Full opinion text
REAVLEY, Circuit Judge: This is a class action suit brought by six present or retired black employees of Gulf’s Port Arthur, Texas Refinery, all of whom are or were members of the Oil, Chemical and Atomic Workers’ International Union Local 4-23 (“OCAW”), against Gulf and the OCAW for declaratory, in-junctive, and monetary relief under Title VII of the Civil Rights Act of 1964, 42 U.S.C. § 2000e et seq., and the Civil Rights Act of 1866, 42 U.S.C. § 1981. The suit was filed on May 18, 1976, within 90 days of plaintiffs Bernard, Brown and Johnson’s receipt of right-to-sue letters from the Equal Employment Opportunity Commission (EEOC). Plaintiffs amended their complaint to join as defendants other unions which represent employees at the refinery, including: the International Association of Machinists and Aerospace Workers, Port Arthur Lodge No. 823; the International Brotherhood of Electrical Workers and its Local 390; the United Transportation Union and its Local 1071; and the International Union of Bricklayers and Allied Craftsmen and its Local 13 (collectively referred to as the “Trade Unions”). In April 1976, Gulf and the EEOC entered into a conciliation agreement in which Gulf agreed to cease various allegedly discriminatory practices, to establish an affirmative action program with respect to hiring and promotion, and to offer backpay to alleged victims of discrimination based on a set formula. Gulf sent notices to the 632 employees eligible for backpay asking in return for the execution of signed statements releasing Gulf from any possible claims of employment discrimination occurring before the date of the release, including any future effects of past discrimination. This suit was filed one month later seeking to vindicate the rights of many of the employees who were receiving settlement offers from Gulf under the conciliation agreement. The district court entered an order limiting communication by the named parties and their counsel with prospective class members. This court, sitting en banc, held that the order was an unconstitutional prior restraint on expression accorded First Amendment protection. Bernard v. Gulf Oil Co., 619 F.2d 459 (5th Cir.1980) (en banc), aff'd on other grounds, 452 U.S. 89, 101 S.Ct. 2193, 68 L.Ed.2d 693 (1981). On writ of certiorari, the Supreme Court, while not reaching the First Amendment issue, held that the district court’s order was an abuse of discretion under Fed.R.Civ.P. 23(d). Gulf Oil v. Bernard, 452 U.S. 89, 101 S.Ct. 2193, 68 L.Ed.2d 693 (1981). The case was remanded to the district court for further proceedings. On January 24, 1983, the district court tentatively and provisionally certified the class as all blacks who worked or applied for work at the refinery in union covered jobs at any time after December 26, 1966. On April 2, 1984 the court modified the class to exclude unsuccessful job applicants, employees not represented by the OCAW, and employees who signed releases in exchange for backpay under the conciliation agreement between Gulf and the EEOC. The court also limited the relevant time period to employment practices occurring between December 26, 1966, which is 180 days prior to the date on which plaintiffs Bernard, Brown and Johnson filed their charges of discrimination with the EEOC, and May 18, 1976, the date upon which the complaint was filed. The court dismissed the Trade Unions because neither the named plaintiffs nor the modified class contained members of these unions. At trial, plaintiffs advanced a number of classwide claims as well as individual claims on behalf of the named plaintiffs. Plaintiffs contended that: Gulfs seniority system was non-bona fide; a reclassification of employees pursuant to a stipulation (“Stipulation 29”) between Gulf and the OCAW had an unlawfully discriminatory impact on blacks; certain tests administered by Gulf had an unlawfully discriminatory impact on blacks; Gulf intentionally applied its Sickness and Accident (S & A) policy to blacks in a discriminatory fashion; Gulf intentionally discriminated against blacks in its selection of temporary and permanent supervisors; Gulf intentionally discriminated against the named plaintiffs; and the OCAW breached its duty of fair representation. Following a bench trial, the court held for Gulf and the OCAW on all of the classwide and individual claims, 643 F.Supp. 1494. On appeal, plaintiffs assert that the district court incorrectly disposed of each of the classwide and individual claims, but concede that the only issues upon which the OCAW could be liable are those relating to the seniority system and Stipulation 29. In addition, plaintiffs contend that the district court improperly: decertified the class to exclude employees who had signed releases; dismissed the Trade Unions; and refused to consider certain union and company records. We hold that the district court failed to address essential issues of the classwide claims concerning the seniority system, Stipulation 29 and the craft tests as well as certain individual claims of intentional discrimination. We therefore vacate the court’s judgment and remand for further findings. We find no error in the other rulings, which we now address. I. Partial Decertification of the Class, Dismissal of the Trade Unions, and Exclusion of Business Records Plaintiffs challenge the district court’s partial decertification order to the extent that it excluded from the class employees who signed releases in exchange for back-pay under the conciliation agreement between Gulf and the EEOC and dismissed the Trade Unions. The basis for the court’s decision was its view that the named plaintiffs lacked standing both to represent employees who signed the releases and to pursue claims against the Trade Unions. In reaching its decision, the court noted that class representatives must “possess the same interest and suffer the same injury” as the class members, Schlesinger v. Reservists Committee to Stop the War, 418 U.S. 208, 216, 94 S.Ct. 2925, 2930, 41 L.Ed.2d 706 (1974), and that a plaintiff lacks standing to litigate injurious conduct to which he was not subjected. Blum v. Yaretsky, 457 U.S. 991, 999, 102 S.Ct. 2777, 2783, 73 L.Ed.2d 534 (1982). See Vuyanich v. Republic Nat’l Bank of Dallas, 723 F.2d 1195, 1200 (5th Cir.), cert. denied, 469 U.S. 1073, 105 S.Ct. 567, 83 L.Ed.2d 507 (1984). Because none of the named plaintiffs signed the releases or were members of the Trade Unions, the court found that they lacked standing to assert class claims on behalf of these employees. Accordingly, the court limited the class to OCAW members who had not signed releases, and dismissed the Trade Unions. Plaintiffs do not challenge the court’s exclusion of non-OCAW employees, but instead challenge the dismissal of the Trade Unions on the ground that these unions discriminatorily prevented black OCAW members from transferring into jobs covered by these unions. The district court found that plaintiffs did not allege that the Trade Unions acted unlawfully against them. After reviewing the complaint and amended complaints filed by plaintiffs, we find support for the district court’s conclusion and hold that the Trade Unions were properly dismissed. Plaintiffs attack the validity of the releases on the ground that they were obtained from employees during the period when the district court’s order prevented counsel for the class from communicating with prospective class members. The validity of