Plaintiffs-appellants (“plaintiffs”), four white inmates at Illinois’ Pontiac Correctional Center (“Pontiac”),1 instituted this class action on behalf of themselves and all other present and future white inmates housed in the maximum security units of Pontiac. On March 16, 1985, plaintiffs filed a motion for a preliminary injunction to compel defendants-appellees, Illinois Department of Corrections and Pontiac Correctional Center administrators, (collectively “administration”), to adhere to and enforce Illinois Department of Corrections (“IDOC”) regulations prohibiting gang activities at Pontiac. Plaintiffs alleged that the administration’s failure to enforce *1267those IDOC regulations violated their 14th Amendment right to equal protection and certain federal regulations. After implementing a limited protective order for the safety of the inmates who were scheduled to testify, the court heard evidence on the motion for a preliminary injunction. In a thorough opinion, the district court denied plaintiffs’ motion for a preliminary injunction finding that plaintiffs had failed to show a reasonable likelihood of success on the merits of their 14th Amendment claim and their federal regulatory claim. Specifically, the court found that plaintiffs had failed to present evidence that the administration’s policies were intentionally discriminatory or that a nexus existed between the administration’s allegedly discriminatory policies and the federally funded program. Because we agree with the district court’s assessment that the appellants are not reasonably likely to succeed on the merits of their claims and because we find no error in the court’s conduct of the hearing, we affirm.
I.
Although we do not intend to reiterate the district court’s lengthy factual findings, a brief exposition of the facts is necessary.
A. Factual Background
Pontiac Correctional Center, located in Pontiac, Illinois, is one of four maximum security facilities in Illinois. On the average, Pontiac houses 2000 inmates. At the time the district court made its findings, there were approximately 1800 inmates in Pontiac’s general population and 260 inmates in the Pontiac protective custody unit. The protective custody unit is designed “to ensure the safety of inmates determined to be vulnerable to attack and intimidation.” Inmates in the protective custody unit are separated from the galleries housing all other Pontiac inmates, collectively referred to as the “general population.” Generally, an inmate in protective custody is entitled to most of the privileges to which an inmate m the general population is entitled although an inmate in protective custody is usually confined more hours per day, has less opportunity to obtain certain prison jobs, and receives privileges at different times than the general population. All these measures are necessary for the inmate’s own protection.
An inmate may request a transfer to protective custody but is usually somewhat reluctant to do so because of the stigma attached to such a request. For example, an inmate seeking protective custody may often be branded as a “stoolie”, “wimp” or homosexual by the general population. Although a request for protective custody is immediately honored by the prison administration, that request is subject to an initial evaluation and reevaluation every 30 days thereafter. If an inmate is not able to justify his initial or any subsequent request for protective custody, he is returned to the general population.
Only 12% of the total inmate population at Pontiac is white, yet forty percent of the total white inmate population is in protective custody while only 9% of the total black population and 13% of the hispanic population are in protective custody. The inordinately2 high number of white inmates in protective custody is directly related to the predominately black gang-activity at Pontiac.
At the hearing on the motion for a preliminary injunction, it was estimated that anywhere from 75% to 99.5% of the total inmate population at Pontiac are gang members. Although the district court found that the number was probably closer to 75%, defendant-appellee, Michael Lane, the director of the Illinois Department of Corrections, testified that about 90% of all Pontiac inmates are gang members or are gang-affiliated. A portion of the inmates entering Pontiac, who are not gang members, ultimately “choose” to affiliate or, in the vernacular, “ride” with or “aid and assist” a gang. Inmates, who ride with gangs are offered protection by the gang *1268but must pay for that protection literally and figuratively by performing any number of deeds, ranging from carrying weapons to performing a “hit” for gang members.
Virtually all gangs are black or hispanic.3 White inmates do not normally become members of the black or hispanic gangs but may only “ride” the non-white gangs. All inmates, who are not gang members, are “strongly urged” and, more often than not, physically coerced to affiliate with a gang. Although a number of both black and white inmates refused to affiliate with a gang, “very few white inmates in general population were not gang affiliated.” That is to say, the white inmates, who decline to accept an invitation to become gang-affiliated, normally may be found in the Pontiac protective custody unit.
