These prison inmates assert claims pursuant to 42 U.S.C. § 1983 against their custodians, the prison officials, for permitting them to be homosexually raped by other inmates in violation of their Eighth Amendment rights.' Plaintiffs Frank Led-ferd, David Corder, Hershel Marsh and Jay French appeal from the trial court’s1 denial of their motions for a new trial on damages and for declaratory and injunctive relief. Defendant Denis Dowd cross-appeals, claiming that the trial court erred in denying his motion for judgment notwithstanding the verdict and that the trial court abused its discretion in awarding plaintiffs’ counsel almost $95,000 in attorneys’ fees. Dowd also argues that a jury instruction erroneously set out the standard to determine whether there was an Eighth Amendment violation. We affirm.
I.
The Farmington Correctional Center (FCC) is a near-maximum security facility located in Farmington, Missouri, that houses high-level security inmates serving long sentences for serious crimes. During 1988-89, all four plaintiffs were inmates of FCC.2 Denis Dowd was the Superintendent at FCC from January 1, 1987, through March 30, 1990. On February 17,1989, the four plaintiffs3 filed a complaint under the *664Civil Rights Act, 42 U.S.C. § 1983, against Dowd and four of his staff members, alleging, inter alia, that they had been homosexually raped by other inmates and Dowd had failed to protect them. After counsel was appointed, plaintiffs filed an amended complaint against Dowd, three of his superiors and seven staff members. The complaint alleged six causes of action. Only one cause of action against one defendant, however, was submitted to the jury: the claim that Dowd violated plaintiffs’ constitutional rights by recklessly or knowingly subjecting them to a pervasive risk of sexual assault.4 The evidence presented by both parties focused on whether there was a pervasive risk of sexual assault at FCC, whether Dowd had acted with deliberate indifference or reckless disregard by subjecting the plaintiffs to such a risk and whether Dowd failed to protect the plaintiffs from assault.
The setting in which plaintiffs claim to have been sexually assaulted is as follows. Each 100-foot, two-story wing of the prison, containing sixty-five to seventy-square-foot, double-occupancy, cells, is controlled from a central “bubble.” From inside this bubble, the control officer monitors all activity in the hall and second-floor walkway of the wing. The officer has a view of the doors to the cells. In the door of each cell is a window about two-thirds the size of a legal pad. The window is covered with mesh. There are four microphones spaced twenty-five feet apart in the ceilings of each wing, and if a prisoner calls out from a cell the officer in the bubble will hear him. For example, if the prisoner calls out his cell number, the officer, who controls unlocking the cell doors from the bubble, can press a switch to unlock the cell. The officer cannot hear conversations that take place in a normal tone of voice. There are set periods of time called “open wing” when prisoners can move freely within the wing and visit each other if they wish. During the time period relevant to this suit, the rules applicable to open wing were that all cell doors were to be left open if an inmate who did not live there was visiting, but otherwise the doors had to be closed.5 At 10:00 p.m., after evening open wing, the prisoners are locked into their cells for the night. This is called “lock down.” The guards then conduct a custody count of each cell by switching the cell light on from the outside and looking to see if there are two. prisoners in each cell. They do not look at the prisoners’ faces, so they do not know if the inmates are in the correct cells.
Prisoners who are threatened within the general population of FCC can request protective custody. The policy for the entire Missouri prison system is that an inmate requesting protective custody must spend three days in administrative segregation pending a hearing before the Classification Committee6 before being accorded protective custody. Administrative segregation is a closely-monitored area generally used to punish prisoners for rules violations. Usually the wait in administrative segregation is longer than three days because there is a waiting list for bed space in protective custody. The wait can be up to several weeks. There was testimony from several witnesses about the differences in conditions between administrative segregation and protective custody. For example, prisoners in administrative segregation receive very little of their personal property, no phone calls for the first thirty days, and have five minutes every other or every second day to either shower or clean their cell. The cells do not contain any furniture. Prisoners in protective custody, in contrast, get most of their personal proper*665ty, some phone privileges, an hour a day out of the cell, and the cells contain furniture.
The evidence at trial indicated that the following happened to the plaintiffs after their arrival at FCC. Hershel Marsh, a nineteen-year-old, white, first-offender convicted of child abuse, arrived at FCC in July 1988. According to Marsh’s testimony, prisoner William Stapleton, a black man, approached Marsh within llh hours of his arrival at FCC and told him Stapleton would “fuck” him that night. Marsh was scared, but only told his cellmate about the encounter. That evening Stapleton came to Marsh’s cell and talked to him for “several hours,” but did not make any threats. The next day, Marsh met Stapleton in the gym and agreed to lift weights with him. That night, Marsh went to Stapleton’s cell to discuss a weight-lifting schedule. Sta-pleton closed the cell door and raped him.7 A week later, Stapleton told Marsh to come to his cell that evening. Marsh did so, and Stapleton raped him again. Marsh did not tell anyone about these attacks. In October or November, Stapleton told Marsh to request a cell change so that they could be cellmates. At that time they were living in different wings of the prison, so Marsh only saw Stapleton outside in the yard. Marsh asked for the cell change and his request was granted even though it was prison policy not to assign white and black inmates to the same cells. Although Marsh’s caseworker discussed the request with him, Marsh claimed that Stapleton was present during the interview. The caseworker, Tom Gibson, testified that he talked to Marsh alone about the requested cell change and Marsh told him that there were no problems, that he and Stapleton were weight-lifting buddies and wished to live together. Marsh, on rebuttal, claimed that the one time he and Gibson discussed the cell change without Stapleton present, Gibson had called Marsh to his office on the loudspeaker and, therefore, Stapleton knew that he was talking to Gibson. Marsh testified that Stapleton raped him repeatedly after they became cellmates.,
During this time, Stapleton also raped two of the other plaintiffs, Ledferd and Corder. In both instances, Stapleton told Marsh to switch cells with the victim for the night and Marsh did so. Marsh kept his face hidden when the guards passed so they could not tell that he was in the wrong cell.
