Plaintiffs-appellants Ann Brunet and Denise Sachs, female firefighters of the Columbus Division of Fire (the “CDF”), and the class they represent (the “Brunet plaintiffs”), appeal three separate orders of the District Court involving the CDF’s hiring of entry level firefighters. First, they appeal the District Court’s order of March 18, 1992, setting aside a February 27, 1989 consent decree between the Brunet plaintiffs and defendant-appellee City of Columbus (the “City”) and enjoining the City from selecting firefighters *393in the manner provided by the consent decree. Second, they appeal the District Court’s order of July 24, 1992, holding that strict rank-order hiring on the basis of an applicant’s total score on the firefighter examination is valid and may be used to select firefighters. Third, the Brunet plaintiffs appeal the District Court’s final judgment entered September 24,1992, which incorporates its order of September 23,1992, reducing the seniority of two female firefighters. These appeals result from a cla'ss action by Stuart Tudor, James Meyer, Joseph Hilleary and Guy Tucker' (the “Tucker plaintiffs”), male CDF firefighters or candidates, against the City1 alleging that the City’s hiring - of firefighters in accordance with the consent decree of February 27,1989, violates the Equal Protection Clause of the Fourteenth Amendment to the United States Constitution by discriminating against male applicants in the selection of entry level firefighters. The District Court consolidated the two class actions.
On appeal, the issues are (1) whether the District Court erred in determining that the Tucker plaintiffs had standing to pursue their action; (2) whether the District Court erred in holding in May, 1986 that the Brunet plaintiffs failed to. prove intentional discrimination in their claim brought under 42 U.S.C. § 1983; (3) whether the District Court erred in holding that the method of hiring entry level firefighters provided for by the consent decree violated the Tucker plaintiffs’ Equal Protection guarantees not to be discriminated against on the basis -of their gender; (4) whether the District Court erred in finding that the City discriminated against the Tucker plaintiffs by agreeing to the February 27,1989 consent decree bn the grounds that no such allegations were made by the Tucker plaintiffs in their complaint and that that claim is outside the applicable statute of limitations; (5) whether the District Court erred in determining that strict rank-order hiring on the basis of a candidate’s total score on the 1986 et seq. firefighter examination is a valid, job-related method of choosing qualified candidates for the position of firefighter; (6) whether the District Court erred in determining that the Brunet plaintiffs had failed to show the existence of an alternative selection device that has a less adverse impact on women; and (7) whether the District Court erred in holding that an appropriate remedy for the Tucker class of plaintiffs was to alter the seniority dates of two female firefighters. For the reasons stated below, we affirm in part, reverse in part and remand for further proceedings.
I.
A.
Brunet Litigation
The background of this case is lengthy. Plaintiffs Ann Brunet and Denise Sachs are currently CDF firefighters. In 1980 and 1984, Brunet took the City’s firefighter examination and was not selected as a firefighter. In 1984, Brunet, together with three other plaintiffs, brought a class action against the City2 alleging that the City’s entry level firefighter examination discriminated against women. The District Court in that case certified a class of all past, present, and future female firefighter candidates.
The 1980 and 1984 firefighter examination consisted of a physical capability test (“PCT”) and a written examination called a cognitive ability test (“CAT”), which included a mechanical reasoning test. The CAT constituted 70 percent of the applicant’s total score. The PCT constituted 30 percent of the applicant’s total score. Each candidate was ranked on the basis of his or her total score on the examination. Candidates who' had taken the 1975 and 1978 examinations were ranked on the basis of their scores on the CAT alone. The PCT was administered on a pass/fail basis. The District Court in Brunet determined that this change in the *394ranking procedure occurred as a result of a Civil Service Commission study conducted by the City that found that “the work of firefighting was largely physical, and that better firefighters were distinguished by the ability to excel while performing physical tasks.” Brunet v. City of Columbus, 642 F.Supp. 1214, 1236 (S.D.Ohio 1986), appeal dismissed, 826 F.2d 1062 (6th Cir.1987), cert. denied, 485 U.S. 1034, 108 S.Ct. 1593, 99 L.Ed.2d 908 (1988). Pursuant to a court order in Dozier v. Chupka, 395 F.Supp. 836 (S.D.Ohio 1975), applicants who took the 1980 and 1984 examination were ranked in the order of their total score on separate eligibility lists for white and black applicants. The purpose of these separate lists was to increase the percentage of black firefighters in the CDF.
In their complaint, the Brunet plaintiffs challenged the 1980 and 1984 PCT and the mechanical reasoning portion of the CAT as having a disparate impact on female candidates and as not being job related in violation of Title VII of the Civil Rights Act of 1964, 42 U.S.C. § 2000e et seq. Plaintiffs further contended in a 42 U.S.C. § 1983 claim, that the City in 1980 and 1984 acted "with intentional discrimination against female firefighter candidates in violation of the Equal Protection Clause of the Fourteenth Amendment.
The District Court’s decision in Brunet consisted of an opinion and order entered on May 13, 1986, and a supplemental opinion and order entered on May 30, 1986. The District Court held that the 1980 PCT and the 1980 and 1984 mechanical reasoning tests did not violate Title VII because they had not had a disparate impact on the class of female candidates. Brunet, 642 F.Supp. at 1221. The District Court also held that the plaintiffs failed to prove intentional discrimination in connection with their Equal Protection claim brought under section 1983: Id. at 1223. However, the District Court held that the 1984 PCT did violate Title VII because the 1984 PCT had a disparate impact on female candidates, and the City had failed to demonstrate that the PCT, as it was then designed, was job related. Id. at 1247-50. The District Court also held that the City’s practice of rank-order hiring from a single list grouping together males and females was impermissible under Title VII because the City could not establish that higher scores on the test meant better job performance. Id. at 1248-49.
The Brunet Court enjoined the City from hiring any new firefighters until a new PCT could be validated for the 1986 examination. Id. at 1253. It also enjoined the use of rank-order selection from a single list comprised of female and male candidates until strict rank-order hiring could also be validated. Id. at 1252-53. Until validation, the City was required to establish a pass/fail scoring system for the PCT based upon the scores of incumbent firefighters and to hire men and women in proportion to the number of each sex passing the 1986 examination. , Id. at 1253. Finally,- the District Court held that upon adoption of a content valid examination, the City was required to offer the new test to the 1984 female applicants and hire the number of women who would have been hired in 1984 had the test been content valid. Id.
However, on July 14, 1986, before the 1986 test was content validated, the District Court entered an order, also reported at 642 F.Supp. 1214, in which it modified its injunction to allow the City to hire men and women firefighters in proportion to the number of males and females who passed the 1986 test and to select from within either category, either by rank order or on a random basis. Id. at 1254-58. This action was necessary to satisfy the City’s urgent need for new firefighter recruits. Plaintiffs appealed and the City cross-appealed.
While the appeals were pending, the City developed a new 1986 PCT and submitted a content-validity study. Stated simply, a content-valid test identifies those skills important to a particular job and tests those skills. On May 21,1987, the District Court held that the new 1986 PCT was not content valid because it included a task that had an adverse impact on women candidates that was not critical to the performance of a firefighter, namely, the “hose hoist event.” The District Court determined that absent this “noncritical task,” the 1986 PCT was content valid. It then approved 'the use of the PCT without this task in selecting future firefight*395er elasses. Id. It also directed the City to continue to hire male and female firefighters in proportion to their numbers in the pool of candidates that passed the firefighter examination.
Moreover, the District Court established a minimum score or “cut” score that all persons taking the examination would have to achieve in order to pass the examination and be eligible for hiring. The cut score was designed to ensure that all persons on the lists of eligible candidates for firefighter were qualified for the position. The District Court found that under this cut score, 15 percent of the incumbent firefighters would fail the PCT. In an order issued on June 12, 1987, the District Court found that had a valid examination been given, a total of four women would have been hired from the 1984 examination, and, noting that the City had already hired three women from that class, it ordered the City to hire one more woman.
Subsequently, this Court dismissed all pending appeals and cross-appeals on the May 13 and 30,1986 orders on the basis that the appeals were not from final judgments and were moot to the extent that the City had complied fully with the District Court’s injunctions. The parties then appealed and cross-appealed from the May 21, 1987 order. While that appeal was pending, the City filed with the District Court a detailed study called a concurrent criterion-related validity study prepared by an expert in industrial psychology, Dr. Frank J. Landy, which purported to validate strict rank-order hiring. No hearing regarding this study was held at this time.
On September 7, 1988, the Brunet plaintiffs filed a motion for contempt against the City alleging that the City had administered the hose hoist event as part of the PCT. While the appeals from the May 21, 1987 order and the contempt motion were pending, the City initiated settlement discussions with the Brunet plaintiffs that culminated in the consent decree of February 27, 1989.
Paragraph 12 of the consent decree provides:
The City and the Safety Director shall continue to appoint male and female applicants in proportion to the relative proportions of males and females achieving passing scores on the firefighter entry level test as a whole, for a period of at least 20 years from the date of this Order. The number to be hired if a fraction shall be rounded up to the next whole number.
The decree also provided that both the CAT and the PCT would be scored on a pass/fail basis and that the minimum “cut” score required by the District Court would be used to determine which candidates had passed the entire examination and were eligible for hiring. Consent decree, ¶¶ 2, 11. The City also agreed to withdraw its concurrent criterion-related validity study on rank-order hiring, although the District Court had neither examined nor ruled on the study. This consent decree was entered into by the parties and approved by the District Court without a public hearing.
B.
