FAGG, Circuit Judge.
These appeals and cross-appeals follow the entry of judgment partially in favor of and partially against the Missouri Highway and Transportation Commission in a suit combining individual and class sex discrimination claims. Our review will focus on five major issues: (1) the estoppel effect of the jury verdict on a district court acting as factfinder on coexisting equitable claims; (2) the sufficiency of the evidence of class discrimination in a case combining statistical and anecdotal evidence; (3) the calculation of class back pay awards; (4) the propriety of orders imposing gender-conscious goals with respect to future hiring; and (5) the enhancement of attorney fees. We affirm the state’s liability to the class but reverse the state’s liability to the individual plaintiffs and remand for modification of the relief awarded and for further proceedings consistent with this opinion.
The allegations of discrimination in this case focus on Missouri’s practices and policies in hiring highway maintenance workers. Requirements for this entry-level position are an eighth-grade education and an ability to operate lightweight motor equipment, and duties include mowing highway right-of-ways, plowing snow, filling potholes, maintaining rest areas, and performing minor service and repairs on equipment. The individual plaintiffs, Jane Catlett, Grace Tuter, Adeline Kallemyn, and Patricia Leembruggen, all applied for and were denied maintenance positions in District Eight of the Missouri Highway and Transportation Commission. These women are also members of the plaintiff class, which consists of all females who applied or might have applied for maintenance positions in District Eight between January 1, 1975, and May 31,1980. District Eight had never employed a female maintenance worker as of July 1975, when Catlett filed the state administrative complaint that evolved into this litigation, and the District during the time period defined by the class [1263]*1263hired eighty-nine males and only eight females.
The individual plaintiffs and the class base their claims on two federal statutes, section 1983 (that is, 42 U.S.C. § 1983) and Title VII of the Civil Rights Act of 1964, 42 U.S.C. § 2000e-2. To prevail under section 1983 the individuals and the class had to prove Missouri intentionally treated them less favorably because of their sex. See Foster v. Wyrick, 823 F.2d 218, 221 (8th Cir.1987). Under Title VII these claimants had the option of proving disparate treatment, that is, Missouri engaged in intentional discrimination as under section 1983, or disparate impact, that is, Missouri employed facially neutral hiring practices that in fact operated to limit job opportunities for women. See generally International Bhd. of Teamsters v. United States, 431 U.S. 324, 335 n. 15, 97 S.Ct. 1843, 1854 n. 15, 52 L.Ed.2d 396 (1977) (distinguishing disparate treatment and disparate impact theories of discrimination). The individual and class section 1983 claims were tried to a jury while the individual and class Title VII claims simultaneously were tried to the court. See Harmon v. May Broadcasting Co., 583 F.2d 410, 410 (8th Cir.1978) (per curiam).
The jury on the individual section 1983 claims returned a verdict in favor of Missouri, thus finding no intentional discrimination. The district court, however, reached a contrary conclusion when under Title VII it held for the individual plaintiffs on a disparate treatment theory. Catlett v. Missouri Highway & Transp. Comm’n (Catlett I), 589 F.Supp. 929, 947-49 (W.D.Mo.1983). The jury and the court both found intentional discrimination by Missouri against the class, and the court also found Missouri’s employment practices had a disparate impact on the class. Id. at 944-47. This disparate impact conclusion pinpointed Missouri’s word-of-mouth recruiting system as the neutral factor artificially limiting opportunities for women. Id. at 947.
At the remedial stage the district court awarded back pay to the individual plaintiffs and the class and retained jurisdiction over the suit until January 1,1990, to monitor the implementation of a broad order of affirmative relief. Catlett v. Missouri State Highway Comm’n (Catlett II), 627 F.Supp. 1015 (W.D.Mo.1985); see also Catlett v. Missouri Highway & Transp. Comm’n (Catlett III), 643 F.Supp. 321 (W.D.Mo.1986) (setting actual back pay amounts). The court in this order granted a classwide hiring preference, thus giving priority for future job openings to all class members still desiring employment as maintenance workers, and set a goal of thirty-seven to forty-eight percent female for the maintenance applicant pool and for District Eight maintenance hires during the pendency of the court order. Catlett II, 627 F.Supp. at 1020, 1021. The court also called for the designation “maintenance-man” to be replaced with a sex-neutral job title and detailed new recruiting, application, and interview procedures to be followed by Missouri. Id. at 1020-22. Finally, the district court in a later order awarded the plaintiffs attorney fees of $505, 907.62 based on a fifty-percent enhancement of the hours-times-rate “lodestar” figure.
I.