Marsh did not report any of these incidents. In November, Marsh attempted to check into protective custody. Marsh testified that he told the prison officials his cellmate was forcing him to have sex. The check-in form filled out .by prison officials recorded only that Stapleton was making Marsh pay his debts to other prisoners. Marsh was put in administrative segrega: tion pending a hearing and available bed space in protective custody. Marsh testified that, shortly afterward, Stapleton, who worked in administrative segregation, contacted Marsh and threatened to harm him and his family if he maintained his allegations against Stapleton and stayed , in administrative segregation. Stapleton provided Marsh with a typed letter stating that his allegations against Stapleton were not true, which Marsh signed. Although prisoners in administrative segregation are not allowed access to pens, paper or typewriters, no investigation into the origin of the letter took place. When his request for return to general population came before the Classification Committee, Marsh reaffirmed that his allegations against Staple-ton were untrue. -He then signed a statement saying that he was not in any danger and- did not need protective custody. Marsh was returned to the same housing unit as Stapleton in general population,8 where Stapleton again raped him. In Janu-, ary 1989, in preparation for this lawsuit, *666Marsh informed FCC officials about the sexual assaults and entered administrative segregation and then protective custody. After he informed FCC officials of the assaults, he was not assaulted again.
Frank Ledferd, an eighteen-year-old first-offender convicted of two counts of sodomy, entered FCC in October 1988. According to his testimony, approximately two weeks after his arrival at FCC he was sexually threatened by an inmate whose identity he did not know. He approached prisoner Rob Johnson, whom he had seen in the other inmate’s company, because he wanted to discuss the threat and try to “straighten things out.” Johnson closed the door of the cell and raped him. The next day Ledferd checked into protective custody and was placed in administrative segregation. He told the prison official who checked him in that he was being pressured for sex, but he did not tell the official about the assault. He claimed that the Classification Committee refused to assign him to protective custody because he ■couldn’t name the individuals who were threatening him. Ledferd signed a statement saying that he was not in any danger and did not need protective custody. The Committee returned him to general population.
Ledferd then discussed the situation with his cellmate, Corder, who told him to go to Stapleton because Stapleton was “ok” and “not like all the rest.” On Corder’s advice, Ledferd went to Stapleton for help. Sta-pleton closed the door on the cell and told him he would be staying there that night. When the guards came by, Stapleton told him to lie facing the wall so the guards could not see his face and he did so. Sta-pleton then raped him. After that, about once every week, Stapleton would tell him to come to Stapleton’s cell, where Stapleton would rape him. In January 1989, in preparation-for this lawsuit, Ledferd informed FCC officials of the assaults. The prison staff checked him into administrative segregation pending bed space in protective custody, but Ledferd requested to be returned to general population two weeks later. In April, he entered protective custody at the insistence of the prison staff.
David Corder, a twenty-three-year-old first-offender convicted of burglary and sodomy, arrived at FCC in July 1988. According to his testimony, a few days after he arrived at FCC, he went to Stapleton’s cell because Stapleton told him to go there. Stapleton told Corder that he was going to spend the night with Stapleton. Corder was there for several hours before lock down, but he did not call out or try to leave. Marsh, Stapleton’s cellmate, was in and out several times getting things so that he could stay in Corder’s cell for the night. When the guards came around later that evening, Stapleton told him to sit on a footlocker facing away from the door. He did this. Afterward, Stapleton raped him. A few days later, another inmate came to Corder’s cell and threatened him. Corder hit the other inmate and locked the man in his cell. He then found a guard and told the guard about the incident. Corder was put into administrative segregation for punishment for thirteen days. While there, he requested protective custody. Stapleton threatened him, however, and made him sign a statement that he and the inmate who threatened Corder in his cell were not enemies and he did not need protective custody. Corder was then returned to general population. In January 1989, in preparation for this lawsuit, he informed FCC officials of the sexual assault. Corder was not sexually assaulted after that time. He has refused to enter protective custody, however, because he has to go through administrative segregation first.
Jay French, twenty-one and convicted of burglary, arrived at FCC in December 1988. French had been in jail before, but not in prison. According to French’s testimony, when he arrived at FCC he bought a television from prisoner George Anderson, contrary to prison rules. Although he paid for the television, Anderson claimed that he had not and told him he would have to pay “homosexually.” French attempted to check into protective custody, complaining that he was being pressured for sex, and was put in administrative segregation pending a hearing and available bed space in protective custody. He requested to be *667returned to general population, however, because he learned that Anderson was going to follow him into protective custody and he was afraid Anderson would kill him.9 Soon after his return to general population, Anderson’s “associate,” Darryl Saddler, raped French in French’s cell. Later, Anderson approached French. When French started to cry, Anderson only required French to touch his penis. A few days later, however, Anderson told French to come to his cell. French went, and Anderson forced him to perform oral sex on Anderson. In January 1989, in preparation for this lawsuit, French informed FCC officials of the assaults. He has not been assaulted since that time. During his testimony, French stated that he did not think Dowd intended to hurt him and that Dowd probably didn’t know about French’s specific situation until French had made his statement. He believes, however, that Dowd must have known about the “general situation.”
Plaintiffs introduced several pieces of evidence intended to show that Dowd must have known about the threat of sexual assault at FCC. Plaintiffs contended that there is a certain profile that can be used to identify prisoners who are likely to be victimized by other prisoners and that they fit this profile. They also argued that Dowd knew this, and that he either knowingly or recklessly failed to take steps to protect vulnerable prisoners from others whose records show that they are aggressive and violent.10 Dr. Ronald Sable, plaintiffs’ expert witness, testified that younger inmates, who are smaller and slightly built, who are first offenders or naive with respect to the correctional system, and who have generally passive personalities, tend to be victimized. Dr. Sable testified that he believed plaintiffs ■ fit this description. He stated that, given the record of the men who assaulted plaintiffs, the administration should have known that this was a very real danger. Plaintiffs introduced into evidence an affidavit signed by Dowd to show that Dowd knew the plaintiffs were at risk. The affidavit read, in relevant part:
Plaintiffs in this action are typical of the. type of inmate who normally receives such intimidation from other inmates. Plaintiffs exhibit generally passive personalities and are unable to stand up to these inmates threatening them. It is likely that plaintiffs will be subject to such treatment by other inmates wherever they go.
Aff. of Denis D. Dowd, Appellant’s App. at 61.