the releases, however, is not at issue. Their validity can be attacked in a separate suit by employees who signed releases during the period when the court’s order was in force. The question presented to the district court was whether the named plaintiffs had standing to represent employees who had signed releases. Because the named plaintiffs sought to obtain relief already obtained by employees who accepted backpay and signed releases, the court found that the plaintiffs lacked standing to represent these employees. We find no error in the court’s determination. Plaintiffs complain of the trial court’s refusal to consider documents, described as business records, which plaintiffs say were given to the clerk during the trial. We find the documents in the record, but they bear no file stamp, and the transcript contains no reference to them. The parties did not agree to their submission. Exhibits come into evidence either by the agreement of the parties or by offer, the laying of a predicate subject to counter by the opposing party, and a ruling admitting the evidence by the trial court. The court correctly disregarded documents that were merely handed to the clerk. II. Gulfs Seniority System Under § 703(h) Plaintiffs contend that Gulf’s pre-Title YII Act seniority system had a discriminatory impact on certain black employees after the effective date of Title VII and that the defendants failed to prove that the seniority system was bona fide under § 703(h). An examination of Gulf’s seniority system is necessary to resolve this claim and plaintiffs’ next claim concerning Stipulation 29. A. Background The Port Arthur refinery was established in 1903 and its labor force was not organized until 1943. Three basic work related areas have existed at the refinery throughout its history: operations, maintenance and labor. The operations group manufactures and processes oil based products and the maintenance group provides repair and construction services. Both groups require skilled employees and until 1954 only whites were hired into them. The district court broadly used the term “craft” to denote the operations and maintenance groups, their employees, lines of progression and the departments and divisions associated with operations and maintenance. We adopt this usage generally, but use the more specific labels where appropriate. The labor group is composed of unskilled employees who assist the craft groups and until 1956 was exclusively comprised of blacks. In 1943 the refinery was organized by the Oil Workers International Union (“OWIU”), the predecessor to the OCAW. At this time, many departments existed and each was functionally related to operations or maintenance. A loosely constructed labor department also existed whose employees were assigned to these various craft departments. Each of these departments had a line of progression for its white employees which indicated the promotional scheme from the entry level position to the top position. It does not appear that clearly defined lines of progression had been established for black laborers. The articles of agreement between the OWIU and Gulf incorporated the existing lines of progression. A seniority system was established based upon plant seniority, defined by the length of service at the refinery, and promotions and demotions along the lines of progression were determined by seniority and ability to perform. A six month residency in one’s particular job classification was a prerequisite to advancement to the next highest job classification in a particular line of progression. Two union locals were established, Local 23 which represented white employees and Local 254 which represented black employees. Members of both locals comprised the union negotiating team. In 1950 the refinery had thirty-four departments: thirty-three craft departments and one labor department. The labor department was composed of many subde-partments, and employees in these subde-partments were assigned to work in twenty of the thirty-three craft departments. Lines of progression had been established in the labor subdepartments, and these lines generally varied according to the structure of the craft departments to which they were assigned. This meant that the twenty integrated craft departments functionally had two separate lines of progression, one for their white employees and the other for the black employees assigned to their departments from the labor subde-partments. These integrated departments also had distinct seniority rosters reflecting this division, one denoted “White” for its employees and the other denoted “Colored” for the assigned labor subdepartment employees. Although black laborers worked in many of the same departments as white craft employees, they could not bid into the more desirable white lines of progression. A gentlemen’s agreement, to which Gulf was not a party, existed between Local 23 and Local 254 which prevented members of either local from bidding on the jobs of the other. This agreement remained in force until 1954, at which time it was temporarily suspended due to pressure from black employees. Ten black employees were permitted to enter white lines of progression in 1954, but no further transfers were allowed until 1956. Between 1954 and 1956, the craft and labor departments were assigned to one of two divisions. The craft departments were assigned to the Operating and Mechanical (“0 & M”) Division, and the labor department, with its various subdepartments, was assigned to the Labor Division. The composition of these departments and subde-partments and their lines of progression remained basically unchanged. Approximately twenty of the craft departments, now in the 0 & M Division, had labor subdepartments assigned to them, and the two separate lines of progression found in these integrated departments before the adoption of the divisional system remained. The labor department employees were technically in the Labor Division even though they physically worked in craft departments in the 0 & M Division. The labels on the seniority rosters in these integrated departments were changed from “White” to “Operating and Mechanical Division” and from “Colored” to “Labor Division.” As of 1956, both black and white employees progressed up their promotional lines based on plant seniority and ability to perform. Newly hired blacks would be assigned to one of the labor subdepartments in the Labor Division, while newly hired whites would be assigned to one of the craft departments in the 0 & M Division. The gentlemen’s agreement between the two locals prevented blacks who progressed to the top positions in their promotional lines from bidding into one of the 0 & M Division lines of progression. The 1956 contract between Gulf and the OCAW made the “Laborer” classification in the Labor Division the entry level position for all employees, both black and white. A high school diploma or its equivalent and adequate performance on a number of tests were required before one could gain employment. After progressing to the top position in one of the Labor lines of progression, an employee could transfer to an entry level position in one of the 0 & M lines of progression after bidding into the mechanical helper pool, which was a classification apparently created to provide apprentice-type instruction for employees entering an 0 & M line of progression. Bids were selected on the basis of plant seniority. The high school diploma