Duke University appeals from the district court's order awarding Heather Sue Mercer almost $350,000 in attorney's fees in her Title IX action against Duke. Duke contends that because Mercer's recovery was ultimately limited to an award of nominal damages, she is not entitled to attorney's fees, or at least not entitled to such a large award of fees. For the reasons set forth below, we affirm the decision of the district court.

Heather Sue Mercer was an all-state place kicker on her high school football team. In 1994, as a freshman at Duke University, Mercer tried out (as a walk-on) for Duke's Division I-A men's football team. She did not make the team, but she did serve as one of the team's managers, and Head Coach Fred Goldsmith allowed her to attend practices and work out with the kickers. Mercer also participated in the winter and spring conditioning programs. In April 1995, the team's seniors selected Mercer to participate in an intra-squad scrimmage game, and Mercer kicked a field goal that won the game for her squad. Shortly thereafter, Goldsmith announced that Mercer was a member of the team.

Not surprisingly, Mercer's game-winning kick and Goldsmith's announcement received an enormous amount of media attention. Mercer was the first female to be a member of a men's Division I-A football team. Duke was quite receptive to the media attention, with Duke's Sports Information Director pressuring Mercer to give interviews and appear on television, which Mercer declined to do. The outside attention, however, apparently caused Goldsmith to question his decision to make Mercer a member of the team. The district court explained the situation in its 2001 order denying Duke's post-trial motion for judgment as a matter of law:

As a result of extensive publicity and widespread interest in Mercer being the first female to make the Duke football team, Goldsmith became concerned that Mercer's presence on the team might have an adverse effect on his players and recruiting. He became more concerned as a result of the publicity that arose from an article published in a Georgia newspaper. The article made light of the fact that Duke had a female football player and caused Goldsmith to express a belated concern that Mercer's presence on the team could be more harmful than helpful.

Mercer v. Duke Univ., 181 F. Supp. 2d 525, 531 (M.D.N.C.2001), vacated in part & remanded, 50 Fed.Appx. 643 (4th Cir.2002). Goldsmith's treatment of Mercer thereafter took a turn for the worse. For example, Goldsmith refused to let Mercer participate in pre-season camp and he refused to let her dress for games or sit on the sidelines with the rest of the team. He made numerous comments that were offensive to her, such as telling her to sit in the stands with her boyfriend and asking her why she was interested in football instead of beauty pageants. Goldsmith eventually cut Mercer from the team, an action he had never taken with any male player.

Mercer brought this action against Duke University, contending that Duke discriminated against her because of her sex, in violation of Title IX. See 20 U.S.C.A. § 1681(a) (West 2000) ("No person in the United States shall, on the basis of sex, be excluded from participation in, be denied the benefits of, or be subjected to discrimination under any education program or activity receiving Federal financial assistance...."). The district court dismissed Mercer's claim, concluding that, by virtue of a regulation creating an exemption for single-sex contact-sport teams, Title IX did not require Duke to give women an opportunity to play on the men's football team.1 And because Duke was not required to give Mercer an opportunity to play, the district court concluded that Title IX did not prohibit Duke from changing its mind once it made Mercer a member of the team. See Mercer v. Duke Univ., 32 F. Supp. 2d 836, 839-40 (M.D.N.C.1998).

Mercer appealed that decision, and this court reversed and remanded for trial. We concluded that while the contact-sport exemption would have shielded Duke from liability had it refused to allow Mercer to try out for the team, the exemption did not give Duke license to discriminate against Mercer because of her sex once Duke decided to allow her to join the team. See Mercer v. Duke Univ., 190 F.3d 643, 647-48 (4th Cir.1999) ("Mercer I").2

At trial, a jury found in favor of Mercer. The jury awarded her one dollar in compensatory damages and two million dollars in punitive damages. Because Mercer was the prevailing party, the district court awarded Mercer more than $380,000 in attorney's fees and costs.