Richard Jones, an employee of FCC, testified that he had read approximately 100 reports of sexual assaults at FCC over a three^year period beginning in 1987. The evidence showed, however, that Dowd had never referred a case of sexual assault for prosecution and had referred only a very few cases for investigation. Plaintiffs’ assaults were not investigated after plaintiffs reported them in January 1989. Dr. Sable testified that, according to documents he reviewed, there were fourteen reported rapes within an eighteen-month period at FCC, and he believed the incidence of unreported rapes to be much higher. Doyle Kirkman, a prisoner at FCC, testified that he had overheard conversations in the presence of guards where some inmates demanded sexual favors, money, and clothing from other inmates and discussed sexual activity on the prior night with their “wives.” Dowd was not present during these conversations, but he testified that *668his staff was instructed to report any such major violations to him.
FCC employees testified that they had had no training or instruction about how to handle sexual abuse or assault problems between prisoners. The orientation program for new prisoners does not contain a component relating to sexual abuse of prisoners by other prisoners; it neither warns prisoners of the occurrence of such abuse nor apprises them that such abuse will be punished. Plaintiffs claimed that the orientation program also did not inform them about how to obtain protective custody. They testified that they did not know how to request protective custody until other inmates told them.
In addition, plaintiffs introduced evidence that at the time relevant to this suit, prisoners were assigned to cells randomly and there was no effort to separate passive inmates from aggressive inmates. Dr. Sable testified that he believed it was reckless to randomly assign people to cells.
Dr. Sable also testified that requiring prisoners to spend time in administrative segregation before they are assigned to protective custody is a heavy burden on inmates seeking protective custody. FCC employee Jones estimated that more than one inmate per month decided not to go into protective custody because they had to go through administrative segregation.
Denis Dowd testified that he had never read in the professional literature about specific characteristics that indicated some inmates are more prone to sexual assault than others. He stated that the affidavit he signed referred to verbal intimidation and was in response to plaintiffs’ demand for transfers to another prison, not their allegations of sexual assault.11 Dowd said that he believes FCC has fewer rapes than other prisons, and he does not know what further steps the prison could take to stop these occurrences. Dowd said that he has never referred a reported prison rape for prosecution because he has never had a reported case that included physical evidence or an eyewitness. He did not account for the fact that there was no investigation of plaintiffs’ rapes after they were reported in January 1989. He explained that he approved the return of prisoners to the general population after they claimed they were being pressured for sex if the Committee recommended that course of action. Despite the Committee review process, prisoners who insist they need protective custody are not returned to general population.12 If an inmate insists that he wants to return to general population, however, and signs a statement saying he is not in any danger, there is very little the prison staff can do to help that prisoner. He said that the inmates have to report incidences before the prison staff can help them, and these plaintiffs did not report that they had been assaulted until January 1989. He emphasized that none of them had been assaulted after they had informed the FCC staff of the problem.
Dowd testified that it is not prison policy to deny an inmate protective custody because he will not or cannot name his enemies. Dowd also testified that Stapleton could not have gotten close enough to either Marsh’s or Corder’s cells in administrative segregation to be able to intimidate them.
In response to Dr. Sable’s testimony about assigning prisoners to cells at random, Dowd pointed out that none of the plaintiffs except Marsh were assaulted by their assigned cellmate, and Marsh had requested a cell change to live with his attacker.
Several prison officials testified that it was standard procedure to inform incoming inmates about the availability and procedures for obtaining protective custody. The evidence showed that all four plaintiffs signed forms when they arrived at FCC *669saying they did not need protective custody. '
After a four-day trial, the jury found in favor of plaintiffs, but awarded them only $1 each in nominal damages. Plaintiffs moved for a new trial on damages and requested that the district court grant them injunctive and declaratory relief. Dowd moved for judgment notwithstanding the verdict (j.n.o.v.). The district court denied all post-trial motions and awarded plaintiffs attorneys’ fees in the amount of $94,680, pursuant to 42 U.S.C. § 1988.
Plaintiffs appealed, arguing that the district court erred in denying them a new trial on the issue of damages and in denying their requests for declaratory and in-junctive relief. Dowd cross-appealed, arguing that the district court erred in denying his motion for j.n.o.v. Dowd also argued that the district court abused its discretion when it awarded plaintiffs attorneys’ fees because it failed to consider the plaintiffs’ lack of overall success. A panel of this court heard argument, and held that plaintiffs were entitled to a new trial solely on the issue of damages and a further hearing on injunctive relief. The panel left the question of attorneys’ fees undecided pending the new trial on damages. It affirmed the trial court on all other issues. Dowd moved for rehearing and rehearing en banc. We granted the motion for rehearing en banc and vacated the panel opinion. We now affirm the trial court on all issues.
II.
A. New Trial on Damages
Plaintiffs claim that the jury’s award of nominal damages was inadequate as a matter of law. Specifically, plaintiffs argue that the evidence shows that they were raped, and that they suffered physical and emotional distress as a result. Indeed, this court has recognized that rape is devastating. Vosburg v. Solem, 845 F.2d 763, 768 (8th Cir.), cert. denied, 488 U.S. 928, 109 S.Ct. 313, 102 L.Ed.2d 332 (1988). Only when a jury’s award is so inadequate as to constitute “plain injustice” or a “shocking” result, however, may we overturn it. Solomon Dehydrating Co. v. Guyton, 294 F.2d 439, 447-48 (8th Cir.), cert. denied, 368 U.S. 929, 82 S.Ct. 366, 7 L.Ed.2d 192 (1961). When making-this determination, “a comparison of verdicts in other cases is of limited value as each case must be evaluated according to the evidence peculiar to that case.” Century “21” Shows v. Owens, 400 F.2d 603, 611 (8th Cir.1968); Perry v. Bertsch, 441 F.2d 939, 944 (8th Cir.1971).
Plaintiffs contend that the only explanation for an award of nominal damages to all four plaintiffs is the jury’s bias against their status as prisoners; thus, the award constitutes “plain injustice.” We disagree. The plaintiffs’ status as prisoners is not the only explanation for the jury’s award. The jury rationally could have concluded that many of the plaintiffs’ injuries would have occurred even if the defendant’s conduct had met constitutional standards. Moreover, plaintiffs failed to produce at trial objective medical evidence supporting their physical injuries or detailing the extent of their emotional injuries. The amount of damages, therefore, depended largely upon the credibility of the plaintiffs’ testimony concerning their injuries. We find that based upon these factors, the jury’s decision to award only nominal damages was not “plain injustice” or so “shocking” as to require a new trial on the issue of damages.
