “core skills” basic to all jobs performed by those in each job title. But Judge Lasker found that defendants had not established the existence of such a common core of skills, or that the job analyses prepared by HRA successfully identified them, or that the examinations given tested for them.- 391 F.Supp. at 1084-85. These findings were clearly permissible on the record before the trial court.
In short, we repeat what we. said in Chance, supra note 1, 458 F.2d at 1175:
While not all of us might have made the same factual finding on the question of job-relatedness as the district judge did, his finding was not clearly wrong.
Ill
Plaintiffs argue that the district court erred in denying them attorneys’ fees. They maintain that plaintiffs su*700ing pursuant to 42 U.S.C. § 1983 are entitled to recover attorneys’ fees as a matter of course, because that statute incorporates the remedial mechanisms of other civil rights statutes, including Title VII of the Civil Rights Act of 1964. The argument is ingenious, but it is clear from Kirkland, supra note 1, 520 F.2d at 430 & n.37, and Bridgeport Guardians, Inc. v. Members of the Bridgeport Civil Service Comm’n, 497 F.2d 1113, 1115 (2d Cir. 1974), cert. denied, 421 U.S. 991, 95 S.Ct. 1997, 44 L.Ed.2d 481 (1975), that this circuit has not accepted the view that attorneys’ fees are routinely available in section 1983 suits.
Plaintiffs also attempt to bring themselves within the exceptions left open by the Supreme Court in Alyeska Pipeline Service Co. v. The Wilderness Society, 421 U.S. 240, 258-59, 95 S.Ct. 1612, 44 L.Ed.2d 141 (1975), to the general rule that attorneys’ fees may not be awarded absent statutory authorization. In light of the explanation of the “common benefit” cases in F. D. Rich Co. v. Industrial Lumber Co., 417 U.S. 116, 130, 94 S.Ct. 2157, 2165, 40 L.Ed.2d 103 (1974), as involving a “shifting of fees . to spread the cost proportionately among the members of the benefited class,” we are not at all sure that plaintiffs come within that exception.6 But even if we were convinced that the exception applied here, the award of fees would still be within the discretion of the district court, see Hall v. Cole, 412 U.S. 1, 93 S.Ct. 1943, 36 L.Ed.2d 702 (1973), and on this record we could not find that the district court abused its discretion in failing to award fees. Finally, plaintiffs argue that Judge Lasker should have awarded attorneys’ fees because of defendants’ “bad faith.” We see no basis on this record for overturning the judge’s finding that “defendants had made reasonable efforts to comply with constitutional requirements.” 391 F.Supp. at 1087.
The judgment of the district court is affirmed.
. See Kirkland v. New York State Dep’t of Correctional Servs., 520 F.2d 420, 425-26 (2d Cir. 1975); Vulcan Society of the New York City Fire Dep’t, Inc. v. Civil Serv. Comm’n, 490 F.2d 387 (2d Cir. 1973); Bridgeport Guardians, Inc. v. Members of the Bridgeport Civil Serv. Comm’n, 482 F.2d 1333 (2d Cir. 1973); Chance v. Board of Examiners, 458 F.2d 1167 (2d Cir. 1972).
. For each of three HRA job titles — Human Resources Specialist, Supervising Human Resources Specialist, and Senior Human Resources Specialist — three examinations were given: an open competitive examination, open to anyone who met certain general qualifications; a promotional examination, open only to HRA employees in the grade next below that for which the examination was given; and a specialty examination in Manpower Development and Training.
. Plaintiffs in Jones challenge the open and promotional examinations for Supervising Human Resources Specialist, and those in Williams attack the open and promotional examinations for Senior Human Resources Specialist and the open examination for Human Resources Specialist. Although Judge Lasker’s reported opinion denies class action certification, 391 F.Supp. at 1086, in subsequent unreported opinions, dated April 4, 1975 and May 9, 1975, class action status is granted and the class defined as “all Blacks and Hispanics who failed any of the five examinations in issue.”
. We recognize that defendants also offered expert testimony which tended to support their view of this question. But it was for the trial court to decide which testimony was more persuasive.
. For a discussion of the different types of test validity, see Vulcan Society, supra note 1, 490 F.2d at 394-95.
. See also Alyeska, 421 U.S. at 257-58, 95 S.Ct. 1612. Lewis v. Texaco Inc., 527 F.2d 921, 926-927 (2d Cir. 1975).