Duke appealed, arguing, inter alia, that punitive damages were not available under Title IX. We held the appeal in abeyance pending the Supreme Court's decision in Barnes v. Gorman, 536 U.S. 181, 122 S. Ct. 2097, 153 L. Ed. 2d 230 (2002). In Barnes, the Supreme Court held that punitive damages are not available for private actions brought under Title VI. See id. at 189-90, 122 S. Ct. 2097. Because Title IX is interpreted consistently with Title VI, see id. at 185, 122 S. Ct. 2097, the Supreme Court's decision in Barnes compelled us to vacate Mercer's punitive damage award. However, we rejected Duke's argument that, as a matter of law, Mercer was no longer entitled to an award of attorney's fees. We concluded that the district court should determine whether, in light of Mercer's substantially more limited success, an award of attorney's fees remained appropriate. See Mercer v. Duke Univ., 50 Fed.Appx. 643, 645 (4th Cir.2002) ("Mercer II").

On remand, the district court concluded that Mercer was entitled to attorney's fees in spite of the fact that her recovery had been reduced to the nominal damage award. The district court first reduced the total amount sought by Mercer's attorneys (more than $430,000) by three percent, a figure intended to approximate the amount of time Mercer's attorneys devoted to the ultimately unsuccessful damages aspect of the case. The district court then reduced the resulting amount by twenty percent, to reflect Mercer's limited degree of success, yielding a final attorney's fee award of $349,243.96.3 This appeal followed.

Congress has authorized the award of reasonable attorney's fees to prevailing parties in certain civil rights cases, including actions brought under Title IX. See 42 U.S.C.A. § 1988(b) (West 2003) ("In any action or proceeding to enforce... title IX of Public Law 92-318 ..., the court, in its discretion, may allow the prevailing party, other than the United States, a reasonable attorney's fee as part of the costs."). A district court's decision to grant or deny attorney's fee under section 1988 is reviewed for abuse of discretion. See Randall v. Prince George's County, 302 F.3d 188, 212 (4th Cir.2002).

"[A] plaintiff `prevails' when actual relief on the merits of his claim materially alters the legal relationship between the parties by modifying the defendant's behavior in a way that directly benefits the plaintiff." Farrar v. Hobby, 506 U.S. 103, 111-12, 113 S. Ct. 566, 121 L. Ed. 2d 494 (1992). That standard is satisfied by a "judgment for damages in any amount, whether compensatory or nominal." Id. at 113, 113 S. Ct. 566. Thus, there is no doubt that the award of nominal damages suffices to make Mercer a prevailing party within the meaning of section 1988(b).

This conclusion, however, means only that Mercer is eligible for, rather than entitled to, an award of attorney's fees. Although Mercer is a prevailing party, the district court has discretion to determine what constitutes a reasonable fee, a determination that requires the court to consider the extent of the plaintiff's success. See Farrar, 506 U.S. at 114, 113 S. Ct. 566 ("Once civil rights litigation materially alters the legal relationship between the parties, the degree of the plaintiff's overall success goes to the reasonableness of a fee award...." (internal quotation marks omitted)). If the prevailing party has recovered only nominal damages, the Supreme Court has explained that "the only reasonable fee is usually no fee at all." Farrar, 506 U.S. at 115, 113 S. Ct. 566 (emphasis added); see also id. ("In some circumstances, even a plaintiff who formally `prevails' under § 1988 should receive no attorney's fees at all. A plaintiff who seeks compensatory damages but receives no more than nominal damages is often such a prevailing party...." (emphasis added)).

Duke contends that, in view of Mercer's very limited degree of success, the only reasonable award is an award of no fees at all. Mercer, however, argues that fee awards are appropriate in some nominal-damage award cases even after Farrar and that the award was justified in this case. Mercer contends her case was an important one that will provide guidance to schools nationwide and that the public purpose served by her case makes the district court's fee award reasonable.

Because the Court in Farrar held that plaintiffs recovering only nominal damages usually or often will not be entitled to an award of attorney's fees, it is clear that such plaintiffs will at least sometimes be entitled to a fee award. Our cases have recognized as much. See Clark v. Sims, 28 F.3d 420, 424-25 (4th Cir.1994) (remanding for district court to consider attorney's fee request by plaintiff who was awarded only nominal damages); see also Mercer II, 50 Fed.Appx. at 645.