Tucker Litigation
The Tucker plaintiffs sat for the Spring 1990 firefighter examination. On July 26, 1991, Tudor, Meyer and Tucker filed a complaint raising a section 1983 challenge to the hiring pursuant to the consent decree on equal protection grounds. The complaint also requested class certification on behalf of all similarly-situated males. The complaint was amended on August 27, 1991, to, inter alia, add Hilleary as a named plaintiff. On March 18, 1992, the District Court held that the consent decree discriminated against the Tucker plaintiffs on the basis of their gender in violation of the Equal Protection Clause. Consequently, the District Court set aside the consent decree and enjoined the City from selecting firefighters in the manner provided by the consent decree. A timely appeal followed. However, the District Court reserved for future consideration the validity of rank-order hiring, i.e., whether there is a linear relationship between test scores on the PCT and job performance.
Subsequently, the District Court entertained further evidence in a hearing on May 21, 1992, on the issue of strict rank-order hiring. To support its argument that strict rank-order hiring based on the current fire*396fighter examination is valid, the City resubmitted the concurrent criterion-related validity study prepared by Dr. Landy. Based on this study, the District Court determined in its memorandum opinion of July 24, 1992, that a higher score on the PCT is likely to result in better job performance, and that therefore the test results could be used to select new firefighters in order of their PCT scores. The court also concluded that there existed no comparable test that had a lesser disparate impact on female candidates. It also modified its May 21, 1987 preliminary order to allow the City to hire firefighters in strict rank order from a single list comprised of males and females. The court further directed that the ranking be based on the passing candidates’ combined scores on the PCT and the CAT, giving equal weight to each. The Brunet plaintiffs also timely appealed this order.
Thereafter, on September 23, 1992, the District Court ordered that all male applicants hired from the 1990 test receive seniority ahead of the two female firefighters, Fox and Sachs, who had been hired from the 1990 examination. The District Court based this decision on the fact that both of the two female firefighters’ test scores were lower than the scores of all the males hired from the 1990 examination. On September 24, 1992, the District Court entered a final judgment in favor of the Tucker plaintiffs. On October 19,1992, the Brunet plaintiffs filed a notice appealing from the District Court’s final judgment of September 24, 1992, and from the District Court’s orders of May 13, 1986, May 30,1986, and May 21,1987. All of the appeals were consolidated.
II. STANDING
A.
The Brunet plaintiffs challenge the Tucker plaintiffs’ standing to bring their action and to act as class representatives. If they lacked standing, then the District Court had no case or controversy over which it, or we, may exercise federal jurisdiction. “The constitutional power of federal courts cannot be defined, and indeed has no substance, without reference to the necessity ‘to adjudge the legal rights of litigants in actual controversies.’ ” Valley Forge Christian College v. Americans United for Separation of Church & State, Inc., 454 U.S. 464, 471, 102 S.Ct. 752, 757-58, 70 L.Ed.2d 700 (1982) (quoting Liverpool S.S. Co. v. Commissioners of Emigration, 113 U.S. 33, 39, 5 S.Ct. 352, 355, 28 L.Ed. 899 (1885)). The doctrine of standing serves to identify those disputes that satisfy the case-or-controversy requirement of Article III. The party invoking federal jurisdiction bears the burden of establishing standing to sue. Warth v. Seldin, 422 U.S. 490, 508, 95 S.Ct. 2197, 2210, 45 L.Ed.2d 343 (1975).
[3] First, the plaintiff must have suffered an “injury in fact.” The Supreme Court has made clear that the injury must be “concrete and particularized” and “actual or imminent, not conjectural or hypothetical.” Lujan v. Defenders of Wildlife, — U.S. -, -, 112 S.Ct. 2130, 2136, 119 L.Ed.2d 351 (1992). If the injury is not actual, but imminent, the plaintiff cannot simply allege possible injury at some indefinite, future time. Whitmore v. Arkansas, 495 U.S. 149, 158, 110 S.Ct. 1717, 1724, 109 L.Ed.2d 135 (1990). The Supreme Court acknowledges that “ ‘imminence’ is concededly a somewhat elastic concept,” but stresses that the purpose of the requirement “is to insure that the alleged injury is not too speculative for Article III purposes—that the injury is ‘certainly impending.’ ” Lujan, — U.S. at - n. 2, 112 S.Ct. at 2139 n. 2 (quoting Whitmore, 495 U.S. at 158, 110 S.Ct. at 1724). Next, the litigant must show that the injury is “fairly ... trace[able] to the challenged action” and “is likely to be redressed by a favorable decision.” Simon v. Eastern Ky. Welfare Rights Org., 426 U.S. 26, 41, 38, 96 S.Ct. 1917, 1926, 1924, 48 L.Ed.2d 450 (1976).
In Regents of the University of California v. Bakke, 438 U.S. 265, 98 S.Ct. 2733, 57 L.Ed.2d 750 (1978), a non-minority medical school applicant challenged the University’s affirmative action program. The Court found that Bakke did not have to show that in the absence of the program he would have been admitted to medical school. The Court found injury “in the University’s decision not to permit Bakke to compete for all 100 places *397in the class, simply because of his race.” Id. at 281 n. 14, 98 S.Ct. at 2743 n. 14. Thus, in the context of affirmative action programs, the challenger need only show that, but for the program, he would have been considered for the job, to satisfy standing requirements. The United States Supreme Court recently reaffirmed this principle in Northeastern Florida Chapter of the Associated General Contractors of America v. City of Jacksonville, — U.S. -, -, 113 S.Ct. 2297, 2303, 124 L.Ed.2d 586 (1993):
When the government erects a barrier that makes it more difficult for members of one group to obtain a benefit than it is for members of another group, a member of the former group seeking to challenge the barrier need not allege that he would have obtained the benefit but for the barrier in order to establish standing. The “injury in fact” in an equal protection ease of this variety is the denial of equal treatment resulting from the imposition of the barrier, not the ultimate inability to obtain the benefit.
The Brunet plaintiffs argue that none of the named plaintiffs in the Tucker action have standing to challenge the consent decree because the decree caused no harm to any of the named Tucker plaintiffs and because any ease or controversy that existed was mooted before class certification. Further, they argue that any threatened future injury was not imminent, but rather was speculative at the time the court ruled on the certification issue.
The Columbus City Charter requires the City to hire entry level firefighters according to the “Rule of Three.” Under the Rule of Three, the Columbus Civil Service Commission presents the Director of Public Safety the three highest ranking candidates'for each position from a certified list; the Director may choose any one of the three. In addition to passing the firefighter examination, eligibility is also dependent upon the applicant successfully passing a background investigation, physical examination and polygraph test. An applicant who passes these additional requirements is then, ranked on a certified list. Not all applicants passing the civil service exam appear on the certified list.
In order to hire in proportion to the males and females who passed the test, the City maintained separate lists for males and females. The female and male candidates who passed the examination were then ranked on their respective lists according to their scores on the CAT and the PCT.3 However, for informative purposes, one overall list that grouped males and females together and ranked each candidate on the basis of his or her score was also maintained. On the overall list for the Spring 1990 examination, Tudor ranked 108, Meyer ranked 110, Hilleary ranked 120, and Tucker ranked 176. The first firefighter class hired from' the 1990 eligibility list was the March, 1991 class. By operation of the consent decree, this class consisted of 33 men chosen from the top 50 men appearing‘on the certified list and 3 women. Due to their placement on the overall eligibility list, none of the named plaintiffs were eligible for consideration or entry in this class.
In the Summer of 1991, the City prepared to hire another class of 36 firefighters from the 1990 examination list. Pursuant to the consent decree, this class was to consist of the top 34.ranked and qualified men and 2 women. The named plaintiffs were ranked as follows on the certified list for the August, 1991 class:
Tudor: 36
Meyer: 38
Hilleary: 44
Tucker: Not included on this list.
In the absence of the consent decree, 36 men would have been hired for this class. On the *398overall list, with one exception,4 no female candidate ranked higher than any of the Tucker plaintiffs.- Hence, in the absence of the consent decree, and in light of the Rule of Three, both Tudor and Meyer would have been considered for a position, and Tudor would probably have been selected.5
After receiving notice of the Tucker plaintiffs’ intent to sue the City, the City Attorney made an offer to Tudor and Meyer to expand the August, 1991 class from 36 -to 38 firefighters in order to include them both. Tudor and Meyer rejected the offer and filed their complaint and a motion for a temporary restraining order challenging the constitutionality of the consent decree the next day. The District Court granted the Tucker plaintiffs’ motion to intervene in the Brunet proceedings and consolidated the two eases.
On July 30, 1991, the District Court granted a temporary restraining order enjoining the operation of the consent decree. The Brunet plaintiffs appealed the District Court’s grant of the temporary restraining order and sought an emergency stay from the District Court, which it denied on August 15, 1991. The Brunet plaintiffs then sought an emergency order from this Court, which stayed the District Court’s temporary restraining order on August 16, 1991. Additionally, we ordered that, pursuant to the City’s offer, the class that would begin training on Monday, August 19, 1991, would be expanded from 36 to-38 firefighters and include Tudor, Meyer, Vicki Fox, and Denise Sachs.