and test performance requirements were applicable to transferees who had not previously fulfilled these requirements because hired before 1956. An employee who successfully bid into an 0 & M line of progression was assigned dual seniority: 0 & M Division seniority and plant seniority. Regardless of how much plant seniority a transferee accumulated in the Labor Division, he started with one day’s 0 & M Division seniority for purposes of promotions and demotions in his particular 0 & M line of progression. Plant seniority was important if workforce reductions occurred because a transferee could use this seniority to return to the Labor Division, where plant seniority determined promotions and demotions within that Division. The effect of this system on black employees hired before 1956 is, in part, the basis for plaintiffs’ claim concerning the seniority system. A black employee, hired into the Labor Division in 1950 and who transferred to an 0 & M line of progression in 1960, would have less 0 & M Division seniority than a white employee hired in 1955 into the same 0 & M line of progression. For the purpose of obtaining a promotion to the next highest job classification in the 0 & M line of progression, the black transferee would have only one day’s divisional seniority, compared to the white employee’s five years’ divisional seniority. If workforce reductions were required in 1961, then the black employee could be required to return to the Labor Division, where his eleven years of plant seniority would control his status in that Division. In 1963, the one-day divisional seniority rule and the high school diploma and test performance requirements were eliminated. Plant seniority became the basis for all promotions, demotions and lay offs. The effect of this change was that blacks with more plant seniority, but less 0 & M Division seniority, than whites in the same job classification could compete for promotions to the next highest classification based upon their longer plant seniority, thereby bypassing white employees with greater 0 & M Division seniority. Also in 1963, the two OCAW locals that had separately represented black and white employees merged into Local 4-23, and the facilities at the refinery, such as bathrooms, locker rooms, cafeterias and drinking fountains, were integrated. B. Analysis Plaintiffs contend that Gulf’s pre-Title VII Act (“pre-Act”) seniority system had a discriminatory impact on senior black employees after the effective date of Title VII, and that the defendants failed to prove that the seniority system was bona fide under § 703(h). Before examining plaintiffs’ contentions, we review the relevant law. Section 703(a), 42 U.S.C. § 2000e-2(a), proscribes discriminatory employment practices, including discrimination “with respect to ... compensation, terms, conditions, or privileges of employment” on the basis of race. Section 703(h) insulates bona fide seniority systems from the dictates of § 703(a), by providing that, [notwithstanding any other provision of this subchapter, it shall not be an unlawful employment practice for an employer to apply different standards of compensation, or different terms, conditions, or privileges of employment pursuant to a bona fide seniority or merit system, ... provided that such differences are not the result of an intention to discriminate because of race, color.... Until the Teamsters decision in 1977, International Brotherhood of Teamsters v. United States, 431 U.S. 324, 97 S.Ct. 1843, 52 L.Ed.2d 396 (1977), courts generally rejected the contention that if an employer ceased all discriminatory practices on the effective date of the Act and had a seniority system that was facially neutral, then the system was necessarily protected under § 703(h). See Local 189, United Papermakers and Paperworkers, AFL-CIO, CLC v. United States, 416 F.2d 980 (5th Cir.1969), cert. denied, 397 U.S. 919, 90 S.Ct. 926, 25 L.Ed.2d 100 (1970); Quarles v. Philip Morris, Inc., 279 F.Supp. 505 (E.D.Va.1968). The Quarles court held that a seniority system that perpetuated, or locked in, the effects of past discrimination was not bona fide, and concluded that “Congress did not intend to freeze an entire generation of Negro employees into discriminatory patterns that existed before the act.” Quarles, 279 F.Supp. at 516. The Supreme Court rejected this reasoning in Teamsters. In that case, the Court found systematic and purposeful discrimination in hiring, assignment, transfer and promotion policies before and after the effective date of the Act and, consistent with its 1976 decision in Franks v. Bowman Transportation Co., Inc., 424 U.S. 747, 96 S.Ct. 1251, 47 L.Ed.2d 444 (1976), awarded retroactive seniority to post-Act discri-minatees. With respect to pre-Act discrimi-natees, however, the Court held that relief was unavailable, and, rejecting the Quarles rationale, found that “an otherwise neutral, legitimate seniority system does not become unlawful under Title VII simply because it may perpetuate pre-Act discrimination.” Teamsters, 431 U.S. at 353-54, 97 S.Ct. at 1864. In finding that the seniority system was bona fide, the Court noted that it “did not have its genesis in racial discrimination, and ... it was negotiated and has been maintained free from any illegal purpose.” Id. at 356, 97 S.Ct. at 1865. After Teamsters, courts presented with the question of whether a seniority system, neutral on its face, is bona fide must determine “whether there has been purposeful discrimination in connection with the establishment or continuation of [the] seniority system.” James v. Stockham Valves & Fitting Co., 559 F.2d 310, 351 (5th Cir.1977), cert. denied, 434 U.S. 1034, 98 S.Ct. 767, 54 L.Ed.2d 781 (1978). This court in James set forth four factors, used by the Supreme Court in Teamsters, relevant to making this determination. The four factors are: (1) whether the seniority system operates to discourage all employees equally from transferring between seniority units; (2) whether the seniority units are in the same or separate bargaining units (3) whether the seniority system had its genesis in racial discrimination; and (4) whether the system was negotiated and has been maintained free of any illegal purpose. James, 559 F.2d at 352. The district court, after applying these four factors to the seniority system as it existed between 1956 and 1963, concluded that the system was bona fide. Plaintiffs do not take issue with the court’s resolution of the first two James factors. Gulf’s seniority system on the effective date of Title VII was facially neutral. In 1963, plant seniority became the basis for all promotions, demotions and layoffs, and the one-day divisional seniority rule was abolished. The two OCAW Locals, which were always in the same bargaining unit, merged in 1963. Therefore, on the effective date of Title VII, there was only one seniority unit and one bargaining unit. Although the district court should have applied the first two James factors to the seniority system as it existed after the effective date of Title VII, instead of to the 1956-1963 system, the court was clearly correct in resolving these factors in favor of Gulf and the OCAW. The gravamen of plaintiffs’ challenge to Gulf’s seniority system is their contention that the system had its genesis in racial discrimination and that it has been maintained, as evidenced by Stipulation 29, with an illegal (discriminatory) purpose. We first examine plaintiffs’ claim