Gang activity, including extortion, possession of contraband, intimidation, and physical and sexual abuse, is expressly prohibited by Pontiac prison rules. Further, the administration, as a matter of departmental policy, does not officially recognize the existence of various prison gangs. The fact is, however, that the existence of those gangs is a reality with which the administration is confronted daily. In an attempt to realistically deal with gang-related problems, prison officials discipline gang “activity” but do not discipline non-violent displays of gang “membership.” However, these non-violent exhibitions of gang membership create an environment in which the gangs flourish and as a consequence, an environment in which prohibited gang activities can and do take place with alarming intensity and frequency. It is not an overstatement that the gang “situation in Illinois prisons” has reached “the crisis stage.”
B. The District Court’s Decision
Plaintiffs filed a motion for a preliminary injunction to enjoin the administration’s failure to enforce the rules against gang activities in Pontiac. Plaintiffs argued the administration predominantly favored black and hispanic gang members by adopting a policy of least resistance to gang activity. Thus, plaintiffs say that the administration purposefully and intentionally discriminated against white inmates, who generally were not gang members, in violation of plaintiffs’ 14th Amendment right to equal protection. Additionally, plaintiffs claim that the administration’s discriminatory practices violated 28 C.F.R. § 42.101 et seq., prohibiting discrimination in any program receiving federal funds.
After hearing the evidence in support of the motion for a preliminary injunction, the district court, relying on our decisions in Roland Machinery Co. v. Dresser Industries, Inc., 749 F.2d 380 (7th Cir.1984) and Lawson Products, Inc. v. Avnet, Inc., 782 F.2d 1429 (7th Cir.1986), initially outlined the legal requirements to which any motion for a preliminary injunction is subject. The court found that before a preliminary injunction will issue, the movant must show (1) that he has no adequate remedy at law and that the movant faces irreparable harm if the preliminary injunction is not granted; (2) the movant has some likelihood of success on the merits of his claims; (3) the benefit to the movant of granting the preliminary injunction outweighs any harm the non-movant might suffer and (4) the public interest will not be disserved by the issuance of the injunction. The district court then noted that if plaintiffs could show, by some evidence, that they had been deprived of a constitutional right they would, as a matter of law, also satisfy the burden of showing irreparable harm. Thus, the keystone to plaintiffs’ motion for a preliminary injunction was whether plaintiffs might reasonably be expected to succeed on the merits of the claim that the administration deprived them of their 14th Amendment right to equal protection.
To prevail on their 14th Amendment claim, the district court ruled that plaintiffs were required to show that the administration’s actions not only had a disparate impact on them as a class, but that those actions were purposefully undertaken. *1269Based on the evidence adduced at the hearing on plaintiffs’ motion, the district court found that plaintiffs had failed in both regards. First, plaintiffs failed to make a “threshold showing that there is an invidious classification as they have claimed.” The court based this finding on the fact that plaintiffs presented no evidence of the treatment received by black or white inmates who, though not gang members, continued to live in general population. Nor did the plaintiffs show that all white inmates in protective custody were there because of pressure to become gang members. Finally, plaintiffs failed to present evidence that black inmates in protective custody were not forced into protective custody by gang members. In short, while the white inmates, who testified that they had been forced into protective custody, were credible, no evidence by other inmates, black or white, was offered and the court simply could not find that only one group— namely white non-gang member inmates in protective custody — were in protective custody because of gang activity.
Second, the court found that even if defendants’ actions resulted in an invidious classification of inmates, there was no evidence that defendants acted with the intent to discriminate. In so holding, the district court rejected plaintiffs’ argument that a 14th Amendment equal protection claim can be based on something less than purposeful discrimination. Plaintiffs argued that the administration had several alternative means of dealing with gang-related activities. According to plaintiffs, the administration chose an alternative in reckless disregard of the impact that alternative would have on white non-gáng member inmates. Because the impact on white inmates was clearly foreseeable, the administration must have intended the impact and therefore must have been discriminatorily motivated. The administration, on the other hand, argued that an intent to discriminate cannot be inferred from the decision to adopt a certain policy despite the fact that the resulting racially disparate impact might have been foreseeable. Without conceding that reckless disregard is the appropriate standard, the administration also argued that they made their policy decisions in good faith.