The jury could have awarded nominal damages because it concluded that the plaintiffs’ actions, not those of the defendant, were the cause in fact of most of the plaintiffs’ injuries. In order to establish a violation of constitutional rights under § 1983, the plaintiff must prove that the defendant’s unconstitutional action was the “cause in fact” of the plaintiff’s injury. Carey v. Piphus, 435 U.S. 247, 263, 98 S.Ct. 1042, 1052, 55 L.Ed.2d 252 (1978); Cowans v. Wyrick, 862 F.2d 697, 703 (8th Cir.1988) (McMillian, J., concurring). Conduct is the cause in fact of a particular result if the result would not have occurred but for the conduct. Similarly, if the result would have occurred without the conduct complained of, such conduct cannot be a cause in fact of that particular result. Carey, 435 U.S. at 263, 98 S.Ct. at 1052; *670see W. Page Keeton, Prosser and Keeton on the Law of Torts, § 41, at 264 (5th Ed.1984). In the context of this case, plaintiffs have to prove that the defendant’s unconstitutional conduct was the cause in fact of the sexual assaults they suffered. If, however, an assault would have occurred even if the defendant had not been “deliberately indifferent to a pervasive risk” of sexual assault, the defendant cannot be liable for that assault.
The jury did find that the defendant’s unconstitutional actions were the cause in fact of at least one of the attacks on each plaintiff. The jury could,. however, have found the defendant’s unconstitutional actions were not the cause in fact of many of the plaintiffs’ injuries and still have returned a verdict for plaintiffs. The evidence presented to the jury detailed numerous assaults against each plaintiff. The jury did not, however, have to believe that the defendant’s actions caused each attack in order for it to find for each plaintiff. Each attack was a separate event, and the relevant causal inquiry is whether that attack would have been prevented if the defendant’s actions had met constitutional standards.
Furthermore, the jury charge would allow the jury to so find. The jury specifically was instructed that it could find the defendant liable to each plaintiff if it found one incident of assault for which the defendant was responsible. Tr., Vol. V at 47. Moreover, the jury was charged that it could award damages only for those injuries that were a direct result of the defendant’s actions. Tr., Yol. V at 48.
Because the jury returned a general verdict, we cannot tell which attacks or injuries the jury concluded were caused by the defendant’s unconstitutional actions, or whether the jury believed that certain of the attacks would have occurred anyway.13 We also cannot tell what evidence the jury found necessary to conclude that the defendant breached his constitutional duty. Given this uncertainty, the jury could likely have believed that the plaintiffs’ actions, and not the defendant’s unconstitutional conduct, caused many of their injuries.
Ample evidence in the record supports such a conclusion. First, in many of the assaults complained of, the plaintiff went to the cell of his attacker. The assailant did not lead him there. Instead, on several occasions a significant period of time elapsed before the plaintiff walked by himself to his assailant’s cell. The record demonstrates that on these occasions, the plaintiff failed to notify prison officials of his situation. Plaintiffs declined to pursue their requests for protective custody and returned to the general population during the period in which they were being assaulted. Plaintiff Marsh actually assisted in the rapes of two of the other plaintiffs by switching cells with them and not showing himself to the guards. Three of the plaintiffs failed to reveal themselves to guards when they were in the wrong cell; each subsequently was assaulted. Moreover, one of the plaintiffs’ main contentions is that prison officials’ failure to classify them as “passive” and to segregate them from violent prisoners led to their assaults. They claim that the defendant’s policy of assigning inmates to cells at random put them at a risk of harm. The record demonstrates, howeyer, that none of the assaults • occurred in the plaintiffs’ own cells at the hands of their randomly assigned cellmates. The jury could have concluded, based on this evidence, that many of the attacks would have occurred even if the defendant’s conduct met constitutional standards.
The record further reflects that once plaintiffs did inform prison officials that they had been sexually assaulted (as opposed to simply being threatened), they were placed in administrative segregation or protective custody. No plaintiff ever *671was sexually assaulted while in administrative segregation or protective custody. From this evidence, the jury may have concluded that although the defendant’s actions caused the first assault as to each, the plaintiffs’ failure to seek protection was the cause in fact of all further assaults they faced.
As additional support for the nominal damages award, the jury could have concluded that plaintiffs failed to prove their damages for the injuries it did find to be caused by the defendant’s wrongful conduct. In addition to holding that a jury award of damages must be “shocking” to be inadequate, we also believe that credibility questions concerning damages should be decided at the district court' level. We have stated that:
[Inadequacy or excessiveness of a verdict is basically, and should be, a matter for the trial court which has had the benefit of hearing the testimony and of observing the demeanor of the witnesses and which knows the community and its standards; that this is a responsibility which, for better working of the judicial process and for other seemingly obvious reasons, is best placed upon its shoulders ....
Taken Alive v. Litzau, 551 F.2d 196, 198 (8th Cir.1977) (citing Solomon Dehydrating Co. v. Guyton, 294 F.2d 439, 447-48 (8th Cir.), cert. denied, 368 U.S. 929, 82 S.Ct. 366, 7 L.Ed.2d 192 (1961)). The court went on to state:
An appellate court should be extremely hesitant to overturn a jury verdict which includes damages for pain and suffering. ... There is no precise or exact measuring stick for calculating general damages for pain and suffering. Although the jury should not pick a figure out of the air, exact compensation for pain and suffering is impossible.
Id.
These dictates apply with particular force to the present case. In order to recover compensatory damages in a § 1983 action, the plaintiff must prove them. See Carey, 435 U.S. at 263, 98 S.Ct. at 1052; Cowans, 862 F.2d at 700. In this case, nearly all of the evidence of the plaintiffs’ damages consists of their own testimony. The issue of damages thus depends upon the credibility of witnesses, a matter particularly within the province of the jury.