that the seniority system had its genesis in racial discrimination. Plaintiffs argue that before 1956 black and white employees performed basically the same type of work within the integrated departments and that the separation of blacks into the Labor Division, and whites into the 0 & M Division, was expressly racial. Plaintiffs also argue that the 1956 one-day divisional seniority rule and diploma and test performance requirements were designed to inhibit blacks from transferring into the 0 & M lines of progression by subjecting blacks to seniority forfeiture upon transfer and by imposing requirements not applied to incumbent whites in the 0 & M Division. Plaintiffs conclude that these considerations show an intent to discriminate, thereby making Gulf’s seniority system non-bona fide. The one-day divisional seniority rule of the 1956 contract effectively operated to penalize blacks who attempted to move into the 0 & M Division. Although the rule was not as egregious as the “seniority suicide” provisions found in Terrell v. United States Pipe & Foundry Co., 644 F.2d 1112, 1118 (5th Cir.1981), vacated, 456 U.S. 955, 102 S.Ct. 2028, 72 L.Ed.2d 479 (1982); United States v. Georgia Power Co., 634 F.2d 929, 931 (5th Cir.1981), vacated, 456 U.S. 952, 102 S.Ct. 2026, 72 L.Ed.2d 477 (1982); Swint v. Pullman-Standard, 624 F.2d 525, 527 (5th Cir.1980), rev’d, 456 U.S. 273, 102 S.Ct. 1781, 72 L.Ed.2d 66 (1982), which required total seniority forfeiture upon transfer, the mere fact that blacks faced with layoff could regain their Labor Division seniority by returning to the Labor Division is insufficient to refute pre-Act discriminatory intent. In addition, the district court’s characterization of Gulf’s diploma and test performance requirements as nondiscriminatory on the ground that they applied to all employees is unconvincing. First, as the district court itself recognized later in its opinion, a test applied to all employees may nevertheless be discriminatory. Second, because these requirements applied only to newly hired employees and employees seeking to transfer into the O & M Division, the court’s finding that they applied equally to all employees is incorrect. Employees in the O & M Division in 1956 (all of whom were white) were not required to have a diploma or take any tests. Although we infer discriminatory intent from Gulf’s pre-Act seniority system, we question whether a seniority system which is neutral as of the effective date of Title VII, which is based on plant seniority, and which has a single bargaining unit could ever be held unlawful solely because of pre-Act discrimination. That the 1963 changes did not rectify the effects of past discrimination, and in fact operated in some ways to lock those effects in, does not imply, in the absence of purposeful discrimination in connection with the post-Act system, that this system was not bona fide under § 703(h). See Teamsters, 431 U.S. at 353, 97 S.Ct. at 1863 (seniority system does not become illegal “simply because it allows the full exercise of the pre-Act seniority rights of employees of a company that discriminates before Title YII was enacted”). The record does not, however, warrant our determination of post-Act bona fides, and we must remand for a district court finding. The problem is Stipulation 29, discussed below, which was implemented in 1967 after Title VII was in effect. The district court must make findings relative to the discriminatory effect of Stipulation 29 and, considering Stipulation 29, the bona fides of the seniority system after 1965. III. Stipulation 29 Plaintiffs contend that the implementation of Stipulation 29, which was an agreement negotiated by Gulf and the OCAW reclassifying certain employees, not only constituted purposeful post-Act discrimination in violation of the fourth James factor, but also had an unlawfully discriminatory impact on black employees under the rationale set forth in Griggs v. Duke Power Co., 401 U.S. 424, 91 S.Ct. 849, 28 L.Ed.2d 158 (1971). Before examining this claim, a review of Stipulation 29 and the 1967 administrative reorganization is in order. A. Background In 1967 the refinery was administratively reorganized, and Gulf and the OCAW implemented Stipulation 29. The combined effect of these two events was to substantially reduce and simplify the lines of progression. In describing these events, we focus on the eleven OCAW maintenance departments because it is the employees in these departments, or who were about to enter these departments, that were affected by Stipulation 29. The structure of the refinery as it existed prior to 1967 has already been described. Before turning to the 1967 changes, one additional fact is necessary. In 1947 a four-year on-the-job mechanical training program had been established to train skilled employees for the number one position in thirteen of the craft departments, eleven of which were the OCAW crafts. To gain entrance into one of the programs, employees in the number two positions in their craft lines of progression were required to bid for trainee job vacancies and pass certain reading comprehension and arithmetic tests. A successful bidder became a “Mechanical Trainee” and upon the completion of the program became a “No. 1 Mechanic” or “Craftsman” in the craft for which he was trained. The 1967 reorganization split the 0 & M Division, which had been comprised of various operations and maintenance departments, into two divisions; an Operating Division to which the operating departments were assigned, and a Maintenance Division to which the maintenance departments were assigned. The eleven OCAW craft departments were assigned to the Maintenance Division. The Labor Division was eliminated, and its subdepartments, which had their own lines of progressions and which had throughout the refinery’s history been assigned to work with the various operations and maintenance departments, were combined with these departments. The mechanical helper pool and certain job classifications in the labor lines of progression were abolished. . Incumbent employees in the mechanical helper pool were reclassified as “Utility Men” and incumbent employees in the eliminated job classifications were reclassified as either “Utility Men” or “Laborers.” The effect of this reorganization on the lines of progression cannot be understood without first examining Stipulation 29. This Stipulation, which was incorporated into the 1967 articles of agreement between Gulf and the OCAW and which became effective on January 1, 1967 with the reorganization, eliminated all job classifications below the mechanical training program (the “Craft Helper” positions) in most of the eleven OCAW departments and moved the employees in these classifications into the mechanical training programs. For these employees, the training programs were reduced from four to two years and the reading and arithmetic tests were suspended in lieu of a simple two-page test which only one employee failed. The new lines of progression can best be illustrated by example. Before the reorganization and Stipulation 29, a new employee would be assigned to the Labor Division as a “Laborer” in one of the labor subdepartment lines of progression. On the basis of plant seniority, the employee would eventually promote to the top classification in his line of progression, which was, in most labor lines of progression, called “Utility Man No. 1.” He would then bid into the mechanical helper