Although the majority opinion in Farrar provides little guidance for courts considering whether an award of attorney's fees is warranted, Justice O'Connor in a separate concurring opinion addressed the question in more detail. According to Justice O'Connor, "[w]hen the plaintiff's success is purely technical or de minimis, no fees can be awarded. Such a plaintiff has either failed to achieve victory at all, or has obtained only a Pyrrhic victory for which the reasonable fee is zero." Farrar, 506 U.S. at 117, 113 S. Ct. 566 (O'Connor, J., concurring). Justice O'Connor recognized, however, that not "all nominal damages awards are de minimis. Nominal relief does not necessarily a nominal victory make." Id. at 121, 113 S. Ct. 566. She suggested that when determining whether attorney's fees are warranted in a nominal-damages case, courts should consider "the extent of relief, the significance of the legal issue on which the plaintiff prevailed, and the public purpose served" by the litigation. Id. at 122.

Several circuits have applied the factors identified by Justice O'Connor when considering whether a plaintiff who received nominal damages is nonetheless entitled to an award of attorney's fees under section 1988. See, e.g., Murray v. City of Onawa, 323 F.3d 616, 619 (8th Cir.2003); Brandau v. Kansas, 168 F.3d 1179, 1182 (10th Cir.1999); Morales v. City of San Rafael, 96 F.3d 359, 363 (9th Cir.1996); Briggs v. Marshall, 93 F.3d 355, 361 (7th Cir.1996). When considering the question of attorney's fees in nominal-damage cases, this court has likewise referred to the factors identified by Justice O'Connor, see Sheppard v. Riverview Nursing Center, Inc., 88 F.3d 1332, 1336 (4th Cir.1996); Mercer II, at 645, and the district court in this case analyzed the question under the framework of the O'Connor factors.

We believe that the factors set forth by Justice O'Connor help separate the usual nominal-damage case, which warrants no fee award, from the unusual case that does warrant an award of attorney's fees. Accordingly, we will consider the district court's decision to award attorney's fees by way of the factors identified by Justice O'Connor.

The first factor identified by Justice O'Connor is one that is of primary importance in all cases where a court is asked to award fees to the prevailing party — the extent of the relief obtained by the plaintiff. When considering the extent of the relief obtained, we must compare the amount of the damages sought to the amount awarded. See Farrar, 506 U.S. at 114, 113 S. Ct. 566 (explaining that the "district court... is obligated to give primary consideration to the amount of damages awarded as compared to the amount sought" (internal quotation marks omitted)); see also id. at 121, 113 S. Ct. 566 (O'Connor, J., concurring) ("[A] substantial difference between the judgment recovered and the recovery sought suggests that the victory is in fact purely technical."). Whether we compare the relief obtained by Mercer to the relief she sought in her original complaint (which asserted four causes of action against Duke and Coach Goldsmith and sought compensatory and punitive damages, along with declaratory and injunctive relief) or to the relief sought in connection with the single claim that proceeded to trial (the Title IX claim against Duke), the answer is the same: the relief obtained by Mercer was extremely limited. She sought substantial compensatory and punitive damages, but ultimately received only one dollar in compensatory damages.

The district court, however, concluded that Mercer should not be viewed as having obtained only limited relief, because she obtained the primary relief she sought — a liability finding against Duke. The court explained that

because the primary purpose of Mercer's litigation was to establish that Duke violated her rights, her degree of success cannot be judged solely by the nominal monetary compensation she was awarded.... Because Mercer achieved the primary result that she sought, the Court finds that the degree of success she achieved was not de minimis.

J.A. 580-81. Duke contends that the district court erred when it determined the extent of the relief obtained by focusing on Mercer's subjective motives in pursuing the litigation. We agree.