We first note that an extensive review of the record demonstrates that none of the plaintiffs provided medical testimony regarding the extent of their physical injuries. Similarly, medical proof of the plaintiffs’ emotional distress was limited to the introduction of one-page medical services statements in which a prison psychologist (not a medical doctor) stated merely that two of the plaintiffs, Marsh and Ledferd, suffered from “post traumatic stress disorder,” and that he had prescribed sleeping pills for them. This psychologist did not testify. Plaintiffs provided no testimony explaining what “post traumatic stress disorder” is or what its effects are. Moreover, the evidence regarding Corder’s and French’s injuries is even more sparse. They provided no medical evidence whatsoever supporting either their physical or emotional injuries.
Plaintiffs’ medical expert, Dr. Ronald Sable, likewise provided little support concerning the extent of the plaintiffs’ damages. He testified that he had not examined any of the plaintiffs. Thus, he could not state whether these plaintiffs were in fact damaged by their experience. Moreover, his testimony concerning rape trauma syndrome does not even give the jury enough information to conclude whether plaintiffs suffered from its effects.14 Thus, after hearing the plaintiffs’ sole medical expert at trial, the jury had no basis from which to conclude that plaintiffs were damaged. The record is bereft of any med*672ical testimony either by hypothetical or otherwise based upon reasonable medical probability.
When medical testimony is absent, proof of the nature and extent of a plaintiffs injuries necessarily depends upon the credibility of the plaintiffs testimony. In addition to the medical testimony recited above, the jury heard the following concerning the plaintiffs’ injuries:
Marsh stated that he suffered from ulcers and lack of sleep following repeated assaults by Stapleton. Ill Trial Transcript at 182 [hereinafter Tr.].' A prison psychiatrist prescribed Marsh sleeping pills. Ill Tr. at 196. Inmate Ivron Butler testified that he observed Marsh throwing up several times after being raped. II Tr. at 133. Ledferd described suffering from anal' bleeding for a week and a half after Stapleton raped him, and stated that he suffers from torment and depression continuously. Ill Tr. at 124-25. Corder stated that he was unable to urinate or defecate, and lost his appetite after Stapleton raped him; he also testified that he suffered headaches after Stapleton hit him on the head during an assault. Ill Tr. at 194. French stated that Anderson beat him during an assault. Ill Tr. at 157. Inmate Doyle Kirkman stated that French broke down in tears while describing his ordeal. .II Tr. at 161.
Dissenting op. of Bright, J., infra at 682. Given the lack of medical testimony, the jury’s determination of the extent of the plaintiffs’ injuries must have depended upon their evaluation of this testimony.
There is considerable evidence in the record from which the jury could have concluded that the plaintiffs’ testimony concerning their injuries was not credible and could have disbelieved the extent of their injuries. Despite the plaintiffs’ claims regarding the extent of their mental torment and suffering, all failed promptly to report the incidents-. Marsh even asked to room with his assailant after being attacked. All of the plaintiffs left administrative segregation and returned to the general population. When they did so, each signed a statement providing that “I do not feel my safety to be threatened or my life to be in danger. Therefore, no, I do not need protective custody.” 15 Moreover, three of the plaintiffs stated that they did not want to check into administrative segregation because they would lose radio and television privileges and would have limited spending money.16 From this evidence, the jury reasonably could have disbelieved the plaintiffs’ testimony concerning the injuries received from those attacks it found to be caused by the defendant. Because the proof of the plaintiffs’ injuries is limited almost solely to their own testimony, and thus the issue of damages limited almost exclusively to that of credibility, this court should not upset the determination of the jury and the trial court, which were in a much better position to determine the credibility of witnesses.
After concluding that the jury could have disbelieved the plaintiffs’ testimony regarding the extent of their injuries, the award in this case is not aberrant. This court has approved nominal damages in other cases in which the plaintiffs suffered apparently quantifiable harm. See Warren v. Fanning, 950 F.2d 1370 (8th Cir.1991) (nominal damage award upheld where denial of proper medical treatment resulted in surgical removal of toenails, surgical removal of rod in leg, and surgery on heel), petition for cert. filed, — U.S. -, 113 S.Ct. 111, 121 L.Ed.2d 68 (1992); Cowans v. Wyrick, 862 F.2d 697 (8th Cir.1988) (nominal damage award appropriate when guard slammed door on inmate’s hand resulting in pain, swelling and bruising if jury unable to place monetary value on harm; jury free *673to disbelieve the extent of swelling or other compensable injury); Williams v. Mensey, 785 F.2d 631, 639 (8th Cir.1986) (nominal damages not inadequate as a matter of law when the plaintiff complained that he had been punched in the head with enough force to hit the wall and claimed to suffer headaches, back pain, and blurred vision as a result); cf. Haley v. Wyrick, 740 F.2d 12 (8th Cir.1984) (nominal damage award to plaintiff stabbed thirty-two times by five other prisoners not so inadequate to require its reconsideration in the absence of a motion for a new trial). Although we do not necessarily equate the pain and suffering of a rape victim with the injuries suffered by plaintiffs in the foregoing cases, given that the jury could reasonably refuse to believe the plaintiffs’ testimony regarding the extent of their injuries, we do not find the present award “shocking”' or a “plain injustice.”
B. Equitable Relief17
1. Declaratory Relief
Plaintiffs claim that the district court erred by denying their motion for declaratory relief. They argue that, because the jury found for them on the liability issue, they are entitled to declaratory relief.
In order for the court to have jurisdiction to grant declaratory relief, there must be an “actual controversy” between the parties. Caldwell v. Gurley Ref. Co., 755 F.2d 645, 648 (8th Cir.1985). The test is whether there is a substantial controversy between parties having adverse legal interests, of sufficient immediacy and reality to warrant the issuance of a declaratory judgment. Id. at 649.
In this case, there was no actual controversy left to resolve through declaratory relief when that issue was submitted to the district court. The plaintiffs’ only requested declaratory relief mirrored what the jury was told it must find in order to hold the defendant liable.18 The jury verdict, therefore, amounts to an implicit declaration of the same things that plaintiffs are requesting in their motion for declaratory relief. Plaintiffs admitted this in oral argument.19 Once the jury’s verdict was reached, declaratory relief could not change or clarify any legal relationship between plaintiffs and defendant. See International Wood Processors v. Power Dry, Inc., 593 F.Supp. 710, 736 n. 15 (D.S.C.1984), aff'd, 792 F.2d 416 (4th Cir.1986) (“[Gjranting declaratory relief would serve no useful purpose at this stage because the legal obligations of the parties have been clarified by the jury’s resolution of the legal claims.”). Therefore, the issue is moot and neither this court nor the district court has jurisdiction to grant declaratory relief.