On August 17, 1991, the Tucker plaintiffs sought and received an ex parte order from the District Court temporarily restraining the City from proceeding with its training class on August 19, 1991, until August 20, 1991, in order that this Court might have an opportunity to reconsider the August 16, 1991 order. On August 19, 1991, this Court granted a second emergency stay pursuant to-the Brunet plaintiffs’ motion, and the class of firefighters scheduled to begin training on August 19, 1991, commenced on August 20, 1991, with Sachs, Fox, Meyer, and Tudor in attendance. Thereafter,. this Court dismissed the Brunet plaintiffs’ appeal from the temporary restraining order as moot in light of the commencement of the hiring class.6
The Tucker plaintiffs filed their motion to certify a class on September 19, 1991. On November 25, 1991, the District Court certified the Tucker plaintiffs’ action as a class action pursuant to Federal Rule of Civil Procedure 23(b)(2). The class certified consisted of all males who desired to be in the August, 1991 training class and in future training classes for entry-level firefighters but who would be precluded from such positions or would lose seniority solely on the basis of gender as a result of the consent decree in the Brunet case. In the same order, the District Court addressed the issue of standing. While we agree with the District Court that the Tucker plaintiffs have standing to bring this action, our reasons are other than those given by the District Court.
B. Hilleary and Tucker
The District Court determined that the controversy between Hilleary and Tucker and the City was “very much alive” because “[a]t this point, they have not been hired as firefighters, and they may well be affected in the future by the operation of the 1989 order.” To the extent that the court was referring to their being hired or not, this determination was in error. Tucker was not on the certified list for the August, 1991 class. Thus at this time, Tucker was not eligible for consideration. Moreover, Tucker’s low rank would have eliminated him from consider*399ation for the August, 1991 class. The record is silent as to whether he would appear on the next certified list. Any future injury to him was merely speculative and does not satisfy the Supreme Court’s “imminence” requirement of the standing doctrine in cases of future injury. Lujan, — U.S. at -, 112 S.Ct. at 2136.
Similarly, the consent decree did not adversely affect Hilleary’s selection to the August, 1991 training class. The Tucker action challenged the operation of the consent decree from August 1991 forward.7 Assuming arguendo that the City would have still hired a class of 38 even in absence of the consent decree, the facts show that the last candidate hired was Meyer, number 38. Under the Rule of Three, the City still would have only reached candidate 40 in its consideration for the 38th position in the class; Hilleary ranked 44. Moreover, looking to the facts known to the court at the time it was ruling on the certification motion, unless Hilleary was eliminated for another reason, he would be hired in the next class regardless of the consent decree.
C. Tudor and Meyer
The same was true of Tudor and Meyer, ie., their date of hire was not affected by the consent decree because they would have been hired in the August, 1991 class whether it was in effect or not. In class actions, standing must exist both at the time the complaint is filed and at the time the class was certified. Sosna v. Iowa, 419 U.S. 393, 403, 95 S.Ct. 553, 559, 42 L.Ed.2d 532 (1975). At the time the complaint was filed, July 26, 1991, it appeared the operation of the consent decree would prevent Tudor and Meyer from being considered for admission into the August, • 1991 class because of their gender. Thus, at the time of the complaint, Tudor and Meyer faced imminent injury and had standing to bring the lawsuit.
“Settlement of á plaintiff s claims moots an action.” Lusardi v. Xerox Corp., 975 F.2d 964, 974 (3d Cir.1992) (citing Lake Coal Co. v. Roberts & Schaefer Co., 474 U.S. 120, 106 S.Ct. 553, 88 L.Ed.2d 418 (1985); Hammond Clock Co. v. Schiff, 293 U.S. 529, 55 S.Ct. 146, 79 L.Ed. 639 (1934); 13A C. Wright, A. Miller & E. Cooper, Federal Practice and Procedure, § 3533.2, at 233 (1984)). When Tudor and Meyer entered the August, 1991 class, their claims that they should have been hired became moot. Admittedly, special mootness rules exist for class actions. Once a class is certified, the mooting of the named plaintiffs claim does not moot the action, the court continues to have jurisdiction to hear the merits of the action if a controversy between any class member and the defendant exists. Sosna, 419 U.S. at 399, 95 S.Ct. at 557. Where, on the other hand, the named plaintiffs claim becomes moot before certification, dismissal of the action is required. Board of School Comm’rs v. Jacobs, 420 U.S. 128, 95 S.Ct. 848, 43 L.Ed.2d 74 (1975) (per curiam); Lusardi, 975 F.2d at 974; Tucker v. Phyfer, 819 F.2d 1030, 1035 (11th Cir.1987); Sannon v. United States, 631 F.2d 1247, 1252 (5th Cir. 1980). Since Tudor and Meyer entered the class before even filing the certification motion, neither Tudor nor Meyer had standing to contest the hiring procedures at the time of certification.
*400The District Court relied on Deposit Guaranty National Bank v. Roper, 445 U.S. 326, 100 S.Ct. 1166, 63 L.Ed.2d 427 (1980), and United States Parole Commission v. Geraghty, 445 U.S. - 388, 100 S.Ct. 1202, 63 L.Ed.2d 479 (1980), to find that Tudor and Meyer have standing in this action. However, the holdings in these cases are limited to the question of a proposed class representative’s right to appeal the denial of class certification and are inapplicable to the present case.
We merely hold that when a District Court erroneously denies a procedural motion, which, if correctly decided, would have prevented the action from becoming moot, an appeal lies from the denial and the corrected ruling “relates back” to the date of the original denial.... The “relation back” principle, a traditional equitable doctrine applied to class certification claims in Gerstein v. Pugh, 420 U.S. 103 [95 S.Ct. 854, 43 L.Ed.2d 54] (1975), serves logically to distinguish this ease from the one brought a day after the [named plaintiffs claim on the merits expires].... If the named plaintiff has no personal stake in the outcome at the time class certification is denied, relation back of appellate reversal of that denial still would not prevent mootness of the action.
Geraghty, 445 U.S. at 406-07 n. 11, 100 S.Ct. at 1214 n. 11. In both Roper and Geraghty, the class representatives had a personal stake in the controversy at the time the court certified the action. Subsequent to the certification, their claims became moot.
In contrast, Tudor and Meyer had no personal stake in the hiring procedures at the time the District Court certified the class. We do not read Roper and Geraghty as doing away with the requirement that the proposed class representative have standing at the time of class certification.
The District Court also relied on Roper for the proposition that the City cannot “buy off’ the named plaintiffs’ claims in order to' avoid certification. Again, however, Roper concerns the “voluntary cessation” issue in the limited context of the attempt to moot a class representative’s claims, after certification is denied, that operates to prevent appellate review of the denial of that certification. Additionally, unlike in Roper, Meyer and Tudor 'ultimately accepted the City’s settlement offer by entering the August, 1991 class. Thus, both before the motion to certify was even filed and at the time the court certified the class, neither Meyer nor Tudor had a personal stake in being hired. Moreover, the Roper Court emphasized the importance of the factual context in which the case arose. In Roper, the plaintiffs never accepted the defendant’s settlement offer as satisfaction of their substantive claims. The Court found that “[n]either the rejected tender nor the dismissal of the action over plaintiffs’ objection mooted the plaintiffs’ claim on the merits so long as they retained an economic interest in class cértification.” Roper, 445 U.S. at 333, 100 S.Ct. at 1171. In the present case, Tudor and Meyer’s claims were not mooted by the City’s settlement offer or by this Court’s order to expand the class. Their hiring claims were mooted, however, when Tudor and Meyer entered the August, 1991 class.
Some courts have held that a case does not become moot where a defendant “picks off’ the claims of named plaintiffs with settlement offers in an attempt to avoid a class action. “If a tender made to the individual plaintiff while the motion for certification is pending could prevent the courts from ever reaching the class action issues, that opportunity is at the mercy of a defendant, even in cases where a class action would be most clearly appropriate.” Susman v. Lincoln Am. Corp., 587 F.2d 866, 870 (7th Cir.1978), cert. denied, 445 U.S. 942, 100 S.Ct. 1336, 63 L.Ed.2d 775 (1980). However, these cases are limited to the situation where “a motion for class certification has been pursued .with reasonable diligence and is then pending before the district court.” Id.; see also Zeidman v. J. Ray McDermott & Co., 651 F.2d 1030, 1051 (5th Cir.1981). In the present case, there was no motion for certification pending when the City made its settlement offer. Moreover, Tudor and Meyer accepted the offer before a motion was even filed. We conclude that it was error for the District Court to have based the standing of Tudor *401and Meyer on the Geraghty and Roper decisions.
D. Loss of Seniority Rights
Plaintiffs Tudor, Meyer, and Hilleary did, however, sustain an actual “concrete and particularized” injury in their loss of seniority rights to female firefighters hired pursuant to the consent decree.
Seniority systems and the entitlements conferred by credits earned thereunder are of vast and increasing importance in the economic employment system of this Nation- Seniority principles are increasingly used to allocate entitlements to scarce benefits among competing employees (“competitive status” seniority) and to compute noncompetitive benefits earned under the contract of employment (“benefit” seniority).... We have already said about “competitive status” seniority that it “has become of overriding importance, and one of its major functions is to determine who gets or who keeps an available job.” Humphrey v. Moore, 375 U.S. 335, 346-347 [84 S.Ct. 363, 370, 11 L.Ed.2d 370] (1964). “More than any other provision of the collective[-bargaining] agreement ... seniority affects the economic security of the individual employee covered by its terms.” Aaron, Reflections on the Legal Nature and Enforceability of Seniority Rights, 75 Harv.L.Rev. 1532, 1535 (1962).
Franks v. Bowman Transp. Co., 424 U.S. 747, 766, 96 S.Ct. 1251, 1265, 47 L.Ed.2d 444 (1976).
Seniority within the [Columbus] Division of Fire is determined by date of appointment. This is important in any firefighter’s career for a number of reasons including pay, pensions, vacation time, “Kelly Day” (or days-off work) selection, and promotions. Seniority points are awarded for years of service, which are then added to a firefighter’s competitive score on promotion examinations; in case of a tie or close finish the more senior person thus obtains the promotions.