Missouri argues the district court was bound by the jury verdicts in the state’s favor on the individual section 1983 claims and thus lacked the power to make contrary factual findings and enter judgments against Missouri on the individual Title VII claims. The individual plaintiffs acknowledge the principle of law on which Missouri relies, see Garza v. City of Omaha, 814 F.2d 553, 557 (8th Cir.1987); Goodwin v. Circuit Court, 729 F.2d 541, 549 n. 11 (8th Cir.), cert. denied, 469 U.S. 828, 105 S.Ct. 1194, 84 L.Ed.2d 339 (1984), but they offer three reasons why it should not apply to bar their individual Title VII recoveries here.
First, the individual plaintiffs point to Missouri’s failure to rely on estoppel by jury verdict when submitting to the court proposed conclusions of law on the Title VII claims. Missouri did raise the estoppel issue at the remedial stage, Catlett II, 627 [1264]*1264F.Supp. at 1017, and this satisfies our concern that a district court be given the opportunity to consider and correct perceived errors, see Edwards v. Hurtel, 724 F.2d 689, 690 (8th Cir.1984) (per curiam). The Goodwin case, in which this court declined to apply estoppel, is distinguishable because of its “peculiar circumstances”: Estoppel in that case had not been raised either to the district court or on appeal, and the party against whom estoppel was asserted had not been a party to the claim decided by the jury. See 729 F.2d at 549 n. 11. We believe we may properly reach the estoppel issue here.
Second, the individual plaintiffs argue estoppel does not apply because the district court’s Title VII decisions were based on findings beyond those made by the jury. Specifically, the individual plaintiffs point out that while the jury under section 1983 considered only the existence of intentional discrimination, the court under Title VII also considered the existence of unintentional discrimination resulting from the use of a facially neutral employment practice. As we understand the district court’s opinion, however, the court drew a specific legal conclusion of disparate impact only with regard to Missouri’s reliance on word-of-mouth recruiting, finding that the existing predominantly male work force communicated information on job openings primarily to other males. See Catlett I, 589 F.Supp. at 946-47. Since the individual plaintiffs all applied for maintenance positions, they were not victims of this discriminatory practice, and they cannot benefit from the court’s disparate impact analysis.
Third, the individual plaintiffs would avoid estoppel by voiding the jury verdicts: They argue the district court committed error by failing to grant them judgments notwithstanding the verdict on their individual claims. The judgments notwithstanding the verdict, however, were properly denied if, resolving all factual conflicts in favor of Missouri and giving Missouri the benefit of all favorable inferences, the evidence was sufficient to support the jury’s conclusions. See Gilkerson v. Toastmaster, Inc., 770 F.2d 133, 136 (8th Cir.1985). We assess the sufficiency of the evidence on the ultimate question of whether Missouri intentionally discriminated against the individual plaintiffs. See id. at 135 (citing United States Postal Serv. Bd. of Governors v. Aikens, 460 U.S. 711, 714-15, 103 S.Ct. 1478, 1481-82, 75 L.Ed.2d 403 (1983)); see also Craik v. Minnesota State Univ. Bd., 731 F.2d 465, 468 & n. 5 (8th Cir.1984) (section 1983 employment claim follows Title VII disparate treatment analysis).
The record suggests a variety of nondiscriminatory reasons for Missouri’s refusals to hire the individual plaintiffs. Catlett in the first instance applied for a position after a hiring decision possibly had already been made (although not officially approved) and in future instances refused to update her application with the name of her current employer. Tuter and Kallemyn interviewed for vacancies eventually filled by males with prior highway maintenance experience, and Leembruggen did not list on her application the lightweight equipment skills required for the job. In addition, although each plaintiff’s application remained active for a year, the state had a policy of considering most maintenance candidates only once during that period. The individual plaintiffs respond that certain of these hiring rationales, while facially legitimate, are inherently discriminatory and as a matter of law cannot support the individual verdicts in favor of Missouri. Cf. O’Connor v. Peru State College, 781 F.2d 632, 636 (8th Cir.1986). The individual plaintiffs, however, did not raise this argument in their motion for judgments notwithstanding the verdict, and they do not on appeal find fault with the jury instructions for failing to reflect this theory. Any challenge to Missouri’s hiring rationales as being inherently discriminatory has not been adequately raised and preserved.
Because the record contains ample evidence to support the jury’s verdicts for Missouri on the individual section 1983 claims, we affirm those verdicts and further hold the verdicts bar judgments against Missouri on the individual Title VII claims.
[1265]*1265II.
Turning to the class claims, we focus our review on the jury’s and court’s findings of intentional discrimination. A finding of classwide intentional discrimination requires proof by a preponderance of evidence that the employer engaged in a pattern or practice of unlawful discrimination. Briggs v. Anderson, 796 F.2d 1009, 1019 (8th Cir.1986); see also Hervey v. City of Little Rock, 787 F.2d 1223, 1228 n. 3 (8th Cir.1986) (same test under section 1983 as under Title VII). A pattern or practice is present when the discriminatory acts were not isolated, insignificant, or sporadic, but were repeated, routine, or of a generalized nature; in other words, discrimination must have been “the company’s standard operating procedure — the regular rather than the unusual practice.” Teamsters, 431 U.S. at 336 & n. 16, 97 S.Ct. at 1855 & n. 16.