After reviewing evidence presented by the parties’ expert witnesses, the court concluded that there had been egregious lapses in Pontiac’s security measures and in the administration’s exercise of judgment. However, the court held that the decision to implement a policy which curtailed only gang “activity”, as opposed to gang “membership,” was not made in intentional disregard of the impact that decision would have on white inmates. Although the district court did not view non-violent displays of gang membership at Pontiac “with the same measure of tolerance that the Department of Corrections evinces,” the court concluded that the administration’s chosen policy of limiting overt gang activity was probably the best option available to the administration. Thus, plaintiffs’ equal protection claim was substantially without merit.
Then, based on Personnel Administrator of Massachusetts v. Feeny, 442 U.S. 256, 99 S.Ct. 2282, 60 L.Ed.2d 870 (1979), the district court held that a policy’s disproportionate impact on a suspect class (assuming a class existed) coupled with evidence of the foreseeability of that impact is not sufficient to prove discriminatory intent where the policy selected is a legitimate one. Relying on Feeny, the district court said:
... [Discriminatory purpose requires that ... [defendants’ course of action was undertaken in part ‘because of’ and not merely ‘in spite of’ its adverse effects on white inmates. However, where those adverse effects are inevitable or compellingly foreseeable, there is a strong inference that they were intended.... But ‘the inference simply fails to ripen into proof’ where the impact is an unavoidable consequence of a legitimate policy.
The district court ruled that the administration’s policy, relating to gang activity at Pontiac, was a legitimate one. The court found that although plaintiffs’ expert would not have selected the particular policy implemented by the Pontiac administra*1270tion, even plaintiffs’ expert agreed that the chosen policy was legitimate and, as the court found, adopted in good faith to control violent gang behavior.
As a final matter, the district court also found that given the legitimacy of the defendants’ policy and the crisis state of the gang situation at the Illinois prison, the plaintiffs’ equal protection claim lacked merit because the state had a compelling interest in implementing and ensuring the viability of its present policies.
Based on all these factors, the district court concluded there was only a “remote possibility, not a reasonable likelihood, that plaintiffs [could] succeed in their equal protection claim.”
The district court next ruled that there was only a slight possibility that plaintiffs’ federal regulatory claim would succeed since plaintiffs failed to submit any evidence of a nexus between the administration’s allegedly discriminatory actions and the program under which Pontiac received federal funds.
Having found that there was little likelihood plaintiffs could prevail on the merits of their claims, the court very briefly addressed the balance of equities in issuing a preliminary injunction, finding that the state’s compelling interest in controlling gangs would be hindered if an injunction were issued. The court also ruled that the issuance of a preliminary injunction would disservice the public.
II.
On appeal, plaintiffs contend the administration’s reckless disregard of their well-being in adopting certain policies, is tantamount to intentional discriminatory behavior. Plaintiffs claim that they need not show a nexus between the administration’s actions and the federal funds received by Pontiac in order to prevail on their federal regulatory claim. Plaintiffs also appeal from a limited protective order entered by the district court which allegedly hampered their ability to present their case.
A. The Fourteenth Amendment Claim
Plaintiffs argue that they presented evidence that the administration’s policy of suppressing only gang “violence” and not gang “membership” in effect, created a class of white non-gang members which was forced into protective custody. This class, they allege, was based on race and was therefore a suspect class4 as a matter of law. Plaintiffs maintain that the district court erred in requiring them to show that black non-gang members, in or out of protective custody, were excluded from the effect of the administration’s policy or that white non-gang members in protective custody were all in protective custody because o/the administration’s policies. Additionally, plaintiffs contend they proved that the administration was discriminatorily motivated in selecting a policy of tacit acceptance, by establishing that it did not select the least segregative policy from a number of alternative policies. In other words, plaintiffs argue intent may be inferred where the disparate consequences of the policy were obvious to the administration from the start.
*1271Plaintiffs acknowledge the Feeny decision and the district court’s finding that the administration’s method of dealing with the gang problem at Pontiac was legitimate. However, plaintiffs argue that although the administration’s express policy may have been legitimate, the policy actually implemented was not legitimate and therefore was not subject to the Feeny limitations. Put another way, plaintiffs contend that proof of the obviously discriminatory consequences of selecting a specific policy is sufficient to prove discriminatory intent where the policy in question is not legitimate.