In Texas State Teachers Association v. Garland Independent School District, 489 U.S. 782, 109 S. Ct. 1486, 103 L. Ed. 2d 866 (1989), the Supreme Court rejected the use of a similar subjective approach when determining whether a plaintiff should be considered the prevailing party. In that case, the plaintiffs prevailed on several, but not all, of the claims asserted against the defendants. The Fifth Circuit, however, denied the plaintiffs' request for attorney's fees, concluding that a plaintiff was a prevailing party under section 1988 only if "the plaintiff prevailed on the central issue by acquiring the primary relief sought." Id. at 787, 109 S. Ct. 1486 (internal quotation marks omitted). The Supreme Court rejected this "central issue" test:

By focusing on the subjective importance of an issue to the litigants, [the central issue test] asks a question which is almost impossible to answer.... This question, the answer to which appears to depend largely on the mental state of the parties, is wholly irrelevant to the purpose behind the fee shifting provisions, and promises to mire district courts entertaining fee applications in an inquiry [that has been] described as "excruciating." ... [T]he search for the "central" and "tangential" issues in the lawsuit, or for the "primary," as opposed to the "secondary," relief sought, much like the search for the golden fleece, distracts the district court from the primary purposes behind § 1988 and is essentially unhelpful in defining the term "prevailing party."

Id. at 791, 109 S. Ct. 1486.

Thus, the Court in Garland made it clear that the subjective motives of the plaintiff are not relevant to determining whether the plaintiff is a prevailing party. And contrary to Mercer's suggestion, we do not believe that Farrar requires consideration of the plaintiff's motives to determine the extent of the relief obtained by the plaintiff.

To be sure, Farrar does require a court determining the scope of the relief obtained to consider, to a certain degree, the purpose of the lawsuit. See Farrar, 506 U.S. at 114, 113 S. Ct. 566 ("Where recovery of private damages is the purpose of civil rights litigation, a district court, in fixing fees, is obligated to give primary consideration to the amount of damages awarded as compared to the amount sought." (emphasis added; internal quotation marks and alteration omitted)). The reference to the "purpose" of the lawsuit, however, merely serves to distinguish an action seeking monetary relief from one seeking injunctive or declaratory relief. That distinction was relevant to the Court's immediately-following explanation of how to determine the extent of the relief obtained by the plaintiff: by comparing the damages asked for to those awarded. See id. If a case sought injunctive relief, the relevant comparison, of course, would be the scope of the injunctive relief sought to the relief actually granted.

Thus, Farrar simply requires courts to consider the relief that was sought by the plaintiff, not the relief that was most important to the plaintiff. See id. at 114-15, 113 S. Ct. 566 ("[A] district court, in fixing fees, is obligated to give primary consideration to the amount of damages awarded as compared to the amount sought. Such a comparison promotes the court's central responsibility to make the assessment of what is a reasonable fee under the circumstances of the case...." (emphasis added; internal quotation marks omitted)); id. at 121, 113 S. Ct. 566 (O'Connor, J., concurring) (noting that when determining whether to award fees, "[t]he difference between the amount recovered and the damages sought is not the only consideration" (emphasis added)). If the rule were otherwise, then every plaintiff recovering only nominal damages would claim that the only thing he was really ever interested in was a liability finding, a claim that the defendant would dispute, thus miring the district court in the same "excruciating" and "distract[ing]" inquiry that the Supreme Court has condemned. Garland, 489 U.S. at 791, 109 S. Ct. 1486; see also Hensley v. Eckerhart, 461 U.S. 424, 437, 103 S. Ct. 1933, 76 L. Ed. 2d 40 (1983) ("A request for attorney's fees should not result in a second major litigation.").

As discussed above, when the relief sought by Mercer is compared to the relief she obtained, it is apparent that Mercer obtained only limited relief.4 The first factor identified by Justice O'Connor in Farrar thus weighs against an award of attorney's fees to Mercer.

According to Justice O'Connor, the second factor we should consider is "the significance of the legal issue on which the plaintiff prevailed." Id. at 122, 113 S. Ct. 566 (O'Connor, J., concurring). This factor is concerned with the general legal importance of the issue on which the plaintiff prevailed. See, e.g., Maul v. Constan, 23 F.3d 143, 145 (7th Cir.1994) ("[W]e understand the second Farrar factor to address the legal import of the constitutional claim on which plaintiff prevailed."); Piper v. Oliver, 69 F.3d 875, 877 (8th Cir.1995) (explaining that the plaintiff's "right to be free from illegal detention was a significant one").5

We agree with the district court that the legal issue on which Mercer prevailed is an important one. Mercer's case established that the contact-sports exemption does not permit a school to discriminate against women that the school has allowed to participate in contact sports. Mercer's case was the first to so hold, and it will serve as guidance for other schools facing the issue.