The class in this case presented both statistical evidence revealing a disparity between the number of females hired and the number expected to be hired and anecdotal evidence recounting instances of discrimination against specific class members. Either of these types of proof alone may be sufficient to establish a pattern or practice of discrimination, see Hazelwood School Dist. v. United States, 433 U.S. 299, 307-08, 97 S.Ct. 2736, 2741-42, 53 L.Ed.2d 768 (1977) (statistics); Briggs, 796 F.2d at 1019 (anecdotal evidence); thus, neither is “automatically entitled to reverence to the exclusion of the other.” Coates v. Johnson & Johnson, 756 F.2d 524, 533 (7th Cir.1985). For example, while a class claim may fail despite proof of discrimination against one or more individuals, a class claim may succeed despite employer rebuttal of alleged instances of discrimination involving class representatives and testifying class members. Id. at 532-33. The inadequacy of the class’ statistics by themselves to establish discrimination does not foreclose recovery; rather, the question is whether the totality of the evidence makes it more likely than not that the statistical disparity between expected and actual female hires resulted from discriminatory decisionmaking. See Palmer v. Shultz, 815 F.2d 84, 96-97 (D.C.Cir. 1987); see also Bazemore v. Friday, — U.S. -, 106 S.Ct. 3000, 3009, 92 L.Ed.2d 315 (1986). Missouri’s statistical arguments thus will not alone be dispositive, and we need just examine the class’ statistical evidence (the admissibility of which is not challenged on appeal) for whatever weight it might contribute to the class disparate treatment judgments.
In this case the record of anecdotal evidence, read in the light most favorable to the class, reveals numerous instances suggesting Missouri was not receptive to the idea of female maintenance workers. For example, one supervisor, upon being told that a woman had applied for a maintenance position, became angry and said, “You send that [expletive] in here, and now I’m going to have to hire her.” Another female was advised not to apply at a particular location because “the fellow up there is not about to hire a woman.”
The male highway employees responsible for interviews and hiring tended to discount female applicants’ abilities and expressions of interest in maintenance work: one was told the job was “very physical * * for a woman”; one was told, “I don’t think you want this job”; one was referred to a secretarial job; and one was encouraged to return to teaching. In filling out review sheets interviewers discounted farm and other types of experience listed by female applicants while finding similar experience relevant when possessed by male applicants. One woman was deemed “not one to do dirty work” even though at the time of her interview she lived and worked on a farm and her background included work on a chicken ranch. Other review sheets remarked on female applicants’ looks, including, for example, speculation that one woman interviewee was wearing false eyelashes.
During the interviews the male supervisors emphasized the unpleasant aspects of the maintenance job, such as the lack of restroom facilities and the removal of dead animals from the road; questioned whether women might be offended by the language [1266]*1266of male coworkers; and suggested that the female applicants’ husbands might object to their maintenance work. One woman was told she might have to lift a 100-pound snowplow by herself, even though the men usually didn’t have to do it alone. Occasionally the interviews were conducted not privately but in the presence of a number of male workers who would snicker and tell “scare stories” about unpleasant things that had happened on the job. Female applicants had difficulty interesting interviewers in their qualifications; some were never told of the salary or benefits for the maintenance job; and at least two were informed the highway department already had men in mind for the openings for which they were being considered. As a result of their interviews female applicants generally felt they were not being seriously considered for maintenance positions, and one was even so discouraged by a pre-interview phone conversation of this nature that she decided not to attend her interview.
The class’ statistical evidence, which reveals that females constituted up to forty-eight percent of the relevant work force, compared to less than ten percent of Missouri’s maintenance hires during the class period, serves to reinforce the suggestion that these discriminatory incidents were not isolated or sporadic but were representative of Missouri’s standard operating procedure. While Missouri argues that the class in defining the relevant work force failed to consider the actual interest of otherwise qualified men and women in maintenance work, Missouri bore the burden of introducing evidence to show this failure was significant: “[M]ere conjecture or assertion on [a] defendant’s part that some missing factor would explain the existing disparities between men and women generally cannot defeat the inference of discrimination created by [a] plaintiff[’s] statistics.” See Palmer, 815 F.2d at 101 & n. 13 (interpreting Bazemore, supra). Missouri, however, introduced no direct evidence demonstrating that a lesser interest in highway maintenance work on the part of females accounted for the disparity between the percentage of such positions held by females and the percentage of females in the work force defined by the class. Expert testimony from both sides established that Missouri’s statistics showing no discrimination also did not necessarily accurately reflect male versus female interest. The jury and court were entitled to consider the class’ statistics and judge their credibility. See id.