pool, which was technically in the 0 & M Division, and then enter one of the craft departments, such as the instrument department. (Alternatively, he could bid directly from his Utility Man position into one of the craft departments without first entering the mechanical helper pool.) Promotion from the Labor Division to the 0 & M Division, and promotions within the 0 & M Division, were based on plant seniority and ability to perform. The entry level position in the instrument line of progression was “Instrument Man No. 3,” and the next highest job classification was “Instrument Man No. 2.” From the number two position the employee would bid into the four-year mechanical training program. The senior bidder was required to pass the reading and arithmetic tests before he could enter the program and become a “Mechanical Trainee.” Upon completion of this program, the employee would become an “Instrument Man No. 1,” which was the top position in the instrument department. The 1967 reorganization, while eliminating the Labor Division, did not eliminate labor jobs. After 1967, a newly hired employee would be assigned to the Maintenance Division as a “Laborer” and would promote directly into the position of “Utility Man.” As a Laborer or Utility Man, he would work with any one or more of the maintenance departments in the Maintenance Division. The fact that Stipulation 29 eliminated the Instrument Man No. 2 and No. 3 positions, by moving these employees into the special two-year mechanical training program, and the reorganization eliminated the mechanical helper pool, by moving these employees into the Utility Man classification, meant that the employee, upon reaching the Utility Man position, would bid directly into the instrument me-chanieal training program. The training program that the employee would bid into was not the special two-year program applicable to reclassified employees, but was the normal four-year program established in 1947. Upon completion of this program, the employee would become an “Instrument Man,” the top position in the instrument department. B. Analysis Plaintiffs contend that the Craft Helpers reclassified as Mechanical Trainees by Stipulation 29 were predominently white employees with less seniority than blacks in certain lower job classifications. Before Stipulation 29 these more senior blacks, plaintiffs argue, were in a position to promote into Craft Helper positions and then, once a vacancy opened in a mechanical training program, bypass whites with less seniority who were already Craft Helpers. Plaintiffs’ statistics show that only nine out of the two hundred and three, or 4.4%, of the Craft Helpers reclassified by Stipulation 29 into the special two-year training programs were black. They contend that this proportion is far lower than the proportion of blacks among senior utility men, the proportion of blacks among senior employees throughout the refinery, or the proportion of blacks among the entire plant workforce. Plaintiffs also claim that Stipulation 29 flooded the top jobs with less senior whites and locked the senior black employees in lower job classifications out of these jobs for the next ten years. Plaintiffs conclude that Stipulation 29 constituted purposeful discrimination by abrogating black employees’ seniority rights and had an unlawfully discriminatory impact upon blacks under the rationale of Griggs v. Duke Power Co., 401 U.S. 424, 91 S.Ct. 849, 28 L.Ed.2d 158 (1971). The district court, while not addressing plaintiffs’ claim that Stipulation 29 constituted purposeful discrimination, held that Stipulation 29 did not constitute a Griggs violation. It appears that the court based its decision on either a finding that the plaintiffs did not establish adverse impact, or on a 1971 conciliation agreement between Gulf and the EEOC which elevated black employees, who had been demoted to the Labor Division while the 1956-1963 one-day divisional seniority rule was in force and thus were absent from Craft Helper positions when Stipulation 29 was implemented, to top positions in the maintenance departments at competitive rates of pay. We decline to base our determination upon the 1971 conciliation agreement because we find that its terms were not coextensive with plaintiffs’ present claim, and instead turn to an examination of plaintiffs’ proof of adverse impact. To establish a prima facie case of adverse impact, plaintiffs must, by a preponderance of the evidence, demonstrate that Stipulation 29 had an adverse impact on blacks at a substantially and disproportionately higher rate than whites. No intent to discriminate need be shown. Teamsters, 431 U.S. at 335 n. 15, 97 S.Ct. at 1854 n. 15. Once established, the burden shifts to the defendants to demonstrate, by a preponderance of the evidence, that legitimate business reasons justify the adoption of Stipulation 29. See Albemarle Paper Co. v. Moody, 422 U.S. 405, 425, 95 S.Ct. 2362, 2375, 45 L.Ed.2d 280 (1975); Griggs, 401 U.S. at 431-32, 91 S.Ct. at 853-54; Page v. U.S. Industries, Inc., 726 F.2d 1038, 1053 (5th Cir.1984). If defendants make the requisite showing, then the burden shifts to plaintiffs to prove that other less discriminatory means were available to achieve defendants’ legitimate business purpose. See Albemarle, 422 U.S. at 425, 95 S.Ct. at 2375. Plaintiffs make two claims of adverse impact in connection with Stipulation 29. First, they claim that the proportion of blacks reclassified (into the special two-year training programs) as Mechanical Trainees from Craft Helper positions under Stipulation 29 was less than the proportion of senior black employees in certain lower job classifications. In response to this claim, the district court held that the employees who were reclassified under Stipulation 29 would have reached the mechanical trainee programs first under the old system, and thus there was no adverse impact. Second, plaintiffs claim that as a result of Stipulation 29, blacks were locked out of top positions (the regular four-year training programs and number one jobs) for the next ten years. Although the district court did not address this contention, the evidence conclusively resolves the issue. Turning to plaintiffs’ first claim, we hold that Stipulation 29 had an adverse impact on blacks at a substantially higher rate than on whites. The record indicates that only 4.4% of employees reclassified as Mechanical Trainees by Stipulation 29 were black and that, due to Gulf’s pre-Act seniority system, 100% of the employees in the Utility Man classification in the Labor Division (in 1966) were blacks with, almost universally, more seniority than whites reclassified as Mechanical Trainees. Before Stipulation 29, these senior blacks enjoyed the possibility of promotion into Craft Helper positions and, from there, they could have bid into one of the four-year mechanical training programs. Because of their greater plant seniority, they would have been senior to many of the whites in the Craft Helper positions and therefore would have had the opportunity to bypass them. The defendants argue that many factors could have prevented the senior blacks from exercising their superior plant seniority to bypass junior whites under the pre-Stipulation 29 structure. Such factors include the employee’s absence, sickness and accident, safety, and disciplinary records, as well as the employee’s failure to fulfill the six-month residency requirement in the job classification below