Duke, however, contends that, when applying this factor, we should not focus on the legal issue resolved in Mercer I. Relying on Hewitt v. Helms, 482 U.S. 755, 107 S. Ct. 2672, 96 L. Ed. 2d 654 (1987), Duke argues that Mercer I was an interlocutory appellate decision and that, for purposes of attorney's fees under section 1988, Mercer did not prevail on the issue at stake in Mercer I. Duke contends that Mercer prevailed only because of the jury's verdict, which was a fact-specific determination of liability that is in no way legally significant. This argument is utterly without merit.

In Hewitt, a plaintiff convinced the Court of Appeals that the district court erred by granting summary judgment in favor of the defendants on the plaintiff's § 1983 claim. On remand, however, the district court concluded that the defendants were entitled to qualified immunity and thus entered judgment in favor of the defendants. The Supreme Court rejected the plaintiff's claim that the Court of Appeals' decision reversing the grant of summary judgment entitled him to attorney's fees. The Court concluded that the plaintiff was not a prevailing party because he "obtained no relief.... [H]e received no damages award[,][and][n]o injunction or declaratory judgment was entered in his favor.... The most that he obtained was an interlocutory ruling that his complaint should not have been dismissed.... That is not the stuff of which legal victories are made." Id. at 760, 107 S. Ct. 2672.

Unlike the plaintiff in Hewitt, Mercer did not win on appeal only to lose at trial. Instead, she won on appeal and she also won at trial. Contrary to Duke's suggestion, the decision in Mercer I was not merely an interlocutory decision that had no bearing on Mercer's victory at trial. Mercer I concluded that the facts alleged by Mercer would, if believed by the jury, amount to a violation of Title IX, and Mercer I thus provided the framework within which the jury would operate. The jury's verdict, of course, does represent a factual determination that Duke was legally responsible for violating Mercer's rights under Title IX. But the facts as found by the jury gave rise to a first-of-its-kind liability determination. Thus, contrary to Duke's argument, Mercer succeeded on a significant legal issue. Accordingly, the second of Justice O'Connor's factors weighs in favor of an award of fees to Mercer.

The final factor we must consider is whether the litigation served a public purpose, as opposed to simply vindicating the plaintiff's individual rights. See Farrar, 506 U.S. at 121-22, 113 S. Ct. 566 (O'Connor, J., concurring) (explaining that a plaintiff's "success might be considered material if it also accomplished some public goal other than occupying the time and energy of counsel, court, and client"); cf. Carter v. Burch, 34 F.3d 257, 266 (4th Cir.1994) (affirming denial of attorney's fee to plaintiff who received only nominal damages, in part because the case "involved no broad civil rights issues").

Title IX prohibits sex-based discrimination by educational institutions receiving federal funds, a prohibition that extends to "all programmatic aspects of educational institutions," including athletics. Cohen v. Brown Univ., 991 F.2d 888, 894 (1st Cir.1993). There can be no doubt that Title IX's goal of reducing gender-based discrimination in education is an important public goal. And as we have already explained, Mercer's case was the first to establish that a school cannot rely on the contact-sport exemption to excuse discrimination against a woman the school has permitted to join an all-male contact-sport team. Thus, Mercer's case was important in that it marked a milestone in the development of the law under Title IX. The case likewise serves a significant public purpose, by furthering Title IX's goal of eliminating discrimination in educational institutions.

To be sure, Mercer did not seek declaratory or injunctive relief that would have extended beyond her own case. That fact, however, while relevant, cannot be dispositive of the question of whether the legal issue in Mercer's case was a significant one or whether the case served an important public purpose. In our legal system, with its reliance on stare decisis and respect for precedent, a case involving the claim of a single individual, without any request for wide-ranging declaratory or injunctive relief, can have a profound influence on the development of the law and on society. Because Mercer's case was the first of its kind, Mercer I and the jury's verdict will serve as guidance to other schools facing similar issues. Mercer's case, therefore, serves much more than her own private interests, notwithstanding the fact that the she did not seek or obtain broad equitable relief that would have reached beyond her individual claim.