Furthermore, Missouri cannot shield itself from liability by pointing out that it hired 2.6 percent of female maintenance applicants (8 of 312) and only 2.5 percent of male maintenance applicants (89 of 3,566). Victims of a discriminatory policy cannot be told they have not been wronged because other females have been hired. See Connecticut v. Teal, 457 U.S. 440, 455, 102 S.Ct. 2525, 2534-35, 73 L.Ed.2d 130 (1982); see also, e.g., Craik, 731 F.2d at 474 (males and females were selected at equal rates but no female ever prevailed when another candidate was male). While the equal hiring rate constituted relevant evidence, see Teal, 457 U.S. at 454, 102 S.Ct. at 2534, it did not entitle Missouri to judgment as a matter of law. The jury and court were free to weigh the inference of nondiscrimination arising from the hiring rate against the inference of discrimination arising from the class’ anecdotal and statistical evidence.
This evidence, viewed in its totality, was sufficient to support the jury’s verdict and the court’s conclusion that Missouri in hiring highway maintenance workers engaged in a pattern or practice of discrimination against women. See Anderson v. City of Bessemer City, 470 U.S. 564, 573-76, 105 S.Ct. 1504, 1511-13, 84 L.Ed.2d 518 (1985); Gilkerson, 770 F.2d at 136. We uphold the judgments in favor of the class under both section 1983 and Title VII.
III.
Victims of discrimination under Title VII are entitled to be “made whole,” that is, to be restored, as nearly as possible, to the positions in which they would have been absent the discrimination. Franks v. Bowman Transp. Co., 424 U.S. 747, 763-64, 96 S.Ct. 1251, 1263-64, 47 L.Ed.2d 444 (1976); Briseno v. Central Technical Community College Area, 739 F.2d 344, 347 (8th Cir. [1267]*12671984); see 42 U.S.C. § 2000e-5(g). The scope of the relief, such as back pay, to be awarded each member of a prevailing class generally will be determined through individualized hearings, with the court recreating past employment relationships to determine when a particular class member should have been hired. Teamsters, 431 U.S. at 361, 371-72, 97 S.Ct. at 1872-73; see Hameed v. International Ass’n of Bridge, Structural & Ornamental Iron Workers, 637 F.2d 506, 519 (8th Cir.1980).
A court, however, retains equitable discretion to fashion an appropriate remedy when, for example, the number of qualified class members exceeds the number of openings lost to the class through discrimination and identification of the individuals entitled to relief would “drag the court into ‘a quagmire of hypothetical judgments’ ” and result in “mere guesswork.” Segar v. Smith, 738 F.2d 1249, 1289-90 (D.C.Cir.1984) (quoting Thompson v. Boyle, 499 F.Supp. 1147, 1170 (D.D.C.1980) (quoting Pettway v. American Cast Iron Pipe Co., 494 F.2d 211, 260 (5th Cir.1974))), cert. denied, 471 U.S. 1115, 105 S.Ct. 2357, 86 L.Ed.2d 258 (1985). In such a situation a court may, as the district court did in this case, award relief on a classwide, rather than individual, basis by calculating the number of positions for which class members should have been hired and distributing pro rata among all qualified class members the total amount of back pay for which the defendant may be held responsible. E.g., Hameed, 637 F.2d at 520—21; see Catlett II, 627 F.Supp. at 1019. This method of distribution does not contravene the mandate of Firefighters Local Union No. 1784 v. Stotts, 467 U.S. 561, 579-82, 104 S.Ct. 2576, 2588-90, 81 L.Ed.2d 483 (1984), that make-whole relief be limited to actual victims of discrimination, because qualified members of the class certified by the court are presumptively victims of discrimination absent proof, which Missouri has not made in this case, to the contrary. See Teamsters, 431 U.S. at 361-62, 97 S.Ct. at 1867-68. We find no basis for disturbing the district court’s determination to adopt a classwide back pay scheme. See Segar, 738 F.2d at 1290.
Missouri, however, argues the district court committed error in its execution of the classwide remedy by overestimating the number of female maintenance workers who would have been hired absent discrimination. The court relied on District Eight labor force statistics and the class’ definition of the relevant work force to determine Missouri should have hired forty-six female maintenance workers during the class period. Eight females were hired, so the court calculated Missouri’s total class back pay liability based on thirty-eight “female” positions. See Catlett II, 627 F.Supp. at 1019 (referring to labor pool definition in Catlett I, 589 F.Supp. at 933). Missouri argues the court in identifying the number of “female” positions instead should have relied on the percentage of maintenance applicants who were female.