the classification for which the employee was bidding and the employee’s failure to pass a physical examination. This speculation has no effect; plaintiffs could not now prove that senior blacks would have inevitably bypassed junior whites. Plaintiffs need only prove, as they did, that senior blacks were denied the opportunity to bypass junior whites to establish a prima facie case of adverse impact. See Connecticut v. Teal, 457 U.S. 440, 451, 102 S.Ct. 2525, 2532-33, 73 L.Ed.2d 130 (1982) (“[t]he suggestion that disparate impact should be measured only at the bottom line ignores the fact that Title VII guarantees [the plaintiff] the opportunity to compete equally with white workers on the basis of job-related criteria” (emphasis in original)). Turning to plaintiffs’ second claim, we find that as a result of Stipulation 29 blacks were not locked out of top positions for ten years. Between 1957 and 1984, the Port Arthur Refinery experienced a 51% reduction in force, from 5,841 employees in 1957 to 2,872 employees in 1984. During this period, black representation in the workforce increased from 16% to 21%. The number of available mechanical training positions and number one craft jobs increased from 490 in 1966 to 764 in 1967, and remained relatively constant thereafter. Therefore, Stipulation 29 actually increased the number of top jobs. It also had the effect of eliminating many intermediate job classifications, thereby making upward mobility to the top craft jobs easier. Before 1967, the mechanical training programs contained virtually no blacks, however, the percentage of blacks entering the regular training programs between 1967 and 1977 was 21.6%. In 1977, blacks represented 19% of the workforce. While we reject the plaintiffs’ claim that as a result of Stipulation 29 blacks were locked out of the regular four-year training programs and number one craft jobs for ten years, the first claim of adverse impact was established by plaintiffs. In 1967 many blacks lost the opportunity to use their superior seniority to bypass junior whites and as a consequence arrived at top positions at a later time than they would have absent Stipulation 29. Because the district court held that plaintiffs did not establish adverse impact, it did not consider Gulf’s evidence that legitimate business reasons justified the adoption of Stipulation 29. See Albemarle, 422 U.S. 405, 425, 95 S.Ct. 2362, 2375. As noted earlier, the district court also did not address plaintiffs’ claim that Stipulation 29 evidenced purposeful discrimination, the establishment of which is essential to plaintiffs’ claim that the seniority system was not bona fide under § 703(h). The district court must now resolve these issues. IV. Craft Tests Since 1947 when Gulf established the mechanical training programs in the maintenance departments, senior bidders have been required to perform adequately on a battery of tests. The same tests have been administered for each of the mechanical (or “craft”) training programs. From 1947 to 1971 the “old tests” were used to screen bidders, and since 1971 the “new tests” have been administered. Gulf considered test results along with other qualifications to determine whether senior bidders would be awarded trainee positions; no cut-off scores were used. In the district court, plaintiffs contended that both sets of tests had an adverse impact on blacks. Gulf responded that the tests were job related. The court upheld the tests’ validity, finding that while both sets of tests had an adverse impact, Gulf was not required to validate the old tests and Gulf established, through a criteria validation study, that the new tests were job related. On appeal, plaintiffs argue that the new tests were not properly validated and that Gulf is required to establish that the old tests were job related. Section 703(h) authorizes the use of professionally developed scored tests. The Supreme Court has construed this section “to require that employment tests be job related.” Griggs, 401 U.S. at 435-36, 91 S.Ct. at 856. The plaintiff bears the initial burden of establishing a prima facie case of substantial adverse impact by showing that significantly fewer members of plaintiff’s class pass the test compared to their counterparts. Once established, the burden shifts to the defendant to show that the test is “job related.” The plaintiff can rebut the defendant’s evidence by showing that, although the test is job related, alternative selection devices exist which have a comparable business utility, but a lesser adverse impact. See Albemarle, 422 U.S. at 425, 95 S.Ct. at 2375. The district court found that use of the old and new tests had an adverse impact upon blacks. As stated by the court, plaintiffs’ evidence showed that: 82.5% of whites who took the Old Tests between January 1969 and March 1971 ultimately passed. Only 42.8% of blacks who took the same tests during that period ultimately passed. Between 1971 and 1980, 97.7% of the whites who took the New Tests passed them, while only 66% of the blacks who took those tests passed. Gulf does not contest this finding, but argues that it has met its burden of showing that the tests are job related. Before turning to Gulf’s proof, we must first address the burden imposed upon a defendant to show the job relatedness of a discriminatory test. The district court relied on Contreras v. City of Los Angeles, 656 F.2d 1267 (9th Cir.1981), cert. denied, 455 U.S. 1021, 102 S.Ct. 1719, 72 L.Ed.2d 140 (1982), for the proposition that an employer must, by professionally accepted methods, prove that the test is “predictive of or significantly correlated with important elements of work behavior that comprise or are relevant to the job or jobs for which candidates are being evaluated.” Id. at 1280 (the “significantly correlated” standard). This burden departs somewhat from the stricter standard set forth by this Circuit in Watkins v. Scott Paper Co., 530 F.2d 1159 (5th Cir.), cert. denied, 429 U.S. 861, 97 S.Ct. 163, 50 L.Ed.2d 139 (1976), which requires an employer to prove that its selection device was necessary to the business goals of safety and efficiency. Id. at 1168-69 (the “business necessity” standard). Recent Supreme Court cases suggest that the burden to be borne by employers when establishing a defense to discriminatory selection devices is embodied in the “significantly correlated,” rather than the “business necessity,” standard. An excellent discussion of these standards is contained in Contreras, in which the Ninth Circuit concluded that the business necessity language was inconsistent with Supreme Court case law and the congressional intent underlying Title VII. 656 F.2d at 1280. The Contreras court drew its “significantly correlated” language from Albemarle, in which the Supreme Court “clarified” the “appropriate standard of proof for job relatedness,” 422 U.S. at 436, 95 S.Ct. at 2380, by articulating a standard that is “the same as that of the Griggs case — that discriminatory tests are impermissible unless shown, by professionally acceptable methods, to be ‘predictive of or significantly correlated with important elements of work behavior which comprise or are relevant to the job or jobs for which candidates are being evaluated.’ ” Id. at 431, 95 S.Ct. at 2378 (quoting 29 C.F.R. § 1607.4(c)); Contreras, 656 F.2d at 1279. The court in Contreras addressed a division within the Ninth Circuit between courts which adopted the “significantly correlated” standard in Albemarle and a court which adopted the stricter “business necessity” standard set forth in a footnote in Dothard v. Rawlinson, 433 U.S. 321, 332 n. 14, 97 S.Ct. 2720, 2728 n. 14, 53 L.Ed.2d 786 (1977). In Dothard, the Supreme Court held that “the employer must meet ‘the burden of showing that any given requirement [has] ... a manifest relationship to the employment in question.’ ” Id. at 329, 97 S.Ct. at 2727 (quoting Griggs, 401 U.S. at 432, 91 S.Ct. at 854). While the holding is consistent with prior Supreme Court eases, a footnote specified that “a discriminatory employment practice must be shown to be necessary to safe and efficient job performance to survive a Title VII challenge.” 433 U.S. at 332 n. 14, 97 S.Ct. at 2728 n. 14. The Supreme Court has since indicated that the “significantly correlated” standard, rather than the “business necessity” standard, is controlling. See Contreras, 656 F.2d at 1279-80. In New York City Transit Authority v. Beazer, 440 U.S. 568, 99 S.Ct. 1355, 59 L.Ed.2d 587 (1979), plaintiffs challenged a transit authority’s (TA) refusal to hire methadone users. The Court stated: Respondents recognize, and the findings of the District Court establish, that TA’s legitimate employment goals of safety and efficiency require that exclusion of all users of illegal narcotics, barbiturates, and amphetamines, and of a majority of all methadone users. The District Court also held that those goals require the exclusion of all methadone users from the 25% of its positions that are ‘safety sensitive.’ Finally, the District Court noted that those goals are significantly served by — even if they do not require — TA’s rule as it applies to all methadone users including those who are seeking employment in nonsafety-sensi-tive positions. The record thus demonstrates that TA’s rule bears a ‘manifest relationship to the employment in question.’ Griggs v. Duke Power Co., 401 U.S. 424, 432, 91 S.Ct. 849, 854, 28 L.Ed.2d 158. See Albemarle Paper Co. v. Moody, 422 U.S. 405, 425, 95 S.Ct. 2362, 2375, 45 L.Ed.2d 280. Id. at 587 n. 31, 99 S.Ct. at 1366 n. 31 (emphasis added) (citations omitted); see Contreras, 656 F.2d at 1279-80. Fifth Circuit cases decided after Beazer and Contreras have imposed a less strict standard than business necessity upon defendants seeking to show that their selection devices are job related. In Rivera v. City of Wichita Falls, 665 F.2d 531 (5th Cir.1982), we affirmed a district court’s determination that a police test was job related without requiring a further showing that the test was necessary to the safety and efficiency of the police unit. Id. at 536-38. Likewise, in Cormier v. P.P.G. Industries, Inc., 702 F.2d 567 (5th Cir.1983), we affirmed a district court’s decision that the defendant successfully rebutted plaintiff’s prima facie case of adverse impact by showing that its tests had “a manifest and legitimate and business relationship to the jobs for which the tests were used.” Id. at 568. We conclude that the business necessity standard adopted in Watkins has been modified by subsequent Supreme Court decisions, as recognized by Fifth Circuit panels, and hold that discriminatory tests are impermissible unless shown, by professionally accepted methods, to be predictive of or significantly correlated with important elements of work behavior that comprise or are relevant to the job or jobs for which candidates are being evaluated. A. The Job Relatedness of the New Tests In 1969, Gulf hired Richardson, Bellows and Henry (“RBH”), an outside consulting firm of industrial psychologists, to develop the new tests. This firm conducted three criteria validation studies in 1971, 1974 and 1983 to establish the job relatedness of these tests. Criterion related validation is established when there is a significant positive correlation between comparative success on the test (the “predictor”) and comparative success on some measure of job performance. The degree of correlation between test scores and job performance is expressed by a correlation coefficient. The value of the correlation coefficient can range from +1.0 (employees with the highest test scores always perform better on the job) to -1.0 (employees with the highest test scores always perform worse on the job). A coefficient of zero indicates that there is no correlation between test and job performance. Gulf primarily relied upon the 1983 study to establish the validity of its new tests. Although the same tests were administered to senior bidders seeking positions in all of the OCAW craft training programs, the study only yielded correlation coefficients with respect to the pipefitting and boiler-making departments. Because the tests in the Gulf battery were not used separately, the total battery scaled score comprised the predictor. Plaintiffs’ primary objection to the 1983 study centers upon the criteria, called the “performance elements,” used by supervisors to evaluate the job performance of incumbent employees in the pipefitting and boilermaking departments. In developing the criteria, RBH first directed supervisors to rate the relative importance of 117 duties and 37 abilities. The duty elements were intended to provide RBH with a better understanding of the tasks performed in each job classification, while the ability elements were the focus of the criterion development. The specific criteria selected by RBH for inclusion in the rating form were based upon the importance attached to the various ability elements by the supervisors. Plaintiffs contend that the validation study was improperly conducted because the criteria used to evaluate job performance were related to general cognitive abilities and not specific job duties. In other words, plaintiffs argue that the criteria used did not consist of “important elements of work behavior that comprise or are relevant to the job.” See Albemarle, 422 U.S. at 431, 95 S.Ct. at 2378. Therefore, plaintiffs conclude, the correlation coefficients derived from the 1983 study cannot be used to establish a significant relationship between test and job performance. The district court found that the criteria employed by RBH accurately reflected the important elements that comprise the jobs studied. The court noted that “about one-half of [Gulfs] criteria are actual duty elements of the same character of Plaintiffs.” In reaching its conclusion, the court emphasized that the positions for which adequate test performance was a prerequisite required certain cognitive abilities, and that performance on tests which were designed to correlate solely with the duty elements would be a poor predictor of job performance. We find no error in the district court’s determination. Upon successful completion of a training program, an employee promotes to the top position for which he has been trained. Gulfs proof supports the conclusion that the possession of certain cognitive abilities is important to success as a trainee and craftsman. Plaintiffs offered no evidence to show that the exclusive use of duty elements as criteria would have yielded lower correlation coefficients. To the contrary, Gulfs expert testified that higher correlation coefficients would result had RBH selected duty elements as the measure against which job performance was judged. Under these circumstances, we find that the criteria used to evaluate job performance consisted of important elements relevant to the jobs studied. This holding does not necessarily lead to