Before addressing plaintiffs’ substantive arguments, we briefly examine our standard for reviewing an order granting or denying a preliminary injunction. We recently held that our review of a district court’s decision to grant or deny a preliminary injunction is subject to the abuse of discretion standard. Lawson Products, Inc. v. Avnet, Inc., 782 F.2d 1429 (7th Cir.1986). Acknowledging that the abuse of discretion standard of review cannot be defined precisely in the preliminary injunction context, we held that the ultimate balancing of equities by the district court must be accorded great deference. Nevertheless, a judge making both factual findings and drawing legal conclusions in determining whether to grant or deny a preliminary injunction, may abuse his discretion in any of three ways:
(1) he may apply incorrect substantive law or an incorrect preliminary injunction standard; (2) he may rest his decision to grant or deny a preliminary injunction on a clearly erroneous finding of fact that is material to the decision to grant or deny the injunction; or (3) he may apply an acceptable preliminary injunction standard in a manner that results in an abuse of discretion.
Zepeda v. INS, 753 F.2d 719, 724 (9th Cir.1983).
Thus, as we said in Lawson:
When a court of appeals considers a preliminary injunction order, the factual determinations are reviewed under a clearly erroneous standard and the necessary legal conclusions are given de novo review. [Citations omitted.]
Lawson Products, Inc. v. Avnet, 782 F.2d at 1437. Still, although we may be called upon to review certain of the district court’s factual findings or legal determinations on appeal, that does not “obfuscate the fact that the standard of review of the grant or denial of a preliminary injunction is the deferential ‘abuse of discretion’ standard.” Id.
In this instance, plaintiffs contend the district court abused its discretion by incorrectly applying substantive law to the facts. Plaintiffs also argue the court erred by imposing an incorrect preliminary injunction standard by requiring them to prove that all the members of the allegedly invidious class were in that class as a result of defendants’ policies and not merely by happenstance.
We need not decide whether the administration’s policies actually create a “suspect class” (although clearly the percentage of white inmates in protective custody is disproportionately large in comparison to the total number of white inmates in the general population) since plaintiffs’ equal protection claim fails for another reason. Even if the administration’s policies have a disparate impact on a suspect class, plaintiffs failed to show that the administration harboured a discriminatory motive in implementing those policies.
... [T]he “Equal Protection Clause has long been limited to instances of purposeful or invidious discrimination rather than erroneous or even arbitrary administration of state powers_” Briscoe v. Kusper, 435 F.2d 1046, 1052 (7th Cir.1970). A plaintiff “must demonstrate intentional or purposeful discrimination” to show an equal protection violation. Bloomenthal v. Lavelle, 614 F.2d 1139, 1141 (7th Cir.1980) (per curiam). “ ‘Discriminatory purpose’ however, implies more than intent as volition or intent as awareness of consequences.” Personnel Administrator of Massachusetts v. Feeny, 442 U.S. 256, 279, 99 S.Ct. 2282, 2296, 60 L.Ed.2d 870 (1979). It implies that the decisionmaker singled out a particu*1272lar group for disparate treatment and selected his course of action at least in part for the purpose of causing its adverse effects on an identifiable group.
Shango v. Jurich, 681 F.2d 1091, 1104 (7th Cir.1982).
Plaintiffs presented no evidence that defendants selected a policy of limiting gang activity and permitting gang membership because o/the effect it would have on the allegedly suspect class, in this case the white inmates who were forced into protective custody. Therefore, even though certain IDOC regulations and policies, though facially neutral, may effect certain groups unevenly, such an uneven effect will not give rise to constitutional concern unless the policy is an obvious pretext for discrimination against the suspect class. Cf. Yick Wo v. Hopkins, 118 U.S. 356, 6 S.Ct. 1064, 30 L.Ed. 220 (1886).
... [A]s was made clear in Washington v. Davis, 426 U.S. 229, 96 S.Ct. 2040, 48 L.Ed.2d 597 and Arlington Heights v. Metropolitan Housing Dev. Corp., 429 U.S. 252, 97 S.Ct. 555, 50 L.Ed.2d 450, even if a neutral law has a disproportionately adverse effect upon a racial minority, it is unconstitutional under the Equal Protection Clause only if that impact can be traced to a discriminatory purpose.
Personnel Adm’r. of Massachusetts v. Feeny, 442 U.S. at 273, 99 S.Ct. at 2293 (emphasis added). Nevertheless, a disparate “impact itself would signal that the real classification made by the law [is] ... not neutral” if no neutral explanation for the disparate impact can be offered. Personnel Adm’r. of Massachusetts v. Feeny, 442 U.S. at 275, 99 S.Ct. at 2294.