This court has recognized that general population or labor force statistics in certain cases may be superior to applicant flow data for use in fashioning back pay remedies. E.g., EEOC v. Rath Packing Co., 787 F.2d 318, 336-37 (8th Cir.), — U.S. -, cert. denied, 107 S.Ct. 307, 93 L.Ed.2d 282 (1986). It was incumbent on Missouri to demonstrate that the district court in this case abused its discretion in relying on labor force data, see id. at 336, yet Missouri in attacking the calculation of thirty-eight positions merely revives in disguised form its statistical attempts to exculpate itself by arguing lack of female interest. In light of the court and jury findings that Missouri engaged in discriminatory employment practices, we cannot accept a theory that would result in the denial of all back pay. See Albemarle Paper Co. v. Moody, 422 U.S. 405, 417-18, 95 S.Ct. 2362, 2371-72, 45 L.Ed.2d 280 (1975).
Missouri is entitled to foreclose from sharing in the classwide back pay award individual class claimants who were not qualified for maintenance positions, and the district court granted the state’s request to exclude from the recovery ten women who did not timely assert their abilities to operate lightweight motor equipment. See Cat[1268]*1268lett II, 627 F.Supp. at 1018. We find no basis for reversing the district court’s decisions to retain in the class other females challenged by Missouri, and the state cannot now advance individual objections not included in its motion for summary judgment at the relief stage.
Plaintiffs Catlett, Tuter, Kallemyn, and Leembruggen may suggest that since they have lost their individual recoveries, they now should be allowed as class members to share in the classwide back pay award. The district court had no occasion to consider this argument, and it has not been fully briefed on appeal. Thus, we do not decide the question at this time. The district court may consider this issue if raised and disputed on remand (resolution in the plaintiffs’ favor, of course, would not increase Missouri’s total back pay liability beyond the amount attributable to the thirty-eight “female” positions).
The district court’s grant of a class-wide hiring preference, however, raises different concerns than the classwide award of back pay, because the thirty-eight positions which should have gone to women cannot be so easily divided and distributed among the entire class. Cf. Albemarle Paper Co., 422 U.S. at 416, 95 S.Ct. at 2370-71 (Title VII statutory scheme “implicitly recognizes that there may be cases calling for one remedy but not another”). The district court’s hiring order potentially requires Missouri to give preference not to thirty-eight women but to all of the more than 150 qualified class members among whom the back pay award was split (approximately eighty members of the class have expressed continued interest in highway work). To the degree this hiring order benefits a greater number of females than would have been hired absent discrimination, the order exceeds a make-whole purpose and constitutes a form of affirmative relief we find unwarranted in this case under principles to be discussed infra. See generally Berkman v. City of New York, 705 F.2d 584, 595-97 (2d Cir.1983).
Thus, on remand, the district court should either eliminate the hiring preference or limit it to thirty-eight positions to be allocated equitably among the interested and qualified class members. In either instance class members not receiving preferences will be entitled to reapply and receive full and fair consideration for maintenance positions.
IV.
Affirmative relief, in contrast to individual and classwide relief, serves “not to make identified victims whole, but rather to dismantle prior patterns of employment discrimination and to prevent discrimination in the future.” Local 28, Sheet Metal Workers’ Int’l Ass’n v. EEOC, — U.S. -, 106 S.Ct. 3019, 3049, 92 L.Ed.2d 344 (1986) (plurality). Gender-conscious remedies benefiting individuals not identified as victims of the discrimination (e.g., members of the minority group outside the scope of the class) may be necessary for this purpose in some instances, as when an employer has engaged in persistent discrimination; however, affirmative action will not be an appropriate remedy in a majority of Title VII cases. Id. at 3050. A district court before exercising its discretion to order affirmative action should particularly consider the efficacy of alternate remedies. Sheet Metal Workers, — U.S. -, 106 S.Ct. at 3050, 3053, 92 L.Ed.2d 344; id. at 3056 (Powell, J., concurring); see United States v. Paradise, — U.S. -, 107 S.Ct. 1053, 1067, 94 L.Ed.2d 203 (1987) (plurality); id. at 1075 (Powell, J., concurring). In contrast, the district court in this case justified its establishment of hiring and applicant pool goals only with a reference to Missouri’s “long history of discrimination and the comparatively short history of attempts to end such discrimination.” See Catlett II, 627 F.Supp. at 1020.
Supreme Court precedents illustrate the inadequacy of this justification. For example, in Sheet Metal Workers, supra, the defendant union initially had been found guilty of discrimination in a state administrative proceeding in 1964. The federal action had been instituted in 1971 after the union had flouted state administrative and judicial remedies, plus the union by the time affirmative action was decreed had [1269]*1269twice been found in contempt of federal court remedial orders. Similarly, the affirmative remedy challenged in Paradise, supra, was imposed, only when eleven years after the finding of discrimination the defendant employer still had not developed a suitable plan for making promotions. The Supreme Court in each instance emphasized that the lower court had mandated affirmative action only after the defendant had failed to live up to court-ordered commitments to eliminate discriminatory practices. See Paradise, 107 S.Ct. at 1066. Thus, the race-conscious remedies benefiting nonvictims were supported not only by the societal interest in eradicating discrimination but also by the societal interests in ensuring compliance with court decrees and eliminating the effects of delays in compliance. Id. at 1067; see also Sheet Metal Workers, 106 S.Ct. at 3055 (Powell, J., concurring).