1991 Joint Stipulations, ¶ 10. It is true that Tudor, Meyer, and Hilleary8 were hired on the same date they would have been hired whether or not the consent decree was in effect. Therefore, their “benefit seniority” rights were not affected by the decree, and they are receiving the same pay, pension contributions and vacation time as they would have -if the women had not been hired.
The Tucker plaintiffs’ injury lies in their loss of “competitive status” seniority rights. The women hired in the August, 1991 class achieved a higher place on the CDF’s seniority hierarchy than they would have in absence of the consent decree. Fox and Sachs would be less senior than Tudor, Meyer and Hil-leary if not for the consent decree. Hiring Fox and Sachs, whose test scores were lower than those of any of the Tucker plaintiffs, diminished the employment status of Tudor, Meyer, 'and Hilleary. This is because Fox and Sachs enjoy the same “competitive status” seniority rights as Tudor and Meyer, and greater rights than Hilleary. Tudor and Meyer have lost the advantage of having more seniority points than the women in a future competition for a promotion. They are harmed by having to compete on equal footing with persons who do not belong in their class. Hilleary has not only lost that advantage, but has been put at a disadvam tage because he is less senior than Fox and Sachs. Moreover, when the City expanded the class from 36 to 38, Tudor and Meyer were put in the position of having to compete with two additional persons, and Hilleary, who was hired in the next class, was junior to two more people than he would have been in absence of the consent decree. We hold that the diminution in employment status suffered by the Tucker plaintiffs is a legally cognizable injury for purposes of standing to challenge the City’s hiring practices.
In Lorance v. AT & T Technologies, Inc., 490 U.S. 900, 109 S.Ct. 2261, 104 L.Ed.2d 961 (1989), AT & T adopted a facially non-discriminatory seniority system for an allegedly discriminatory purpose. The allegedly discriminatory effect of the new seniority sys*402tem was not felt until a few years after its adoption, when female employees, who had lost seniority under the new system, were demoted. The Court faced the issue of “when the limitations period begins to run in a lawsuit arising out of a seniority system not alleged to be discriminatory on its face or as presently applied.” Id. at 903, 109 S.Ct. at 2264. The Supreme Court held that the women were time barred from bringing a Title VII action to challenge the adoption of the system. The Court further held that “when a seniority system is nondiscriminatory in form and application, it is the allegedly discriminatory adoption which triggers the limitations period.” Id. at 911, 109 S.Ct. at 2268-2269 (emphasis in original). In other words, the cause of action accrues upon the happening of the discriminatory act, even if the effects of the discriminatory act are not felt until a later time. See Northeastern Fla. Chapter of the Associated Gen. Contractors of Am. v. City of Jacksonville, — U.S. -, -, 113 S.Ct. 2297, 2303, 124 L.Ed.2d 586 (1993) (“The ‘injury in fact’ in an equal protection case of this variety [ (city contract set-aside program)] is the denial of equal treatment resulting from the imposition of the barrier, not the ultimate inability to obtain the benefit.”); Delaware State College v. Ricks, 449 U.S. 250, 258, 101 S.Ct. 498, 504, 66 L.Ed.2d 431 (1980) (“proper focus is upon the time of the discriminatory acts, not upon the time at which the consequences of the acts became most painful”) (emphasis in original) (quoting Abramson v. University of Haw., 594 F.2d 202, 209 (9th Cir.1979)). Thus, under Lorance, the Tucker plaintiffs do not have to wait until they lose out on a promotion to a woman that they would have received if the woman’s date of hire had been later than theirs.
The Civil Rights Act of 1991, Pub.Law No. 102-166, 105 Stat. 1071 (1991) (the “Act”), provides additional support for this conclusion. November 21, 1991, was the effective date of the Act and thus, it was in effect on November 25, 1991, the date the class was certified. The Act provides in relevant part:
[A]n unlawful employment practice occurs, with respect to a seniority system that has been adopted for an intentionally discriminatory purpose in violation of this subchap-ter (whether or not that discriminatory purpose is apparent on the face of the seniority provision), when the seniority system is adopted, when an individual becomes subject to the seniority system, or when a person aggrieved is injured by the application of the seniority system or provision of the system.
42 U.S.C.A. § 2000e-5(e)(2) (West Supp. 1992). In response to Lorance, Congress overruled the Court’s holding that the adoption of the challenged seniority system is the sole triggering event. Section 2000e-5(e)(2) provides for the accrual of a cause of action when the consequences of the discriminatory act are felt by the plaintiff; it did not alter that portion of the decision recognizing that a cause of action accrues at the time of the discriminatory act. Thus, we hold that when the discriminatory act occurred, ie., when the decision was made to hire two women in the August, 1991 class, pursuant to the consent decree, Tudor, Meyer and Hilleary were injured and, therefore, have standing to sue.
III.
The Brunet plaintiffs argue that the Tucker plaintiffs never made a claim in their original or amended complaints for relief based upon the City entering into the February 27, 1989 consent decree and that consequently, the District Court erred in determining that the City had intentionally discriminated against male applicants by entering into the consent decree. This argument is meritless. See Tucker plaintiffs amended complaint, ¶20. The consent decree forms the basis of the Tucker plaintiffs action, and the Brunet plaintiffs should have been aware of this during the proceedings below. Because this issue was not raised before the District Court, it is considered abandoned on appeal and is not reviewable,, See Boyd v. Ford Motor Co., 948 F.2d 283, 284 (6th Cir. 1991), cert. denied, — U.S. -, 112 S.Ct. 1481, 117 L.Ed.2d 624 (1992); First State Bank v. City & County Bank, 872 F.2d 707, 715 n. 5 (6th Cir.1989). The Brunet plaintiffs also argue that any claim regarding the consent decree is barred by the applicable two-year statute of limitations. This argument is an affirmative defense and must be pled. *403See Fed.R.Civ.P. 8(e). Having raised this argument for the first time in this appeal, the Brunet plaintiffs have waived this affirmative defense. See id.; Great Southwest Life Ins. Co. v. Frazier, 860 F.2d 896, 903 (9th Cir.1988). See also Hoover v. Langston Equip. Assocs., Inc., 958 F.2d 742, 744 (6th Cir.1992) (“Under normal rules of pleading, the statute of limitations is an affirmative defense and must be raised in the answer.” (quoting 5 Wright & Miller, Federal Practice & Procedure, § 1308, at 695 (1990)).
IV.
Next we consider whether the District Court erred in determining that the terms of the February 27, 1989 consent decree discriminate against the Tucker plaintiffs in violation of the Equal Protection Clause. In determining that the consent decree violated the Equal Protection Clause, the District Court stated in relevant part;
[i]n Croson, 488 U.S. at 507 [109 S.Ct. at 729], the Supreme Court held that race based set aside programs undertaken by municipalities are subject to strict scrutiny and that to satisfy the Equal Protection Clause, they must be “narrowly tailored to remedy prior discrimination.” Gender based preferences are likewise subject to strict scrutiny under the Equal Protection Clause.
The District Court noted that the Brunet plaintiffs in their underlying claim had failed to prove intentional discrimination by the City against female applicants and further stated that “[t]he only finding of discrimination against women was the finding that the PCT portion of the 1984 entrance examination had an adverse impact against women and that the City had failed to show that it was job related.” The District Court further stated that:
[although the Brunet plaintiffs argue that the consent decree is supported by past and present disparity of women in the position of firefighter, neither the City nor the Brunet plaintiffs attempted to prove past discrimination with evidence of disparity in the present [consolidated] proceedings. The. Court in the original Brunet case apparently did not accept .plaintiffs’ claims in that regard. In any event, mere statistical imbalance' alone would not suffice. Croson, 488 U.S. at 501 [109 S.Ct. at 725]; Hazelwood School District v. United States, 433 U,S. 299 [97 S.Ct. 2736, 53 L.Ed.2d 768] (1977). There was no evidence introduced here or in the original Brunet case which would justify a finding of a pattern or practice of discrimination against women. There was no evidence that the City itself had made any determination of past discrimination as the predicate for an affirmative action program for the hiring of female firefighters.
The District Court concluded that “[b]y agreeing to this order, the City has intentionally discriminated against male applicants” and that “[t]he consent decree violates the Equal Protection rights of male applicants .., because it is not narrowly tailored to remedy the prior discrimination found in Brunet.”9
A.
First, the Brunet plaintiffs argue that the District Court erred in subjecting the consent decree to the strict scrutiny standard of review. In Mississippi Univ. for Women v. Hogan, 458 U.S. 718, 724, 102 S.Ct. 3331, 3336, 73 L.Ed.2d 1090 (1982), the Supreme Court applied an intermediate standard of review to the university’s decision not to admit a male applicant to the nursing school solely on the basis of his gender. The Court examined this gender preference to determine if it possessed an important governmental objective and means substantially related to the achievement of the objective. Id. However, in Conlin v. Blanchard, 890 F.2d 811, 816 (6th Cir.1989), we relied on Wygant v. Jackson Bd. of Educ., 476 U.S. 267, 106 S.Ct. 1842, 90 L.Ed.2d 260 (1986) and City of Richmond v. J.A. Croson Co., 488 *404U.S. 469, 109 S.Ct. 706, 102 L.Ed.2d 854 (1989), to apply the strict scrutiny standard to an affirmative action plan based on a gender classification. Under the precedent in this Circuit, gender based affirmative action plans are subject to strict scrutiny when challenged under the Equal Protection Clause. Both we and the District Court are bound by the prior decision of this Court in Conlin. But see Coral Constr. Co. v. King County, 941 F.2d 910, 931 (9th Cir.1991) (applying intermediate scrutiny to gender based affirmative action plan), cert. denied, — U.S. -, 112 S.Ct. 875, 116 L.Ed.2d 780 (1992).