Here, much like in Feeny, the prison administration offered a neutral explanation for the disparate impact its policy had on white inmates. Like the state statute in Feeny, which drew distinctions between veterans and non-veterans but had a disparate impact on women because so few women were veterans, the administration’s gang-related policies, which draw a distinction between gang membership and gang activity, have a disparate impact on white inmates because a large majority of those inmates are not gang members. Given that there is a neutral explanation for the disparate impact of the administration’s policies, plaintiffs have failed to show that the administration’s policies are a pretext for preferring black and hispanic inmates over white inmates. As the Supreme Court stated:
Just as there are cases in which impact alone can unmask invidious classification, cf. Yick Wo v. Hopkins, 118 U.S. 356, 6 S.Ct. 1064, 30 L.Ed. 220 (1886), there are others, in which — notwithstanding the impact — the legitimate noninvidious purposes of a law cannot be missed.
Id. at 442 U.S. at 276, 99 S.Ct. at 2294.
Therefore, as in Feeny, “the dispositive question” in this instance is whether the plaintiffs have shown that a racially-based “discriminatory purpose has at least in some part shaped” the Pontiac administration’s gang activity policy. Id. The fact that the administration may have been aware that its policy would have a greater effect on white non-gang member inmates is not enough. Arguably, the administration’s decision to implement a policy, which does not prohibit gang membership, might be expected to benefit gangs to the extent that the gangs’ continued existence is assured. It might also be expected that the administration’s decision would adversely affect non-gang members who are largely white. On the other hand, there is no evidence that the administration intended to harm white non-gang members, though such harm might have been an “inevitable concomitant of the chosen scheme.” Feeny, 442 U.S. at 278, 99 S.Ct. at 2295. In short, the foreseeability of the adverse consequences of the administration’s chosen policy is not sufficient to prove that the administration intentionally sought to harm white non-gang member inmates. Id. Nothing in this record would indicate that any policy, having a preferential effect on gang member inmates at Pontiac, was adopted or devised with the collateral goal of discriminating against white inmates.
*1273Plaintiffs argue the more stringent Fee-ny burden of proving discriminatory intent should be applied only where the challenged policy is a legitimate one. Plaintiffs rely on footnote 25 in the Feeny opinion in which the Court said:
This is not to say that the inevitability or foreseeability of consequences of a neutral rule has no bearing upon the existence of discriminatory intent. Certainly, when the adverse consequences of a law upon an identifiable group are as inevitable as the gender-based consequences of ch. 31, § 23, a strong inference that the adverse effects were desired can reasonably be drawn. But in this inquiry-made as it is under the Constitution-an inference is a working tool, not a synonym for proof. When, as here, the impact is essentially an unavoidable consequence of a legislative policy that has in itself always been deemed to be legitimate, and when, as here, the statutory history and all of the available evidence affirmatively demonstrate the opposite, the inference simply fails to ripen into proof.
Plaintiffs' position is that any prison policy which permits the continued existence of gangs is not legitimate. The very fact that gangs exist permits those gangs to flourish and wreak havoc on the system.
The district court, after hearing testimony from the expert witnesses, concluded that a number of alternatives for handling the gang problem at Pontiac existed and that the policy chosen by the administration was one of these alternatives. Although there was conflicting testimony on whether the chosen policy was the best one available, there was no evidence that the policy was not legitimate. Thus, the perceived distinction between Feeny and this case is non-existent.
Plaintiffs also argue that even if the chosen policy for dealing with the gang problem at Pontiac was legitimate on its face, that policy was not implemented. Plaintiffs contend the administration may have adopted a legitimate policy but then implemented it in a haphazard fashion or not at all. In contravention to its own policy, the administration permitted certain activities to occur and to remain unpunished. Thus, the "actual" policy put into effect by the administration was not legitimate and no direct evidence of discriminatory intent, as required by Feeny, is necessary.
Once again however, the district court expressly found that the administration adhered to its policy and enforced it more times than not. Recognizing that there were a number of serious security breaches and "lapses" in the administration's judgment concerning the extent of recognition given to the gangs, the district court nevertheless correctly found that even serious violations of the underlying policy did not render the policy itself illegitimate.