Applying these principles to the case before us, we cannot say Missouri has a “long history of ‘foot-dragging resistance’ to court orders” such that simply enjoining the state from further discrimination would clearly be futile. See id. at 3051. The Supreme Court has not yet approved a court-ordered race- or gender-conscious remedy against a defendant which has not been given a chance voluntarily to bring its personnel practices into compliance with a generalized injunction against further discrimination. We decline to take that step here.
Thus, we vacate those portions of the district court’s order establishing a goal of thirty-seven to forty-eight percent female for Missouri’s applicant pool and new maintenance hires. In addition, it will not be necessary for the district court to retain jurisdiction to monitor compliance with its order, and Missouri will not be required to provide to the court the statistics set out in the court’s appendix A. See Catlett II, 627 F.Supp. at 1021. Missouri, however, should maintain all applications, interview sheets, and related records in case further litigation becomes necessary.
The order imposing hiring goals shall be replaced by a generalized injunction prohibiting further discrimination by Missouri in hiring highway maintenance workers, and nothing in our opinion should be construed as diminishing Missouri’s obligation to take effective, affirmative steps to publicize employment opportunities and manifest its receptiveness to female maintenance applicants. Wé particularly observe that the thirty-seven to forty-eight percent female goal selected by the district court reflects the jury’s and the court’s view of the evidence in this case and may prove a useful standard for Missouri in evaluating its own compliance with our injunction. Because we do not understand Missouri to challenge any other portions of the award of injunctive relief, we uphold the district court’s remedial order in all aspects not mentioned.
V.
The class as a prevailing party is entitled to recover attorney fees. See Hensley v. Eckerhart, 461 U.S. 424, 429, 103 S.Ct. 1933, 1937, 76 L.Ed.2d 40 (1983); 42 U.S.C. §§ 1988, 2000e-5(k). Calculation of the amount to be awarded begins with a determination of the hours reasonably expended on the matter plus an appropriate billing rate. Hensley, 461 U.S. at 433 & n. 7, 103 S.Ct. at 1939 & n. 7. The district court here allowed all of the 4,658 hours claimed by class counsel and multiplied those hours by a billing rate varying from $50 per hour for time expended in 1978 up to $100 per hour for time expended in 1986. The court then enhanced this award by fifty percent, citing the “substantial risk” incurred by class counsel “in expending so many billable hours over a nine-year period” and the success of the class in securing hiring preferences and gender-conscious relief. Missouri on appeal argues only (1) that enhancement based on success or risk of loss was inappropriate as a matter of law; and (2) that the district court committed error in failing to disallow those hours representing wheelspinning and duplication, those hours spent unsuccessfully pursuing the individual claims, and those hours spent unsuccessfully opposing Missouri’s move to limit the class, which was originally certified statewide, to District Eight.
[1270]*1270A district court, because of its “special understanding” of the degree of complexity of a case, is best equipped to determine the reasonableness of the hours expended by counsel, and an appeal on this issue will succeed only where the district court abused its discretion or committed error in implementing the controlling legal standards. Moore v. City of Des Moines, 766 F.2d 343, 345-46 (8th Cir.1985), cert. denied, 474 U.S. 1060, 106 S.Ct. 805, 88 L.Ed.2d 781 (1986). The fee award should not compensate counsel for those hours expended in pursuit of unsuccessful claims “distinct in all respects” from the successful claims. Hensley, 461 U.S. at 440, 103 S.Ct. at 1943. Where the hours expended are not easily allocable because the unsuccessful and successful claims were related, the court may merely reduce the award in its discretion to an amount reasonable in relation to the result on the merits, id. at 435, 440, 103 S.Ct. at 1940, 1943, but a plaintiff who has achieved substantial success should receive a fully compensatory fee despite the failure of the district court to adopt every contention raised. Id.