For the foregoing reasons, we affirm the district court’s denial of plaintiffs’ motion for declaratory relief.
2. Injunctive Relief
Corder claims that the district court erred in denying his motion for injunctive *674relief.20 In order for him to obtain injunc-tive relief, he must show “some substantial likelihood that past conduct alleged to be illegal will recur.” Sterling v. Calvin, 874 F.2d 571, 572 (8th Cir.1989). A showing that unconstitutional practices have taken place in the past is not enough: He must show that such practices are likely to affect him in the future. See id. Corder has not made any such showing. Although he claims that he is still subject to threats from fellow inmates, Appellants’ App., Tab 7 p. 511146, Pis.’ First Amended Compl., he does not claim that he is still subject to sexual assault. The testimony was that none of the plaintiffs has been sexually assaulted since January 1, 1989. Trial Tr., Vol. I pp. 25 et seq.; Appellants’ App., Tab 7, Pis.’ First Amended Compl. In fact, the evidence demonstrates that Corder feels' secure enough from assault that he filed a grievance about a policy change that was apparently designed to prevent such assaults from occurring.21 Corder thus has not shown a substantial likelihood that the unconstitutional practices he complained about at trial will recur or affect him in the future. The injunctive relief Corder requests, therefore, will only benefit other inmates, particularly new inmates. Injunc-tive relief, however, must be carefully tailored to remedy the alleged specific harm to the parties involved. See United States v. Articles of Drug, 825 F.2d 1238, 1247-48 (8th Cir.1987); see also Orantes-Hernandez v. Thornburgh, 919 F.2d 549 (9th Cir.1990); Brown v. Trustees of Boston Univ., 891 F.2d 337 (1st Cir.1989), cert. denied, 496 U.S. 937, 110 S.Ct. 3217, 110 L.Ed.2d 664 (1990) (classwide relief appropriate only where there is a properly certified class). This is not a class action. Corder, therefore, is not entitled to the injunctive relief he requests.
In addition, much of the injunctive relief Corder requests in this case is beyond the power of the superintendent of an individual institution to grant. He requests, inter alia, a transfer to the Missouri Eastern Correctional Center,22 orientation programs on sexual assaults for new inmates, a different system to classify and segregate some inmates from others, and no temporary stay in administrative segregation before entering protective custody. All these aspects of institutional life are governed by a centralized authority, the Missouri Department of Corrections, and the superintendent has no power to change centralized decisions. Appellee — Cross-Appellant’s Supplemental App., Tab 3 at 14-16, Aff. of George A. Lombardi, Director of Adult Institutions for the Missouri Department of Corrections and Human Resources.
For the foregoing reasons, we affirm the district court’s denial of the motion for injunctive relief.
C. Denial of J.N.O.V. •
Dowd argues that the district court erred when it denied his motion for j.n.o.v. because plaintiffs failed to prove that they faced a pervasive risk of sexual assault at FCC. He contends that plaintiffs failed to establish that sexual assaults occurred at FCC with such frequency as to reasonably apprise FCC officials of the risk of attack. The jury found that Dowd violated plaintiffs’ Eighth Amendment right to be free from cruel and unusual punishment by knowingly or recklessly subjecting plaintiffs to a pervasive risk of sexual assault. See Jury Instruction 11, supra; see also Bailey v. Wood, 909 F.2d 1197, 1199 (8th Cir.1990); Vosburg, 845 F.2d at 765; Martin v. White, 742 F.2d 469, 474 (8th Cir.1984); cf. Wilson v. Seiter, — U.S. -, -, 111 S.Ct. 2321, 2326-27, 115 L.Ed.2d 271 (1991).
In reviewing the district court’s denial of Dowd’s motion for j.n.o.v., we apply the same standard as the district court. Mor*675gan v. Arkansas Gazette, 897 F.2d 945, 948 (8th Cir.1990) (citing Cleverly v. Western Elec. Co., 594 F.2d 638, 641 (8th Cir.1979) (per curiam)). Therefore, we must consider the evidence in the light most favorable to the prevailing party (the plaintiffs), assume that the jury resolve,d all conflicts in the evidence in favor of the plaintiffs, assume as proved all facts that the plaintiffs’ evidence tends to prove, give the plaintiffs the benefit of all favorable inferences that can reasonably be drawn from the facts proved, and affirm the denial of the motion if reasonable persons could differ about the conclusions to be drawn from it. Id. Generally, “only the evidence favoring the nonmoving party (usually, as here, the plaintiff) should be considered.” Id. (quoting Dace v. ACF Indus., 722 F.2d 374, 376 (8th Cir.1983)).
When we consider the evidence presented at trial in this case under this stringent standard of review, we find that there was sufficient evidence to support the jury’s finding that Dowd subjected plaintiffs to a pervasive risk of sexual assault in violation of their Eighth Amendment rights. Plaintiffs presented evidence that there were at least 100 sexual assaults reported at FCC during a three-year period. During that same time frame, however, there were almost no sexual assaults assigned to the prison investigator for investigation, and Dowd never referred even one sexual assault for prosecution. There was no warning to prisoners about the risk of sexual assault. The prison personnel had received no training or instructions about how to handle sexual assaults or how to prevent them. Prisoners were assigned to general population at random, with no segregation based on their vulnerability to assault. Dr. Sable, plaintiffs’ expert, testified that this practice was reckless conduct. Plaintiffs introduced an affidavit at least implying that Dowd knew plaintiffs, and others like them, were vulnerable. Protective custody was not readily available to inmates because they were required to spend extended periods of .time, in administrative segregation before being admitted to protective custody. Dr. Sable testified that this is a significant barrier to inmates requiring protection. Given this evidence presented by plaintiffs, we cannot say the district court erred by denying Dowd’s motion for j.n.o.v.