Duke, however, insists that Mercer's case is not particularly important, because it is unlikely that many other women will want to play football. Duke also contends that, in any event, schools will be less likely after Mercer's case to allow women to play on all-male teams, because schools now "understand that, by permitting women to participate on such teams, they could unnecessarily open themselves to lawsuits and administrative sanctions." Brief of Appellant at 35. Duke thus seems to suggest that Mercer's action in fact hindered the public interest, by closing off any opportunity women might otherwise have had to participate in contact sports.

It may well be that to avoid the requirement that they treat women fairly once they are on a team, some schools and coaches will decline to allow women to even try out for contact sports. While that would be an unfortunate consequence of Mercer's lawsuit, that possibility does not change the fact that Mercer's lawsuit broke new ground, nor does it diminish the significance of the case.

Moreover, Duke's doomsday predictions about the participation of women in football and other male-dominated sports do not seem to be panning out. A little research reveals that, even after the jury's verdict in Mercer's case, others have continued to hike along the trail that Mercer blazed. For example, in 2001, Ashley Martin, a kicker for Jacksonville State University, became the first woman to participate and score in a Division I-AA football game, kicking three extra points. And in 2003, Katie Hnida, a kicker for the University of New Mexico, became the first woman to score in a Division I-A football game, kicking two extra points. (Hnida transferred to New Mexico from the University of Colorado, where she made the football team as a walk-on kicker. She suited up for Colorado, but never played in a game.) See Woman Kicker Accounts for First, L.A. Times, Aug. 31, 2003, available at 2003 WL 2431227. It is also worth pointing out that in 2003, nearly 3,000 girls in high school (where Title IX is also applicable) played football, and another 10,000 participated in other traditionally male sports like ice hockey and wrestling. See J.A. 487, 516-17. Duke's attempt to minimize the significance of Mercer's case thus falls flat.

Although Mercer ultimately obtained only limited success on her claim against Duke, the effect of the appellate decisions and jury verdict in her case reaches well beyond Mercer herself. And because Mercer's case served a significant public purpose, we agree with the district court that even though Mercer recovered only nominal damages, her victory is not de minimis or purely technical. Cf. Farrar, 506 U.S. at 118, 113 S. Ct. 566 (O'Connor, J., concurring) ("When the plaintiff's success is purely technical or de minimis, no fees can be awarded."). Accordingly, we conclude that the district court did not abuse its discretion by determining that Mercer was entitled to an award of attorney's fees. See Pino v. Locascio, 101 F.3d 235, 239 (2d Cir.1996) (explaining that attorney's fees will be warranted in a nominal-damages case if the "lawsuit create[s] a new rule of liability that serve[s] a significant public purpose"); Sheppard, 88 F.3d at 1336 ("Farrar's concern was not only with whether the extent of recovery accords with the amount of attorney's fees. The decision suggested a more general proportionality consideration as well: whether the public purposes served by resolving the dispute justifies the recovery of fees.").

In addition to challenging Mercer's general entitlement to attorney's fees, Duke also contends that the amount fixed by the district court is unreasonable. Like the question of Mercer's entitlement to fees, the question of reasonableness of the amount awarded is an issue entrusted to the sound discretion of the district court. See, e.g., Carroll v. Wolpoff & Abramson, 53 F.3d 626, 628 (4th Cir.1995) ("It is for the district court in the first instance to calculate an appropriate award of attorney's fees. On appeal, this court has a duty to affirm an attorney's fee award which falls within the district court's broad discretion." (citation, internal quotation marks and alteration omitted)).

When calculating the fee award after remand, the district court started by looking to the fee calculations made in its 2001 post-trial opinion. The district court added to those calculations the number of hours that had been expended since the time of that opinion and multiplied that figure by a reasonable hourly rate to reach the "lodestar" amount. The district court reduced the lodestar amount by three percent, the amount of time that Mercer's attorneys devoted to the issue of damages, and then reduced that amount by another twenty percent, to reflect Mercer's limited success by virtue of her failure to recover anything more than nominal damages.