The class argues the district court properly included in the fee award attorney hours spent defending statewide class certification because the issue constituted an incidental contention that, despite its rejection, was reasonably raised in pursuit of the substantial success achieved on the discrimination claims. Cf. Monroe v. United Air Lines, 565 F.Supp. 274, 285-86 (N.D. Ill.1983) (supplemental order) (unsuccessful motion for preliminary injunctive relief); Easley v. Anheuser-Busch, Inc., 572 F.Supp. 402, 416 (E.D.Mo.1983) (unsuccessful motion for partial summary judgment), aff'd in part, rev’d in part on other grounds, 758 F.2d 251 (8th Cir.1985); Bruno v. Western Elec. Co., 618 F.Supp. 398, 404 (D.Colo.1985) (unsuccessful motion to compel discovery); Laffey v. Northwest Airlines, 572 F.Supp. 354, 362 n. 8, 364 & n. 11 (D.D.C.1983) (unsuccessful motion for revision of the postjudgment interest rate), aff'd in part, rev’d in part on other grounds, 746 F.2d 4 (D.C.Cir.1984), cert. denied, 472 U.S. 1021, 105 S.Ct. 3488, 87 L.Ed.2d 622 (1985).
Hensley, however, emphasizes the degree to which work on an unsuccessful claim may also be deemed to have been expended on the successful claim or claims. 461 U.S. at 435, 103 S.Ct. at 1940. While the district court’s decision regarding the scope of the class necessarily took into consideration the facts and legal theories on which the District Eight action was based, the extra efforts expended in attempting to defend statewide certification did not in general contribute to the advancement of the District Eight claim, and attorney time spent studying hiring practices in other districts apparently was easily segregable, see id. The use at trial of a few pieces of evidence uncovered during statewide class discovery is insufficient to justify a fee award covering all of the approximately 643 attorney hours expended on non-District Eight matters. See Webb v. County Bd. of Educ., 471 U.S. 234, 243, 105 S.Ct. 1923, 1928-29, 85 L.Ed.2d 233 (1985). The statewide discovery was made not so much on behalf of the District Eight class in pursuit of its discrimination claims as on behalf of other female highway maintenance applicants throughout the state, and the non-District Eight women were not for the purposes of this litigation prevailing parties entitled to attorney fees. We conclude that attorney hours spent relative to statewide class certification should have been excluded from the fee award in this case. See Bennett v. Central Tel. Co., 619 F.Supp. 640, 646-47 (N.D.Ill.1985); In re Mid-Atlantic Toyota Antitrust Litig., 605 F.Supp. 440, 447 (D.Md.1984); see also Richardson v. Restaurant Mktg. Assocs., 527 F.Supp. 690, 699 (N.D.Cal.1981) (pre-Hensley case).
We also conclude, however, that fees were properly awarded for time spent pursuing the unsuccessful individual claims. A court should not disallow attorney hours related and necessary to successful claims, Carmichael v. Birmingham Saw Works, 738 F.2d 1126, 1137 (11th Cir. 1984), and our discussion in section II, supra, explains the relevance to the class pattern or practice claim of evidence of individual incidents of discrimination. Fur[1271]*1271thermore, the district court did not abuse its discretion in failing to reduce attorney hours for inefficiency and duplication of services. We cannot find the hours claimed by the class excessive when, as illustrated by comments in the record, the pervasive documentary and statistical evidence and exhibits in this case created special burdens for the parties.
Class counsel’s allowable and reasonably expended hours, if multiplied by a reasonable rate, presumptively resulted in an adequate fee award. See Blum v. Stenson, 465 U.S. 886, 897, 104 S.Ct. 1541, 1548, 79 L.Ed.2d 891 (1984). Time and rate calculations reflect most factors relevant to attorney compensation, and absent a specific showing that an hours-times-rate “lodestar” amount is inadequate, a district court which enhances a fee award engages in impermissible double counting. See id. at 898-900, 104 S.Ct. at 1548-50. The class here did not present evidence that the success achieved through the litigation was not reflected in the lodestar factors. See id. at 900 & n. 16, 104 S.Ct. at 1550 & n. 16. Nor did the class demonstrate that enhancement was necessary to attract competent local counsel in light of the risk of loss arising from the class’ contingency fee arrangement. See Pennsylvania v. Delaware Valley Citizens Council for Clean Air, — U.S. -, 107 S.Ct. 3078, 3089, 97 L.Ed.2d 585 (1987); id. at 3090-91 (O’Connor, J., concurring in part). Thus, the district court’s enhancement of the class fee award may not be upheld on either ground challenged by Missouri.
The risk of losing the case, and thus the fee, however, must be distinguished from the loss a prevailing attorney experiences due to the delay in collecting the statutory fee award. Id. at 3081-82. A delay in the receipt of fee payment, because of inflation and lost interest, “dilutes the eventual award and may convert an otherwise reasonable fee into an unreasonably low one.” Johnson v. University College of the Univ. of Ala., 706 F.2d 1205, 1210-11 (11th Cir.), cert. denied, 464 U.S. 994, 104 S.Ct. 489, 78 L.Ed.2d 684 (1983). A fee award based on the rates in effect at the time attorney hours were expended (“historic rates”) will not be fully compensatory unless the award accounts in some way for the loss due to delay in payment. Daly v. Hill, 790 F.2d 1071, 1081 (4th Cir.1986). A court may account for this loss by enhancing the lodestar. Id.