The Brunet plaintiffs also argue that the consent decree embodies a “gender conscious” plan, which should be subjected to the intermediate standard of review, rather than a “gender preference” plan, which is subjected to the strict scrutiny standard. They rely on Jacobson v. Cincinnati Board of Education, 961 F.2d 100, 102 (6th Cir.), cert. denied, — U.S. -, 113 S.Ct. 94, 121 L.Ed.2d 55 (1992), in which we employed the intermediate level of scrutiny because the school board’s transfer policy was race neutral, had no disparate impact on any race, and was thus, “race conscious” as opposed to being a “racial preference.”
Under the consent decree, some females with lower scores are hired before males with higher scores. The District Court later held (a holding that we affirm below) that applicants scoring higher on the test are better qualified than those with lower scores. Most importantly, the City knew at the time it entered into the consent decree that this was the case. Thus, the consent decree contains a gender preference since it requires that some lesser qualified women be hired before better qualified men.
B.
The Brunet plaintiffs next argue that the District Court erred by placing the burden of proof upon them and the City to prove the constitutionality of the consent decree by producing evidence of past discrimination against women by the CDF. Instead, the Brunet plaintiffs argue, the burden should have been on the Tucker plaintiffs to show an absence of prior discrimination on the part of the fire department against women.
Our decision in United Black Firefighters Ass’n v. City of Akron, 976 F.2d 999 (6th Cir.1992), is instructive. In United Black Firefighters, we considered the Equal Protection challenge of the local firefighters’ union to the district court’s order approving a consent decree between the United Black Firefighters Association (“UBFA”) and the City of Akron. The consent decree was designed to increase the percentage of black firefighters promoted to the position of lieutenant in the Akron Fire Department. In considering this challenge, we first determined that the local union had successfully shown that the consent decree embodied a racial preference. See id. at 1007. Our focus then shifted to whether this racial preference survived equal protection analysis under the strict scrutiny standard as required by the Supreme Court in Wygant and Croson.
Both Wygant and Croson involved Equal Protection challenges to a public employer’s affirmative action plan. We held that under Wygant and Croson, a state actor’s affirmative action plan survives strict scrutiny when (1) a compelling governmental interest underlies its racial preference, and (2) the preference is narrowly tailored to achieve that interest. We further held that, “[u]nder the first prong of the Croson test, a state actor possesses a compelling state interest when its concern is with remedying past discrimination.” United Black Firefighters, 976 F.2d at 1009. We then considered whether the City and/or the UBFA had produced sufficient evidence of past discrimination within the City’s fire department to provide a compelling governmental interest to support the affirmative action plan. Id. at 1010-11. In effect, once the local union proved a racial preference existed, the burden shifted to the defendants to produce evidence of past discrimination sufficient to justify the racial preference. We then held that sufficient evidence of past discrimination exists to support a compelling governmental interest in a racial preference where a “strong basis in evidence” of prior discrimination is shown. Id. *405at 1009; see also hong v. City of Saginaw, 911 F.2d 1192, 1196 (6th Cir.1990).
The Brunet plaintiffs rely on Johnson v. Transportation Agency, 480 U.S. 616, 107 S.Ct. 1442, 94 L.Ed.2d 615 (1987), to support their position that the Tucker plaintiffs must prove the absence of prior discrimination. However, the crucial distinction between Johnson and the present case is that in Johnson, the attack on the affirmative action plan at issue was based on Title YII alone; there was no Equal Protection challenge to the plan. The obligations of a public employer under Title VII and under the Constitution in regard to an affirmative action plan are not identical. Id. at 627 n. 6,107 S.Ct. at 1449 n. 6. Johnson is simply not helpful in resolving this issue. Thus, the District Court did not err in placing a burden of production upon the City and the Brunet plaintiffs to show evidence of past discrimination by the CDF in hiring female firefighters once the Tucker plaintiffs showed that the consent decree embodied a gender preference. Of course, the Tucker plaintiffs retain the ultimate burden of persuasion in proving the uneonstitutionality of the consent decree. Wygant, 476 U.S. at 277-78, 106 S.Ct. at 1848-49.
C.
The Brunet plaintiffs next argue that even if strict scrutiny is applied to the consent decree, they have successfully shown a strong basis in evidence of prior discrimination by the CDF sufficient to support a compelling governmental interest in the consent decree. As part of their efforts to make such a showing, the Brunet plaintiffs now seek to appeal the District Court’s May, 1986 orders in which the court held that the Brunet plaintiffs had failed to prevail on their intentional discrimination claim.
The Brunet plaintiffs had appealed the District Court’s findings within 30 days of the entry of the orders, but we dismissed that appeal on the ground that no final judgment on that issue had yet been entered. The Brunet plaintiffs concede that entry of the consent decree in this ease could be construed as the final judgment on their claims.
When the Brunet plaintiffs entered into the February 27, 1989 consent decree with the City, they agreed to forego all appeals from previous decisions of the District Court. However, because the consent decree has been set aside and they no longer can enjoy the benefit of their bargain, equitable principles require that they be allowed to now raise on appeal issues mooted by the consent decree. In Jansen v. City of Cincinnati, 977 F.2d 238 (6th Cir.1992), cert. denied, — U.S. -, 113 S.Ct. 2344, 124 L.Ed.2d 254 (1993), we stated that “[a] consent decree, although in effect a final judgment, is a contract founded on the agreement of the parties.... It should be construed to preserve the position for which the parties bargained....” Id. at 241 (quoting Vogel v. City of Cincinnati, 959 F.2d 594, 598 (6th Cir.), cert. denied, — U.S. -, 113 S.Ct. 86, 121 L.Ed.2d 49 (1992)). Since the Brunet plaintiffs no longer have the benefit of their bargain, they are entitled to the benefits of their litigation. Accordingly, we address the Brunet plaintiffs’ appeal from the District Court’s 1986 determination that they failed on their section 1983 claim alleging intentional discrimination.
In Brunet, the District Court held that in view of the substantial evidence suggesting the absence of discrimination, plaintiffs had failed to prove intentional discrimination. In reaching this conclusion, the court relied on the City’s extensive efforts at recruiting women and in preparing women for the examination, and that the women firefighters who were employed suffered no discriminatory treatment in their work. It acknowledged that prior to 1975, job announcements for firefighters were restricted to males and that only five of 832 firefighters were women. It noted that plaintiffs presented evidence about bias against women on the part of the Director of the Training Academy. But it also noted that this bias led to his removal. Finally, it recognized that defendants were aware, at various times, of less discriminatory testing methods than those that they were employing, but refused to adopt them.
The district judge did not discuss the change in testing that occurred in 1980 when *406determining the Brunet plaintiffs’ intentional discrimination claim. In 1975, the first year women were included in the firefighter applicant pool, and in 1978, the City administered the PCT on a pass/fail basis and ranked all candidates in the order of their scores on the CAT. In 1980, the City decided to score the PCT and ranked the candidates based on their combined PCT and CAT scores. The PCT was weighted 30 percent and the CAT 70 percent. Later in the same opinion, Judge Kinneary observed that this change was made “without any apparent consideration being given to possible greater adverse impact upon women.” Brunet, 642 F.Supp. at 1236. The court’s opinion does indicate that the purported reason for the change in approaches was a 1980 job analysis that concluded “that the work of firefighting was largely physical, and that better firefighters were distinguished by the ability to excel while performing physical tasks.” Id. The Civil Service Commission adopted the report’s recommendation that the PCT be made part of the ranking of job candidates. The court thus did not make an explicit finding that the City changed from a pass/fail to a scored PCT for legitimate non-discriminatory reasons. However, we believe that this was implicitly found by the court, as the court was aware of the change and of the purported non-discriminatory reasons for the change. We conclude that the District Court was not clearly erroneous in finding that the City did not intentionally discriminate against women in the hiring of firefighters.
D.
The Brunet plaintiffs argue that they can still show a strong basis in evidence to support the need for the affirmative action program set forth in the consent decree. They contend that the District Court, in its March 18, 1992 order, erred in holding that in order to produce evidence sufficient to support the consent decree, there must be a formal judicial finding of past discrimination. The Tucker plaintiffs do not argue that the City could enter into the consent decree only if there was a judicial finding of past discrimination. Rather, there was a finding of no intentional discrimination, and the Tucker plaintiffs argue that the only finding of discrimination in the Brunet plaintiffs’ underlying action was that the 1984 PCT had a disparate impact on women and that the test was not validated in violation of Title VII. Because this “singular” incident of discrimination had already been remedied, there was not a sufficient basis to support the consent decree’s affirmative action plan.
The Brunet plaintiffs are correct that a strong basis in evidence to support a public employer’s affirmative action plan does not require a formal finding of past discrimination. See Wygant, 476 U.S. at 289-92, 106 S.Ct. at 1854-57 (O’Connor, J., concurring); see also United Black Firefighters, 976 F.2d at 1009; Donaghy v. City of Omaha, 933 F.2d 1448, 1459 (8th Cir.1991), cert. denied, — U.S. -, 112 S.Ct. 938, 117 L.Ed.2d 109 (1992); Howard v. McLucas, 871 F.2d 1000, 1008 (11th Cir.), cert. denied, 493 U.S. 1002, 110 S.Ct. 560, 107 L.Ed.2d 555 (1989). In fact, a prima facie case of intentional discrimination is sufficient to support a public employer’s affirmative action plan. See Croson, 488 U.S. at 497, 109 S.Ct. at 724; United Black Firefighters, 976 F.2d at 1010. Thus, less evidence is necessary to justify an affirmative action plan than is necessary to prevail on an individual or class action claim of intentional discrimination. Given the disparity in the required evidentiary standards for each type of relief, a failure to prevail in a previous intentional discrimination claim does not preclude the Brunet plaintiffs from now attempting to show a strong basis in evidence to support the City’s affirmative action plan.