the conclusion that the tests are job related; it only establishes that a proper rating system was used to evaluate job performance. To establish the job relatedness of the tests, the degree of correlation between test scores and job performance ratings must be examined. The district court upheld the validity of the tests without making any findings concerning the sufficiency of correlation. Because a finding of significant correlation between test results and job performance is a prerequisite to a holding that the tests are job related, we assume that the court sub silentio made this finding. Plaintiffs challenge the validity of the tests as they relate to all of the OCAW maintenance departments. The 1983 study yielded correlation coefficients for only the pipefitting and boilermaking departments. The unadjusted coefficients were, for pipefitters .22, and for boilermakers .32. When adjusted for criteria unreliability the coefficients were .23 and .51 respectively, and when adjusted for battery score restriction, the coefficients were .27 and .38 respectively. While Gulfs experts testified that the unadjusted figures underestimated the coefficients’ true values, the record contains little evidence concerning the reliability of the adjusted figures. Gulfs expert, Dr. Hunter, testified that the results of an American Petroleum Institute (API) study, which synthesized forty-two local validation studies performed by companies in the petroleum industry on tests similar to those used by Gulf, indicated that Gulfs tests are job related. Using the methodology of the API study, Dr. Hunter recomputed the results of the 1983 RBH study and derived a correlation coefficient, for maintenance, operations and laboratory technician employees, of .37, From this evidence, it is not possible to determine the basis for the district court’s holding that the tests are job related. We are troubled by the fact that Gulf submitted correlation coefficients for only two maintenance departments and did not attempt to show that the tests were valid with respect to the other OCAW maintenance departments. We therefore remand this cause to the district court for further findings. The court is directed to hold evidentiary hearings as it deems necessary, and to take into consideration the relationship between test scores and job performance in all of the departments which comprise the basis for plaintiffs’ complaint. In determining whether the tests are significantly correlated with job performance, the court should specifically set forth the reasons that justify its determination, including any correlation coefficients upon which it relies and an explanation for its reliance. B. The Old Tests The district court held that, as a matter of law, Gulf was not required to validate the old tests. The court’s analysis of this issue was as follows: So far as the Old Tests are concerned, Gulf was not required to validate them, for the reason that the Old Tests were eliminated four years before Albemarle Paper Co. v. Moody, 422 U.S. 405 [95 S.Ct. 2362, 45 L.Ed.2d 280] (1975), imposed validation requirements; and seven years before The Equal Employment Opportunity Guidelines in Employee Selection Procedures instigated validated studies. We find that while it may not be possible at the present time to conduct a criteria validation study on the old tests, Gulf still bears the burden of establishing that these tests were job related. The old tests were used until 1971, and as early as 1966 Gulf had notice that the administration of non-job related tests was prohibited. The Supreme Court in Griggs invalidated tests which had been in use since 1965. 401 U.S. at 428, 91 S.Ct. at 852. On remand, the district court is directed to determine whether the old tests were job related. Given the impracticality of requiring that a validation study be conducted, Gulf may establish the job relatedness of the old tests through other methods. See e.g. Davis v. City of Dallas, 777 F.2d 205, 221-23 (5th Cir.1985), cert. denied, 476 U.S. 1116, 106 S.Ct. 1972, 90 L.Ed.2d 656 (1986); B. Schlei & P. Grossman, Employment Discrimination Law 14-56 (2d ed. 1983). V. Disparate Treatment Plaintiffs raise two claims of class-wide disparate treatment concerning Gulfs application of its Sickness and Accident (“S & A”) policy and Gulfs selection of supervisors. To establish these claims, plaintiffs must prove that Gulf intentionally discriminated against blacks on a regular basis. Teamsters, 431 U.S. at 366 & n. 16, 97 S.Ct. at 1855 & n. 16. Proof of “more than the mere occurrence of isolated or ‘accidental’ or sporadic discriminatory acts” is required. Id. Statistics establishing a great disparity between Gulfs treatment of blacks and whites may alone justify an inference of discriminatory motive. Payne v. Travenol Laboratories, Inc., 673 F.2d 798, 817 (5th Cir.), cert. denied, 459 U.S. 1038, 103 S.Ct. 451, 74 L.Ed.2d 605 (1982). If statistical evidence is insufficient to establish discriminatory intent, the plaintiffs may introduce historical, individual, or circumstantial evidence. Id. See Walls v. Mississippi State Dep’t of Pub. Welfare, 730 F.2d 306, 321-22 (5th Cir.1984); Page v. U.S. Indus., Inc., 726 F.2d 1038, 1046 (5th Cir.1984); Pouncy v. Prudential Ins. Co. of Am., 668 F.2d 795, 802 (5th Cir.1982). Gulf may rebut plaintiffs’ prima facie case by introducing proof that plaintiffs’ statistics are “inaccurate or insignificant,” Payne, 673 F.2d at 817, or by providing a “non-discriminatory explanation for the apparently discriminatory re-suit,” id. (quoting Teamsters, 431 U.S. at 360 n. 46, 97 S.Ct. at 1867 n. 46). Before turning to a discussion of plaintiffs’ specific allegations, we note that the ultimate determination of the existence of intentional discrimination is a question of fact, reviewed on appeal under the clearly erroneous standard. Anderson v. City of Bessemer City, N.C., 470 U.S. 564, 105 S.Ct. 1504, 1511, 84 L.Ed.2d 518 (1985); Page, 726 F.2d at 1044-45. A. Discriminatory Application of Sickness and Accident Policy An employee’s bid for certain desirable positions, such as supervisory and mechanical trainee positions, triggers an examination of his S & A record to determine whether absences are excessive. A record indicating more than 80 hours of S & A in the preceding year is carefully scrutinized and may result in disqualification. Plaintiffs introduced statistical evidence which purported to show that between 1975 and 1980, 33.1% of all black bidders and 22.0% of all white bidders were rejected because of their unsatisfactory S & A records, and anecdotal testimony from three witnesses which attempted to establish individual acts of discrimination. At trial, plaintiffs argued that Gulf’s policy had an adverse impact on blacks and that Gulf subjected black employees to disparate treatment. The district court held that, “Gulf’s S & A policy was applied equally to blacks and whites. Hence, even if Gulf’s S & A policy resulted in adverse impact on blacks, its use is justified as good business practice under Title VII.” (citation omitted). On appeal, plaintiffs do not advance their adverse impact claim but instead contend only that Gulf intentionally discriminated against blacks in the application of its S & A policy. Gulf’s evidence establishes that the average number of S & A hours for bypassed whites was 188.5 while the corresponding figure for bypassed blacks was 220 hours. These numbers tend to show that Gulf did n