Duke does not quarrel with the manner in which the district court arrived at the amount of fees to be awarded. Indeed, the district court's approach to the fee calculation was entirely consistent with the procedures set forth by the Supreme Court and this court. See Hensley, 461 U.S. at 436-37, 103 S. Ct. 1933; Johnson v. City of Aiken, 278 F.3d 333, 337 (4th Cir. 2002). Moreover, Duke stipulated below that both the hours claimed by Mercer and the hourly rate applied by the district court were reasonable, and Duke did not dispute Mercer's claim that only three percent of the attorneys' time was attributable to the damages issue.

Nonetheless, Duke contends on appeal that the district court erred by awarding Mercer so much of her requested attorney's fees. Duke, however, offers no explanation as to how or why the district court erred, except to state rather cryptically that "[l]imited success should only beget limited fees." Brief of Appellant at 39. Instead, Duke simply re-asserts many of the arguments it made in support of its contention that Mercer was not entitled to any fees.

Duke supports its nebulous argument with a string of citations to civil rights cases where the plaintiffs received only nominal damages. In most of the cases cited by Duke, the courts refused to award any fees to the plaintiffs; in the few cases where a fee award was made, it was in an amount substantially lower than the amount at issue in this case. Given Duke's failure to make any specific argument about the applicability of any of the cases it cites, Duke thus seems to believe that the mere existence of cases awarding no fees or low fees establishes the error of the district court's ways in this case. We disagree.

The no-fees cases cited by Duke are typical civil rights cases, in that they involved nothing more than a moral vindication of the plaintiff's rights, without establishing any new rule or having any farreaching public effect. See Pino, 101 F.3d at 239 (observing that "[t]he vast majority of civil rights litigation does not result in ground-breaking conclusions of law, and therefore, will only be appropriate candidates for fee awards if a plaintiff recovers some significant measure of damages or other meaningful relief"). Accordingly, the no-fees cases are just that — cases where a court has concluded that, under the circumstances of that case, the only reasonable fee award is no fee at all. See, e.g., Pouillon v. Little, 326 F.3d 713, 717 (6th Cir.2003) ("Pouillon has not demonstrated that his case is distinguishable from the `usual' case where a prevailing civil rights plaintiff is not entitled to attorney's fees when all that he has won is a technical vindication of rights in the form of nominal damages."); Johnson, 278 F.3d at 338 ("As in Farrar, [the plaintiffs'] victory against [a single defendant] did little more than provide them the moral satisfaction of knowing that a federal court concluded that their rights had been violated." (internal quotation marks and alteration omitted)); Briggs, 93 F.3d at 361 (affirming denial of attorney's fees in case where "`plaintiffs' victory served little or no public purpose'"); Carter, 34 F.3d at 266 (affirming denial of attorney's fee to plaintiff who received only nominal damages, in part because the case "involved no broad civil rights issues"). In this case, however, the district court concluded, and we agree, that Mercer's is one of the rare nominal-damage cases warranting an award of attorney's fees. Duke's no-fees cases thus shed no light on the question of whether the fees awarded in this case are reasonable.

The cases cited by Duke where fees were awarded, but in lower amounts, are similarly unhelpful. Duke claims that there are only six reported cases where attorney's fees have been awarded in nominal-damage civil rights cases, and in none of those cases did the fee award even approach the amount awarded by the district court in this case.6 Preliminarily, we note that lower fees were awarded in the cases cited by Duke largely because the fees requested were substantially lower than those requested in this case. From that we can readily infer that the cases were not nearly as long- or hard-fought as this case. In addition, the district courts in these cases reduced the requested fees (to reflect the plaintiffs' limited success) by amounts ranging from ten percent to more than sixty percent, and those decisions were affirmed on appeal. See Murray, 323 F.3d at 619-20 (court awarded 50% of the approximately $15,000 in fees requested); Brandau, 168 F.3d at 1183, affirming No. 96-2414, 1997 WL 613333, at *3 (D.Kan.1997) (reducing requested fees by 10% and awarding approximately $40,000 in fees); Muhammad v. Lockhart, 104 F.3d 1069, 1070 (8th Cir.1997) (counsel sought $6,000 in fees; court awarded $4,500); Piper, 69 F.3d at 876-77 (counsel sought approximately $24,000 in fees and costs; court awarded $8,250); Wilcox v. City of Reno, 42 F.3d 550, 555 n. 5 (9th Cir.1994) (court awarded $66,000 of the more than $110,000 in fees requested); see also Jones v. Lockhart, 29 F.3d 422, 423-24 (8th Cir.1994) (counsel requested $31,000 in fees; district court awarded $25,000, which appeals court reduced to $10,000 after concluding that the plaintiff prevailed on only one claim, not two claims as the district court had concluded). Rather than supporting Duke's vague claim that Mercer's fee award was too much, these cases reinforce the principle that the determination of a reasonable fee is an intensively fact-bound matter for the trial court to determine in the sound exercise of its discretion. We simply cannot discern from these cases any over-arching principle that would require us to set aside the district court's carefully calculated fee award.