The District Court noted in its March 18, 1992 order that there was no finding of intentional discrimination in the Brunet plaintiffs’ 1986 trial. However, the District Court also stated that “[t]here was no evidence introduced here ... which would justify a finding of a pattern or practice of discrimination against women.” This statement indicates that the District Court did not simply rely on the prior decision in the underlying Brunet plaintiffs’ action in reaching its conclusion that insufficient evidence exists to support the consent decree. In light of this holding, we next examine the evidence that the Brunet plaintiffs contend provides a strong basis in evidence of discrimination.
*407Appropriate statistical evidence setting forth a prima facie case of discrimination is sufficient to provide a strong, basis in evidence to support a public employers’ affirmative action plan. See Croson, 488 U.S. at 501, 109 S.Ct. at 725; United Black Firefighters, 976 F.2d at 1010. The District Court stated that “[m]ere statistical imbalance alone would not suffice.” The District Court is correct. Rather, as we stated in United Black Firefighters:
[t]he method generally used is to compare the minority percentage in the relevant statistical pool to the minority percentages in the group of persons selected for the position at issue. See Croson, 488 U.S. at 501-03, 109 S.Ct. at 725-26; Hazelwood School District v. United States, 433 U.S. 299, 308, 97 S.Ct. 2736, 2742, 53 L.Ed.2d 768 (1977); Long, 911 F.2d at 1199. The relevant statistical pool is comprised of all persons qualified for the position at issue. Roughly the same percentage of minorities observed in the relevant statistical pool should also be observed in the group of persons selected for the position at issue. Id. Where a gross disparity exists between the expected percentage of minorities selected and the actual percentage of minorities selected, then prima facie proof exists to demonstrate intentional discrimination in the selection of - minorities to those particular positions. See Croson, 488 U.S. at 501, 109 S.Ct. at 725; Hazelwood, 433 U.S. at 307-09, 311 n. 17, 97 S.Ct. at 2741-42, 2743 n. 17....
976 F.2d at 1011.
The Brunet plaintiffs offer the following statistical evidence to support the consent decree. Two percent of all those who passed the 1988 firefighter examination were women. However, had it not been for the court ordered proportional hiring requirement, no women would have been hired from that examination list. Thirty-one females out of 519 or 6 percent passed the examination administered in 1990; however, had it not been for the consent decree, only one woman would have been hired. These statistics do indicate a significant disparity. Nevertheless, as we stated in United Black Firefighters, “a gross disparity is not conclusive as to a finding of discrimination, because othér factors unrelated to race may account for the disparity.” Id. Therefore, those opposing the affirmative action plan may present evidence to rebut the inference, thus defeating the validity of the affirmative action plan. See Stuart v. Roache, 951 F.2d 446, 453 (1st Cir.1991), cert. denied, — U.S. -, 112 S.Ct. 1948, 118 L.Ed.2d 553 (1992).
The Tucker plaintiffs have presented evidence to rebut this evidence. The reason 'why only one woman would have been hired from the 1988 and 1990 firefighter examinations absent the proportional hiring requirement is that the PCT is used to rank candidates for hiring purposes and women perform much more poorly than men as a whole on the PCT. The PCT developed in 1986 is valid and, as we determine below, and as the City knew in 1987, rank-order hiring did not discriminate since the PCT and the CAT are both job-related methods of selecting the most qualified individuals for the position of firefighter. Thus, a factor other than intentional discrimination against females accounted for the disparity.
The Brunet plaintiffs also note than in 1986, before the trial in their underlying claim, only 6 of 822 or .72 percent of the City’s firefighter force were women and that these women had been hired only as a result of the affirmative action plan arising from Dozier designed to correct past discrimination against black candidates. However, this statistical information is not particularly helpful in showing discrimination or ¡rebutting discrimination because there is'no evidence of how many women desired the position, were qualified and yet were not hired. Moreover, the same reason for the srtiall percentage of women on the fire force today explains why there were so few women on the force in 1986 — the PCT has an adverse impact on women.
In its 1986 opinion, the District Court found that the 1980 examination did not have an adverse impact upon women because men and women were hired at about the same rate in proportion to the numbers of males and females eligible for hire. The court declined to attach significance to the fact that all of the women hired were hired from the *408Dozier list. We think that this was error. We would find that the 1980 examination had an adverse impact upon women, because in absence of the Dozier one-for-one hiring requirement no women would have been hired. However, this discrimination had already been remedied as four women were in fact hired in 1980. We can safely assume that the women actually hired in 1980 were the most qualified women and would be hired first from a female list. Likewise, the discrimination that the court found to exist, in the administration of the 1984 examination had also been remedied.
The Brunet plaintiffs are not limited to statistical evidence to support the consent decree. United Black Firefighters, 976 F.2d at 1011. The Brunet plaintiffs refer to other evidence arising as early as 1975 to support the consent decree. The Tucker plaintiffs argue that the consent decree cannot be supported by evidence arising outside the statute of limitations for the Brunet plaintiffs’ underlying claim. The Tucker plaintiffs rely on Campbell v. Dayton, No. C-3-89-098, slip op. at 8 n. 7 (S.D.Ohio Dec. 9, 1991), in which the district court held that the plaintiffs’ claim under Title VII for disparate treatment could not concern acts occurring outside the applicable two-year statute of limitations. This case does not support the Tucker plaintiffs’ position because the Campbell plaintiffs were seeking relief under Title VII and 42 U.S.C. §§ 1981 and 1983.
In contrast, the Brunet plaintiffs have introduced evidence for the purpose of justifying a public employer’s affirmative action program, not for the purpose of prevailing on their Title VII and section 1983 claims. This distinction is important since an affirmative action program serves a very different.purpose than does a claim brought under Title VII or section 1983. As the Supreme Court stated in Local 28 of Sheet Metal Workers v. EEOC, 478 U.S. 421, 474, 106 S.Ct. 3019, 3049, 92 L.Ed.2d 344 (1986):
[t]he purpose of affirmative action is not to make identified victims whole, but rather to dismantle prior patterns of employment discrimination and to prevent discrimination in the future. Such relief is provided to the class as a whole rather than to individual members; no individual is entitled to relief, and beneficiaries need not show that they were themselves victims of discrimination.
In this connection, we as well as other courts have considered evidence of prior discrimination occurring years before the affirmative action plan at issue. In Vogel, 959 F.2d at 600, we considered evidence as far back as eight years before the underlying claim was filed and nine years before the affirmative action plan at issue was instituted, in upholding an affirmative action plan that provided racial preferences. See also Stuart, 951 F.2d at 452 (considering evidence of discrimination occurring eight years before the underlying claim was filed and ten years before the affirmative action plan at issue was instituted).
The Tucker plaintiffs also argue, relying on Quirin v. City of Pittsburgh, 801 F.Supp. 1486 (W.D.Pa.1992), that evidence occurring before 1972, when Title VII was made applicable to state and local governments, cannot be considered to justify an affirmative action plan. However, Quirin does not stand for this proposition. The district court in Quirin found that the only evidence of discrimination against women by the City of Pittsburgh occurred' prior to 1972 and that this evidence was too remote in time from the affirmative action plan implemented in 1987 to provide a sufficient evidentiary basis to support it. See id. at 1492; see also Howard, 871 F.2d at 1007 (considering evidence of discrimination occurring in 1971 to support, in part, the affirmative action plan at issue). Consequently, we may consider evidence occurring prior to 1972 and outside the statute of limitations applicable to the Brunet plaintiffs’ Title VII and section 1983 claims. The question remains, however, whether this evidence constitutes a “strong basis in evidence” to support the consent decree at issue.
The strongest evidence of discrimination against women at the CDF is the testimony of Dr. Kriska, the City’s expert who has been in charge of testing of firefighters since 1974, that prior to 1975 the City’s job announcements for the position of firefighter specified “males.” This designation would clearly indicate that the City preferred, if not required, *409males over females for this position and would have the effect of discouraging women from applying for the job. However, the District Court found in Brunet that there was “substantial evidence” that the City made efforts to encourage women to apply for the job of firefighter after 1975. 642 F.Supp. at 1223. While evidence of discrimination prior to 1975 may be considered, it is conduct that occurred at least 14 years before the February 27, 1989 consent decree was entered. In Hammon v. Barry, 826 F.2d 73, 76-77 (D.C.Cir.), order granting reh’g en banc, 833 F.2d 367 (D.C.Cir.1987), order granting reh’g en banc vacated, 841 F.2d 426 (D.C.Cir.), cert. denied, 486 U.S. 1036, 108 S.Ct. 2023, 100 L.Ed.2d 610 (1988), the Court held that discriminatory conduct occurring 18 years prior to the institution of the affirmative action plan at issue was insufficient by itself to justify the plan under Title VII standards. See also Fountain v. City of Waycross, 701 F.Supp. 1570, 1577 (S.D.Ga. 1988) (if the discrimination is sufficiently remote, it will not create compelling interest).