Dowd also claims that he was entitled to j.n.o.v. because as a state actor he was protected by qualified immunity. In Harlow v. Fitzgerald, 457 U.S. 800, 102 S.Ct. 2727, 73 L.Ed.2d 396 (1982), the Supreme Court held that defendants will be protected by qualified immunity in § 1983 suits only if their actions did not “violate clearly established statutory or constitutional rights of which a reasonable person would have known.” Id. at 818, 102 S.Ct. at 2738. This court has clearly held that prisoners have a constitutional right to be free of sexual attacks by other inmates and that a prison official is liable if they are deliberately indifferent or act with reckless disregard of that right. Vosburg, 845 F.2d at 765-66; Martin, 742 F.2d at 474. Because this is a clearly established constitutional right about which Dowd either knew or should have known, Dowd was not entitled to qualified immunity in this case. The district court did not err in denying his motion for j.n.o.v. on this basis.
D. Attorneys’ Fees
Dowd claims that the district court abused its discretion in awarding plaintiffs’ counsel $94,680 in attorneys’ fees. Dowd argues that because plaintiffs failed on the issue of damages, they achieved limited success. Therefore, the district court should reduce the award to reflect the limited success achieved by plaintiffs. Dowd also argues that the district court failed to consider that one of the two plaintiffs’ attorneys wa,s inexperienced and attended all depositions and proceedings without doing anything, thereby unnecessarily increasing the hours plaintiffs’ attorneys spent on this case.23
*676A district court has a “special understanding” of the complexity of a case, and therefore is best equipped to determine the reasonableness of the hours expended by counsel. Catlett v. Missouri Highway & Transp. Comm’n, 828 F.2d 1260, 1270 (8th Cir.1987), cert. denied, 485 U.S. 1021, 108 S.Ct. 1574, 99 L.Ed.2d 889 (1988) (citing 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)). We will overturn a district court’s award of attorneys’ fees only if the district court abused its discretion. Id.
In this case, the district court awarded plaintiffs $94,680 in attorneys’ fees based on 946.9 hours expended at a reasonable hourly rate of $100.24 When determining the amount of attorney’s fees, the district court must focus on the “significance of the overall relief obtained by the plaintiff in relation to the hours reasonably expended on the litigation.” Hensley v. Eckerhart, 461 U.S. 424, 435, 103 S.Ct. 1933, 1940, 76 L.Ed.2d 40 (1983). The Supreme Court has also stated that “[t]he amount of damages a plaintiff recovers is certainly relevant to the amount of attorney’s fees to be awarded under § 1988. It is, however, only one of many factors that a court should consider in calculating an-award of attorney’s fees. We reject the proposition that fee awards under § 1988 should necessarily be proportionate to the amount of damages a civil rights plaintiff actually recovers.” City of Riverside v. Rivera, 477 U.S. 561, 574, 106 S.Ct. 2686, 2694, 91 L.Ed.2d 466 (1986) (plurality) (citations omitted). We believe that the district court properly followed these mandates. First, the magistrate judge specifically considered the level of success in calculating his award. The magistrate judge stated that “[although at first blush such an award might seem excessive in light of the nominal damages awarded, the measure of success in Civil Rights litigation of this type is not limited to the monetary damage award. There is inherent value in a determination that ones [sic] constitutional rights have been violated. Accordingly, this court finds that plaintiffs’ attorneys are entitled to fees in the amount of $94,680.” Corder v. Dowd, No. 89-276C(6), slip op. at 4 (E.D.Mo. Sept. 19, 1990). The record reflects that the litigation did serve the interest of prisoners other than the plaintiffs, as the defendant subsequently instituted some of the safety measures requested by the plaintiffs.25 Cf. Rivera, 477 U.S. at 585-86, 106 S.Ct. at 2699-2700 (Powell, J., concurring).
The court also considered several other factors, including the complex nature of the litigation, the quality of the work, the zealous manner’in which plaintiffs’ attorneys pursued their cause, and the effect this litigation had on the attorneys’ small firm. Id. at 3. In addition, the court awarded attorneys’ fees to counsel based on. a flat hourly fee of $100 even though plaintiffs’ lead counsel presented substantial evidence supporting a fee of $175 an hour for his time. After reviewing the district court’s calculations and the evidence submitted by plaintiffs, we cannot find that the district court abused its discretion in this case.
*677E. Instruction
Dowd argues that the instruction on the standard to determine whether he violated the Eighth Amendment was erroneous. The instruction read as follows:
Prison officials may be liable for violation of a prisoner’s civil rights, where they are deliberately indifferent to a prisoner’s constitutional right to be free from sexual attacks by other inmates, if they actually intend to deprive them of that right, or if they act with careless disregard of this right. Reckless disregard of the prisoner’s right to be free from sexual attacks by other inmates may be shown by the existence of a pervasive risk of harm to inmates from other prisoners and failure of prison officials reasonably to respond to that risk.
We have clearly held that a prison official may be liable for a violation of a prisoner’s right to be free from sexual attacks by other inmates if that official is deliberately indifferent to that right or acts with reckless disregard for the right. Vosburg, 845 F.2d at 765-66; Martin, 742 F.2d at 474. Dowd argues that Wilson, — U.S. at -, 111 S.Ct. at 2326-27, has changed the law and he should not have been subject to liability for reckless disregard. Dowd contends that he is therefore entitled to a new trial on liability. Dowd argues this issue for the first time on appeal. He did not object to the instruction at the time of trial. We do not address this issue because it has not been preserved and, therefore, is not before us. See Diamonds Plus, Inc. v. Kolber, 960 F.2d 765, 768 (8th Cir.1992); United States v. Standefer, 948 F.2d 426, 430 (8th Cir.1991).
III.
For the foregoing reasons, we affirm the district court on all counts.
. The Honorable Robert L. Kingsland, United States Magistrate Judge for the Eastern District of Missouri.
. Three of the plaintiffs have since been released. Only David Corder is still incarcerated. His first parole hearing is set for December 1992.
.Two other inmates were also plaintiffs in the original action. Their claims were voluntarily dismissed before trial.
. The other defendants were dismissed by plaintiffs prior to trial. The five other causes of action were dismissed by plaintiffs after the evidence had been presented to the jury but before the case had been submitted to them.
. This policy has been changed to require that the cell doors be open at all times during open wing. One of the plaintiffs, David Corder, filed a grievance protesting this change of policy.
.The Classification Committee evaluates requests for protective custody and recommends a course of action to the Superintendent. The Committee also evaluates requests by prisoners in administrative segregation waiting for protective custody or by those already in protective custody to be returned to general population.