The district court in this case used historic rates in calculating the lodestar award, and we believe the court, in its reference to the nine-year duration of this litigation, was basing its enhancement at least in part on the lost value of the fee due to delay in payment. The class argues that the fifty-percent enhancement may be justified totally on this basis. The average rate of about $71 per hour contemplated by the unenhanced award would have been entirely inadequate, see Paxton v. Union Nat’l Bank, 806 F.2d 785, 786 (8th Cir.1986), and the average rate realized by class counsel after enhancement is still less than $110 per hour, a figure well in keeping with current market rates, cf. Ramos v. Lamm, 713 F.2d 546, 555 (10th Cir.1983) (using current market rates as approximating historic rates adjusted for inflation and lost interest).
Shortly after the district court entered its fee order, however, the Supreme Court held a fee enhancement to compensate for delay in payment equivalent to an award of interest. Library of Congress v. Shaw, — U.S. -, 106 S.Ct. 2957, 2965-66, 92 L.Ed.2d 250 (1986). Because interest is not available in suits against the United States absent “express congressional consent * * separate from a general waiver of immunity to suit,” id. at 2961, the Court disallowed an enhanced fee award made against the federal government under Title VII. At least one circuit has extended the “express waiver of interest” rule to eleventh amendment immunity, thus limiting the use of enhancements for delay in assessing attorney fees against states. Rogers v. Okin, 821 F.2d 22 (1st Cir.1987). Furthermore, the district court here used a similar federal-state analogy in finding the eleventh amendment barred the class’ recovery of prejudgment interest on the back pay award. Catlett II, 627 F.Supp. at 1019-20.
Because the parties have not yet briefed or argued immunity in the context of fee [1272]*1272enhancement or seriously explored the reach of Shaw, we remand to allow the district court to consider thoroughly this important question and make a comprehensive disposition of both the enhancement and prejudgment interest issues. Missouri has not waived any of its rights should immunity be found; the eleventh amendment may be raised at any time. See Edelman v. Jordan, 415 U.S. 651, 677-78, 94 S.Ct. 1347, 39 L.Ed.2d 662 (1974); Ford Motor Co. v. Department of Treasury, 323 U.S. 459, 466-67, 65 S.Ct. 347, 351-52, 89 L.Ed. 389 (1945); Rose v. Nebraska, 748 F.2d 1258, 1262 (8th Cir.1984), cert. denied, 474 U.S. 817, 106 S.Ct. 61, 88 L.Ed.2d 50 (1985).
In sum, the district court on remand should eliminate from the fee calculation hours spent unsuccessfully defending statewide certification and should similarly adjust paralegal hours and other expenses included in the award. If the district court has counted certain taxable costs twice, this correction should be made. Enhancement will not be available if Missouri is protected under the eleventh amendment from awards of interest; however, if Missouri is not immune the district court in its discretion may enhance the class fee award to compensate for delay in payment. If Missouri is not immune the district court should also exercise its discretion regarding prejudgment interest on the back pay award. Missouri, in light of the intervening Supreme Court decision in Crawford Fitting Co. v. J.T. Gibbons, Inc., — U.S. -, 107 S.Ct. 2494, 96 L.Ed.2d 385 (1987), will be free to challenge the assessment against it of the class’ expert witness fees.
VI.
We have carefully considered Missouri’s additional challenges, for example, to the jury instructions and to the district court’s conduct at trial, and we find no reversible error. We need not reach Missouri’s argument regarding the failure of plaintiffs Tuter, Kallemyn, and Leembruggen to file complaints with the Equal Employment Opportunity Commission, see 42 U.S.C. § 2000e-5(e), (f), because we reverse the individual judgments in favor of those plaintiffs. Also, we need not reach Missouri’s argument that under the eleventh amendment the Highway and Transportation Commission was not a proper party to the section 1983 actions; all relief awarded is supportable under Title VII and/or against proper section 1983 defendants. Missouri’s argument regarding the absence at the remedial stage of any nonapplicant claimants mistakes the nature of the class-wide award of relief and cannot justify a reduction in Missouri’s total back pay liability; the identity of the class claimants does not change the number of positions which absent discrimination would have been filled by women. On cross-appeal we review the district court’s class certification ruling for abuse of discretion, see Shapiro v. Midwest Rubber Reclaiming Co., 626 F.2d 63, 71 (8th Cir.1980), cert. denied, 449 U.S. 1079, 101 S.Ct. 860, 66 L.Ed.2d 802 (1981), and we find no abuse in the denial of a statewide class here.
We affirm the Title VII and section 1983 judgments in favor of the class, affirm the section 1983 judgments against the individual plaintiffs, reverse the Title VII judgments in favor of the individual plaintiffs, and remand to the district court for further proceedings consistent with this opinion.