In this case, the discriminatory policy against women in the Columbus fire department prior to 1975 is too remote to support a compelling governmental interest to justify the affirmative action plan embodied in the consent decree. In sum, the affirmative action plan, which provides preferential hiring to females, is not supported by a strong basis in evidence of prior discrimination sufficient to pass constitutional muster — any discrimination against women caused by adverse impact has been remedied, and the intentional discrimination against women that occurred prior to 1975 is too remote. . Further, the Brunet plaintiffs have offered no direct evidence of any intentional discrimination against women on the part of the City after 1975. Since 1987, the City has known that the PCT is job related and that a better score on the PCT and the CAT indicates a more qualified candidate. Thus, since 1987, the PCT’s adverse impact on women has been justified by non-discriminatory job-related reasons. In the absence of a “compelling governmental interest,” i.e., any prior unremedied or current discrimination, the City had no constitutional basis on which to enter into the consent decree, which provided for the hiring of lesser-qualified women and denied men the opportunity to compete for all firefighter positions. By doing so, the City has intentionally discriminated against male applicants because of their gender. We affirm the District Court’s holding that the consent decree violates the Equal Protection rights of male applicants for the entry level position of firefighter. Because, the plan is unsupported by a compelling governmental interest, it is unnecessary for us to consider whether the plan is narrowly tailored as required under Croson.
V. RANK-ORDER HIRING
The District Court determined that a linear relationship between test scores on the PCT and success on the job exists. In other words, the District Court determined that higher scores on the PCT are indicative of greater success as a firefighter. The District Court also determined that no comparable selection devices exist that are “substantially equally valid” as strict rank-order hiring and have less of a disparate impact on women. The Brunet plaintiffs challenge both of these determinations. The District Court based its conclusions primarily on the concurrent criterion-related validity study prepared -by Dr. Landy in 1987. A criterion-related validity study seeks to show the correlation between one’s score on a test and one’s subsequent success on the job. In a concurrent criterion-related analysis, an employment test is given to incumbents who have already obtained the necessary job experience and have an established record of performance. The incumbents’ test scores are compared to their success on the job to determine the degree to which the test successfully measures incumbents’ job performances.
In conducting the study, Dr. Landy identified a representative sample of incumbent firefighters who were given the PCT. He then gathered performance ratings from supervisors from these incumbents on eighteen different firefighting performance areas. Dr. Landy next calculated the correlation coefficients, or measures of association, between the incumbents’ test scores and their performance ratings. Both Dr. Landy and the *410City’s expert, Dr. Kriska, determined on the basis of this study that there was a sufficient correlation between test scores and job performance to support using the PCT for ranking purposes. The District Court agreed.
This conclusion is a factual finding reviewed only for clear error. See Police Officers for Equal Rights v. City of Columbus, 916 F.2d 1092, 1103 (6th Cir.1990). “Ranking is a valid, job-related selection technique only where the test scores vary directly with job performance.” Williams v. Vukovich, 720 F.2d 909, 924 (6th Cir.1983). Where a test that has a disparate impact upon a protected class of people is shown to be content valid and valid for purposes of ranking, the plaintiff may still prevail under a Title VII claim if he or she can show that other tests or selection devices would also serve the employer’s legitimate hiring interests and at the same time have a lesser adverse impact upon that protected class. Police Officers, 916 F.2d at 1096.
It is not disputed that the 1986 PCT is content valid. However, the Brunet plaintiffs argue that the PCT is not appropriate for ranking purposes. They point to what they believe are serious flaws in the concurrent criterion-related validity study that purport to show that the PCT is appropriate for ranking purposes.
The Brunet plaintiffs contend that the hose hoist event, a non-eritical task, was considered by Dr. Landy in calculating the correlation coefficients for the PCT. An examination of Appendix K of the study seems to indicate that this contention is correct. Appendix K provides the correlation coefficient between the PCT and overall job performance. The City’s expert, Dr. Kriska, stated at trial that Appendix K was the crux of the concurrent criterion-related validity study and that he relied on the correlation coefficients in Appendix K most heavily in determining that ranking on the basis of the PCT was valid. In addition, the District Court also relied heavily on Appendix K in reaching its conclusion that ranking on the basis of the PCT is valid. The correlation coefficient for the hose hoist event is .42, a value higher than that for all the other events on the PCT. The correlation coefficient for the overall PCT is .29. Other courts have found such correlation coefficients to be predictive of job performance, thus indicating the appropriateness of ranking where the correlation coefficient value is .30 or better. Clady v. County of Los Angeles, 770 F.2d 1421, 1431-32 (9th Cir.1985), cert. denied, 475 U.S. 1109, 106 S.Ct. 1516, 89 L.Ed.2d 915 (1986); Boston Chapter of NAACP v. Beecher, 504 F.2d 1017, 1024 n. 13 (1st Cir.1974), cert. denied, 421 U.S. 910, 95 S.Ct. 1561, 43 L.Ed.2d 775 (1975); see also EEOC v. Atlas Paper Box Co., 868 F.2d 1487, 1502 n. 24 (6th Cir.) (Cook, J. concurring in part and dissenting in part), cert. denied, 493 U.S. 814, 110 S.Ct. 63, 107 L.Ed.2d 30 (1989). The correlation coefficient for the overall PCT is only .29 with the hose hoist event and would most likely be even less without the hose hoist event. Additionally, the Uniform Guidelines provide that a test that “has a low correlation coefficient will be subject to close review if it has a large adverse impact.” 29 C.F.R. § 1607.14B(6); see also Clady, 770 F.2d at 1432 (“[a]s a general principle, the greater the test’s adverse impact, the higher the correlation which will be required”); Atlas Paper Box, 868 F.2d at 1502 n. 24.
In rebuttal, the City and the Tucker plaintiffs argue that use of the PCT for ranking purposes can be justified on the basis of content validity alone and that because the PCT has been determined to be content valid, it is also valid for purposes of ranking. The Uniform Guidelines provide in relevant part:
[w]here a selection procedure supported solely or primarily by content validity is used to rank job candidates, the selection procedure should measure those aspects of performance which differentiate among levels of job performance.
29 C.F.R. § 1607.14C(9). However, the District Court in its May 27, 1987 opinion and order determined that despite the fact that the PCT was content valid, further studies such as a concurrent criterion-related validity study were required to determine the validity of the PCT for ranking purposes.
Next, the Tucker plaintiffs point out that the hose hoist event was performed last during the concurrent criterion-related validity *411study in order to prevent affecting other events. While this may be the case, the fact remains that the correlation coefficient for the hose hoist event was still included in the correlation coefficient for the overall PCT. However, as the District Court noted, it was necessary to evaluate the conflicting testimony of two well-qualified experts in the field of industrial psychology and it found the testimony of the Tucker plaintiffs’ expert, Dr. Landy, more convincing than the testimony of the Brunet plaintiffs’ expert, Dr. Cranny. The court noted that while Dr. Cranny had never prepared an examination for firefighters and had little experience in the testing of public safety personnel, Dr. Landy has had extensive practical experience in the development, validation, and administration of public safety personnel selection in promotion examinations. The court noted that Dr. Landy has studied the firefighter job first hand, living in fire houses and accompanying firefighters inside burning structures. The use of statistical data to validate an examination for ranking purposes is a complicated matter requiring a high level of expertise. A district judge is entitled to rely on the opinion of one expert in examining questions requiring this type of expertise. See Police Officers, 916 F.2d at 1103.
Furthermore, under the Uniform Guidelines, the values for the correlation coefficients in Appendix K are all statistically significant. Under the Guidelines, the correlation coefficient must have a value of at least .05, which means that it would be expected to occur no more than five times out of a hundred by chance. 29 C.F.R. § 1607.14(b)(5). Where the correlation coefficient is statistically significant, the scores from the selection procedure are considered to possess a direct relationship to one’s performance on the job. Id.
We hold that the District Court was not clearly eiToneous in its determination that rank-order hiring is valid.
VI. ALTERNATIVE HIRING PROCEDURES
We next consider whether the District Court erred in determining that there were no alternative selection devices that have less of a disparate impact on females and are “substantially equally valid” methods of choosing qualified firefighters. Uniform Guidelines, 29 C.F.R. § 1607.3(B). It is undisputed that strict rank-order hiring has an adverse impact on female firefighter applicants. Under the auspices of Title VII, once the City has shown that the challenged test is job related, as the City has in this case, the burden shifts to the challengers to prove that the City was using the validated tests as a pretext for discrimination by demonstrating that other tests, without the adverse effects, would also serve the City’s legitimate hiring interests. Police Officers, 916 F.2d at 1096.
The Brunet plaintiffs offer two alternatives. First, they suggest that the use of the PCT should be dropped altogether because the correlation coefficient for the CAT as set forth in Appendix K indicates that it alone is more predictive of job performance than the combined PCT and CAT tests. The correlation coefficient for the CAT alone is .52 while the correlation coefficient for the combined test is .44. The District Court rejected this argument relying on Dr. Landy’s testimony showing how the statistical data for the PCT and the CAT could be manipulated to make the PCT alone appear more predictive of success as a firefighter. The court also determined, based on a multiple regression analysis conducted by Dr. Landy, that the PCT was necessary to test those physical skills required to be a firefighter. The court concluded that without the PCT, those physical skills simply would not be tested, and thus, the CAT alone is not as effective in choosing qualified firefighters as is the PCT and the CAT combined. Because we find nothing in the record that leads us to conclude that the District Court was clearly erroneous in this finding, we agree that the PCT cannot be completely eliminated from the firefighter examination.