. A closed cell door automatically locks. When prisoners wish to leave the cell, they call out their cell number. The officer in the bubble hears this through the microphones in the ceiling of the wing and presses a button that opens, the door. Therefore, a guard is not present when the prisoners leave the cell and would not necessarily notice if the prisoners were not the inmates assigned to the cell.
. He was not, however, reassigned to the same cell.
. Daniel Henry, an official at FCC, testified that, while serving as Acting Superintendent during Dowd’s absence, he approved a Classification Committee recommendation that French, who had requested protective custody because he was being pressured for sex by another inmate, be returned to general population. French said he wished to return to general population because he preferred that to going to administrative segregation pending bed space in protective custody. Henry did not conduct any investigation into French's allegations, and he approved returning French to his prior cell even though that cell was in close proximity to the cell of the inmate French claimed was pressuring him for sex. Henry had no instructions to conduct an investigation or to take any other action in this type of case.
. All the inmates accused of rape in this case were serving long sentences for violent crimes and have histories of violence toward other prisoners.
. Plaintiffs originally sued for, inter alia, transfer to the Missouri Eastern Correctional Center. That prison does not have any protective custody available, and sends those inmates who need protective custody to FCC.
. Several other FCC officials also testified that no prisoner who persists in their request for protective custody throughout the review process is ever denied protective custody.
. We may speculate that the jury could have concluded the plaintiffs did not prove the defendant caused all of their injuries because the plaintiffs submitted to the jury a general verdict form concerning the question of damages — one that makes it impossible to determine which claims the jury did or did not believe. Cf. Williams v. North Carolina, 317 U.S. 287, 63 S.Ct. 207, 87 L.Ed. 279 (1942) (general verdict of guilty cannot stand if one potential ground is unconstitutional because a general verdict does not state what the jury' decided).
. Dr. Sable’s testimony concerning rape trauma syndrome consists solely of the following: "There can be what’s generally described as the rape trauma syndrome, which includes emotional, cognitive and psychological consequences.” Tr., Vol. Ill at 56. He also testified that rape can cause physical injury and infection. Id. Because neither he nor any medical doctor examined the plaintiffs at the time of any of the attacks, there is no medical evidence that the plaintiffs received physical injury.
, Each plaintiff admitted on cross-examination to signing the statements. See cross-examination of Hershel Marsh, Tr., Vol. II at 202; cross-examination of Frank Ledferd, Tr., Vol. Ill at 139; cross-examination of Jay French, Tr., Vol. Ill at 166, 168; cross-examination of David Corder, Tr., Vol. Ill at 216.
. See cross-examination of Hershel Marsh, Tr., Vol. II at 213-17; cross-examination of Jay French, Tr., Vol. Ill at 176; cross-examination of David Corder, Tr., Vol. Ill at 203-09.
. After trial, Jim Purket replaced Denis Dowd as Superintendent of FCC. Plaintiffs moved pursuant to Fed.R.Civ.P. 25(d)(1) to add Purket as a party defendant in his official capacity. Because we find that plaintiffs are not entitled to equitable relief, we affirm the district court’s denial of this motion.
. Plaintiffs’ requested jury instruction concerning liability provided that in order for the jury to find in favor of the plaintiffs, it must find:
One, the plaintiff was subjected to one or more sexual attacks by prisoners in Farming-ton Correctional Center. Two, defendant knew or should have known the prisoners, such as plaintiff, were subject to a pervasive risk of sexual attack. Three, defendant failed to respond reasonably to the risk of such attacks. Four, the defendant’s conduct was a proximate cause of the plaintiffs injury. And five, the plaintiff was injured.
Tr., Vol. V at 48. Plaintiffs’ requested declaratory relief was identical to the liability jury instruction. Appellants’ App. at 28, Pis.’ Mem. in Support of Their Motion for Declaratory and Injunctive Relief.
.At oral argument, Judge Magill commented that the only victory that plaintiffs can claim is $4, and that’s not very much of a victory. Plaintiffs’ attorney responded that they had $4 and a declaration implicit in the verdict that a very important constitutional right was denied these men.
.We address this issue only as it relates to Corder because he is the only plaintiff still incarcerated. The claims for injunctive relief of the other three plaintiffs, who have been released from prison, are moot.
. See Corder Testimony, Trial Tr., Vol. Ill at 218-19. Prison policy changed to require inmates to keep their doors open during open wing period, when free movement of inmates within the wing is allowed.
. See supra Part I. at 668 n. 11.
. The argument before us is that the attorneys’ fees should be reduced. This is a different issue from the question of whether a plaintiff can get attorneys’ fees when they are awarded only nominal damages at trial. The circuits are split on this issue. Compare Domegan v. Ponte, 972 *676F.2d 401 (1st Cir.1992); Romberg v. Nichols, 953 F.2d 1152 (9th Cir.1992) (per curiam) (nominal damages make plaintiff prevailing party for purposes of obtaining attorneys' fees pursuant to 42 U.S.C. § 1988); Ruggiero v. Krzeminski, 928 F.2d 558, 564 (2d Cir.1991) (same); Coleman v. Turner, 838 F.2d 1004 (8th Cir.1988) (per curiam) (same); Nephew v. City of Aurora, 830 F.2d 1547 (10th Cir.1987) (en banc) (same), cert. denied, 485 U.S. 976, 108 S.Ct. 1269, 99 L.Ed.2d 481 (1988); and Garner v. Wal-Mart Stores, Inc., 807 F.2d 1536, 1539 (11th Cir.1987) (same) with Estate of Farrar v. Cain, 941 F.2d 1311 (5th Cir.1991) (nominal damages do not make plaintiff prevailing party for purposes of obtaining attorneys’ fees pursuant to 42 U.S.C. § 1988), cert. granted, — U.S. -, 112 S.Ct. 1159, 117 L.Ed.2d 407 (1992). Because this issue is not before us in this case, we do not address it.
. The court disallowed .1 hour because it was paralegal time and not compensable in attorneys’ fees. Dowd does not dispute that plaintiffs’ attorneys spent 946.9 hours working on the claim on which they succeeded and claims related to it.
. Specifically, the FCC instituted a new policy requiring that the doors of all prisoners be open during its “open wing” period.