We are well aware that the fees awarded by the district court are quite substantial. But we have agreed with the district court that Mercer's victory was not de minimis or purely technical under Farrar, so that Mercer is entitled to an award of fees that is reasonable under the circumstances. Given the nature of this litigation, which was hard fought by both sides, has been on-going for more than seven years, and is now in the middle of its third appeal, we cannot conclude merely from the size of the fee award that the award is unreasonable. And as we have explained, Duke gives us no other basis upon which we could conclude that the award is unreasonable. As noted above, the district court scrupulously followed the framework for determining a reasonable fee award, and Duke does not contend otherwise. Duke does not argue on appeal that the hours or the hourly rate used by the district court are unreasonable, nor does Duke contend that more than three percent of the time of Mercer's attorneys was dedicated to the damages issue. Duke does not even suggest what would be a more appropriate percentage to reduce the award in reflection of Mercer's limited success. Duke, although unable to point to an error by the district court, simply insists that the fees awarded are unreasonable.

We have made it clear that the determination of a reasonable attorney's fee award is a decision for the district court to make, and the district court has broad discretion in that regard:

Had we been the district court in this case, we may well have exercised our discretion differently. The fixing of attorney fees, however, is primarily the task of the district court. We have promoted uniform reasonableness in fixing attorney fees by identifying general standards to guide the decisionmaker.... Beyond providing such standards, however, the appellate courts' role in achieving uniformity on a case-by-case basis is limited. It is important that the district court remain primarily responsible for resolving fee disputes, because it is in the better position to evaluate the quality and value of the attorney's efforts. The very discretion basic to the trial court's duties creates results that inevitably differ in degree. An appellate court, however, cannot quarrel with varying results among independently-minded trial judges that merely reflect differences in their individual judgments. We gauge only whether a trial court abuses its proper discretion.

Ballard v. Schweiker, 724 F.2d 1094, 1098 (4th Cir.1984); see also Hensley, 461 U.S. at 437, 103 S. Ct. 1933 ("[T]he district court has discretion in determining the amount of a fee award. This is appropriate in view of the district court's superior understanding of the litigation and the desirability of avoiding frequent appellate review of what essentially are factual matters."); Plyler v. Evatt, 902 F.2d 273, 277-78 (4th Cir.1990) ("Our review of the district court's award is sharply circumscribed; we have recognized that because a district court has close and intimate knowledge of the efforts expended and the value of the services rendered, the fee award must not be overturned unless it is clearly wrong." (internal quotation marks and alteration omitted)). Although we might have arrived at a different figure, we have no basis upon which we can conclude that the district court abused its discretion when calculating the fees to which Mercer was entitled under section 1988. See Carroll, 53 F.3d at 631 ("Discretion allows district courts latitude in determining fee awards — even those that an appellate court might initially have set in a different amount."). Accordingly, we are constrained to affirm the district court's award of attorney's fees to Mercer.

Mercer's claim was a novel one that established a new rule of law with regard to liability under Title IX. Because Mercer's litigation thus served an important public purpose, we conclude that the district court did not abuse its discretion by concluding that Mercer's victory entitled her to an award of attorney's fees even though she ultimately recovered only nominal damages. And because Duke has identified no error in the district court's calculation of the amount of fees to which Mercer was entitled, we likewise conclude that the district court did not abuse its discretion when fixing the fee amount. Accordingly, the decision of the district court is hereby affirmed.7