The Brunet plaintiffs next suggest that the City should administer the PCT on a pass/fail basis and then rank candidates solely on the basis of their CAT scores, since the CAT does not have an adverse impact on women. The City utilized this procedure in its 1975 and 1978 firefighter examinations. Although the consent decree provided for the pass/fail *412administration of both the CAT and the PCT, men and women were ranked, albeit on separate lists, in the order of their combined CAT and PCT scores. The Brunet plaintiffs argue that there is no evidence indicating that the City was dissatisfied with any of the firefighters hired according to these procedures, implying that a pass/fail examination is a viable alternative to a scored examination. The District Court held that rank ordering is justified because there is a direct correlation between performance on the PCT and performance on the job. We believe that the following conclusion is implicit in this holding: For the purpose of hiring the most qualified candidates, a procedure that selects from a pool of applicants who achieve a passing score on the PCT and who are ranked on the basis of the CAT alone, is not “substantially equally valid” to a procedure that chooses its hires from a pool of applicants ranked according to their scores on both the CAT and PCT. We agree.
The posture of this case at present differs from the more typical Title VII challenge to hiring devices. The Brunet plaintiffs initially brought a Title VII action against the City in 1984, which culminated in the consent decree that instituted the proportional hiring requirement. After the Tucker plaintiffs’ successful challenge to the consent decree on constitutional grounds, the consent decree was set aside. The City then indicated that in absence of the consent decree it wished to rank order all candidates on one gender-neutral list. However, under the Uniform Guidelines, 29 C.F.R. § 1607.3B, before implementing a process of strict rank-order hiring from one list, a process that has an adverse impact upon women, the City is obligated to conduct its own investigation of viable alternatives with lesser or no impact on the female applicants. “[Bjefore utilizing a [selection] procedure that has an adverse impact on minorities, the City has an obligation pursuant to the Uniform Guidelines to explore alternative procedures and to implement them if they have less adverse impact and are substantially equally valid to rank ordering.” Officers for Justice v. Civil Service Commission, 979 F.2d 721, 728 (9th Cir.1992) (citing the Uniform Guidelines, 29 C.F.R. § 1607.3B). There is nothing in the record indicating the City explored alternatives to strict rank ordering or that the District Court looked to see that the City had done so. We believe that this was error. We reiterate that a selection procedure that ranks only on the basis of CAT scores is not acceptable. However, we have found nothing in the record that requires the CAT and the PCT to be weighted equally.10 In 1980 and 1984, the PCT constituted 30 percent of an applicant’s score; the CAT weighed in at 70 percent. The City should be required to demonstrate why the CAT, which arguably is more predictive than the PCT, should not be weighted more than the PCT. We suspect that this change would result in less of an adverse impact on women, although it will not eliminate it. But as Judge Kinneary stated seven years ago when the Brunet plaintiffs’ journey began:
Title VII does not require employers to equalize the probabilities of hiring of the average members of two groups. Rather, it requires that actual individuals enjoy opportunities for employment free from discriminatory barriers.
It is not the province of the Court to determine whether women should be firefighters, or how many women should be firefighters. Rather, it is the Court’s duty to evaluate a test in light of the standards set forth in Title VII. How many women should be firefighters can be decided only by the administration of a validated examination.
Brunet, 642 F.Supp. at 1228, 1250.
VII. REMEDY
Lastly, the Brunet plaintiffs assert that the District Court erred in granting the remedy it did in its September 23, 1992 order. The District Court found that “an order adjusting the seniority of male and female firefighters hired pursuant to the 1990 Firefighters Examination is necessary and appropriate to provide the Tucker class of plaintiffs with complete relief.... But for *413the consent decree of February 27, 1989, the members of the Tucker class would have been hired before Firefighters' Fox and Sachs and would have greater seniority than Firefighters Fox and Sachs.” As a result, the District Court ordered that Fox’s and Sachs’ seniority be adjusted so that it is less than that of all males hired from the 1990 examination.
“A district court has ‘not merely the power but the duty to render a decree which will so far as possible eliminate the discriminatory effects of the past as well as bar like discrimination in the future.’ ” United States v. Paradise, 480 U.S. 149, 183, 107 S.Ct. 1053, 1073, 94 L.Ed.2d 203 (1987) (quoting Louisiana v. United States, 380 U.S. 145, 154, 85 S.Ct. 817, 822, 13 L.Ed.2d 709 (1965)). It is within the sound discretion of the district court to fashion such a remedy. Id. 480 U.S. at 183-84, 107 S.Ct. at 1072-73. However, the remedy must be narrowly tailored to address the discrimination at issue and take into consideration the interests of all parties affected by the remedy. Id. at 185, 107 S.Ct. at 1073.
The Tucker plaintiffs argue that the relief was appropriate because the practical effect of this adjustment was that Fox and Sachs were placed in the seniority order they would have been in had they not been given a preference under the consent decree. The remedy in this action should be to put the Tucker plaintiffs in the same position they would have been in absence of the consent decree. The Brunet plaintiffs argue that the focus of the District Court’s remedy is improperly placed on putting Fox and Sachs in the same or similar position that they would have been in the absence of the consent decree. Moreover, Fox and Sachs retain an important interest in their seniority. See Oliver v. Kalamazoo Bd. of Educ., 706 F.2d 757, 763 (6th Cir. 1983) (“the Supreme Court has recently reiterated the ‘overriding importance’ of seniority provisions.... ”). “[T]o protect the strong expectations in these pervasive and important seniority rights, the remedy must be ‘necessary,’ not merely, ‘reasonable’ to vindicate the constitutional rights” of the victims. Id. (citation omitted).
In this case, the fashioned relief does not go beyond remedying the effects of discrimination suffered by the Tucker plaintiffs. As discussed at length above, plaintiffs Tudor, Meyer and Hilleary and the class they represent-were harmed'by the fact that Fox and Sachs are either as senior or more senior than them. The only way to remedy their injury is to place Fox and Sachs lower in the seniority hierarchy than the Tucker class, as they would have been in absence of the consent decree. Accordingly, we affirm the District Court’s order of September 23, 1992, adjusting the seniority of firefighters Fox and Sachs.
VIII.
The District Court’s order setting aside the consent decree is AFFIRMED. The District Court’s order adjusting the seniority of firefighters Fox and Sachs is AFFIRMED. The District Court’s order validating rank-order hiring is REVERSED. The case is REMANDED for proceedings consistent with this opinion.
. The Tucker plaintiffs also named as defendants in their action: Larry H. James, Director of Public Safety; Harmon J. Dutko, Chief of the Division of Fire; the Civil Service Commission of Columbus; Civil Service Commissioners Daniel D. Connor, Julia F. Johnson, and Frederick L. Ransier.
. The other defendants were the Columbus Civil Service Commission, Columbus Mayor Dana Rinehart, and Columbus Safety Director Alphonso Montgomery.
. We note that while the consent decree provides for the pass/fail administration of the physical and cognitive portions of the examination, the City actually scored the results of both tests and used the scores to create two rank-ordered lists to comply with the proportional hiring requirement of the consent decree. The consent decree did not require this result. The City could have used another method of choosing candidates from the male and female pools of applicants other than maintaining dual rank ordered lists, e.g., a lottery system.
. One woman, Darrylee Potter, who ranked 142, was ranked higher than Tucker. Ms. Potter had been hired in the March, 1991 class.,
. Even if the first 35 ranked applicants were chosen for the first 35 positions, candidates 36, 37 and 38 would have been considered for the 36th position. Tudor and Meyer need not show that they would have been selected for the position, but for the consent decree, only that they would have been considered. See Bakke, 438 U.S. at 280-81 n. 14, 98 S.Ct. at 2742-43 n. 14.
.Shortly after this series of orders and appeals, the Brunet plaintiffs filed a motion requesting that Judge Kinneary disqualify himself for bias. The court denied this motion. On the same day, August 22, 1991, Judge Kinneary did recuse himself sua sponte, and Judge Graham was assigned to the case.
. In their complaint, the Tucker plaintiffs made the following assertions:
The plaintiff class consists of all males who desire to become Columbus firefighters in the August 19, 1991 class, or in subsequent entry-level firefighter classes, and who are affected by the discriminatory practices described herein....
Although two firefighter positions are at issue for the firefighter class commencing August 19, 1991, the plaintiff class is much broader than simply two persons since it includes all persons who will be affected by the discriminatory practices over approximately the next eighteen years. The plaintiff class includes males who desire to become Columbus firefighters in future classes, and who are or would be affected by the discriminatory practices described herein either in respect of hiring or in respect of their seniority within the Division of Fire.
Tucker Complaint, ¶¶ 12, 13. The action did not reach back to the hiring that took place in March of 1991. The men who lost their position to the women hired in March, 1991 are not part of the Tucker class. Indeed, the women’s placement in the March, 1991 class pursuant to the consent decree is not challenged. Therefore, we do not include the March, 1991 hiring in our analysis of the effect of the consent decree on the Tucker plaintiffs in August, 1991.
. At the time of class certification, Hilleary was certain to be hired in the next class absent unforeseen circumstances unrelated to his ranking. We note that Hilleary was in fact hired in the next class.
. The City seeks to challenge the. District Court’s finding that it intentionally discriminated against the male plaintiffs by agreeing to the February 27, 1989 consent decree. The City admits it has not filed a notice of appeal from the District Court's March 18, 1992 order, and we therefore lack jurisdiction to consider the City’s challenge. See e.g., Torres v. Oakland Scavenger Co., 487 U.S. 312, 315, 108 S.Ct. 2405, 2408, 101 L.Ed.2d 285 (1988).
. It may be that in the voluminous record this issue was addressed, but we are unable to find it.