To the extent the district court's legal conclusions concerning the applicability of Feeny depend on its evidentiary determination that the administration's policy is a legitimate one, we must defer to the district court's ruling. Any debate about whether plaintiffs are required to make some showing that the administration purposefully discriminated against them was resolved when the district court found that the administration's chosen policy for dealing with gang activity at Pontiac was legitimate. As we have already said, where a policy is legitimate its disparate impact on a group is not sufficient to sustain an equal protection claim. The district court heard the testimony of expert witnesses and concluded that the administration had indeed chosen a viable and recognized means of dealing with the gang problem. As we said in SEC v. Suter, 782 F.2d 1294, 1800 (7th Cir.1984) "on review `[f]indings of fact shall not be set aside unless clearly erroneous, and due regard shall be given to the opportunity of the trial court to judge of [sic] the credibility of witnesses.'"
We hold therefore that the district court correctly applied Feeny to the facts in this case. We affirm the court's finding that plaintiffs' equal protection claim will not succeed on its merits in the complete absence of any evidence that the administration intentionally discriminated against *1274white inmates at Pontiac when it implemented its gang activity policies.
B. The Federal Regulatory Claim
Likewise, we agree with the district court that plaintiffs are not likely to succeed on the merits of their Title VI claim.
In their complaint, plaintiffs alleged that the administration violated 28 CFR 42.104, a federal regulation prohibiting discrimination in any program receiving federal funds. Plaintiffs contended, and the administration did not dispute, that Pontiac received federal funds from the National Institute of Corrections (“NIC”). According to plaintiffs, the Pontiac administration used those funds to run Pontiac in a discriminatory fashion by allowing black and hispanic gang member inmates to victimize white inmates. Plaintiffs further allege that each of the named inmates suffered monetary damages, although it is not clear from the complaint whether plaintiffs allege that those damages are the direct result of a violation of the federal regulations.5 The administration responded by arguing that plaintiffs failed to present any evidence of a nexus between the federal funds received by Pontiac and its allegedly discriminatory actions.
The district court indicated that it might be assumed that the funds received by Pontiac, used for population forecasting and developing a unified classification system for incoming prisoners, would ultimately inure to the benefit of the Pontiac inmates because funds received and used in forecasting could have some impact on the administration’s protective custody policies. However, the district court found that regardless of the reasonableness of such inferences, plaintiffs had failed to present any evidence on which the court could base those inferences. In other words, the district court found that plaintiffs failed to establish any connection between the use of the federal funds and defendants’ allegedly discriminatory practices.
Plaintiffs assert that the case law should not be read to require them to prove a nexus between the funds received by Pontiac and the administration’s allegedly discriminatory actions in order to prevail on their claim based on federal regulations.
It is clear that plaintiffs may maintain a private cause of action to enforce the regulations promulgated under Title VI of the Civil Rights Act. Cannon v. University of Chicago, 441 U.S. 677, 99 S.Ct. 1946, 60 L.Ed.2d 560 (1979). Moreover, plaintiffs need not show intentional discriminatory conduct to prevail on a claim brought under these administrative regulations. Evidence of a discriminatory effect is sufficient. Guardians Association v. Civil Service Commission of the City of New York, 463 U.S. 582, 103 S.Ct. 3221, 77 L.Ed.2d 866 (1983); 28 CFR 42.104(b)(vii)(2). However, there is some dispute whether evidence of a nexus between the program receiving federal funds and the defendants’ actions is necessary.
The provision upon which plaintiffs’ regulatory claim is based is 28 CFR 42.104 which, in its relevant part, states:
(a) General No person in the United States shall on the ground of race, ... be denied the benefit of, or be otherwise subjected to discrimination under any program to which this subpart applies.
A “program” is defined in 28 CFR 42.102(d) as:
any program, project or activity for the provision of services ... or other benefits to individuals....
More importantly for plaintiffs’ purposes, 28 CFR 42.104(d) states:
the disposition, services ... or benefits provided under a program receiving Federal financial assistance shall be deemed to include any portion of any program or function or activity which ... is directly or indirectly improved, enhanced, enlarged or benefited by such Federal financial assistance.
(Emphasis added.)
Applying these provisions to this case, plaintiffs contend the federal funds *1275received from the NIC, though intended for prison forecasting models, benefit the entire prison system. Therefore, all programs so benefited, including protective custody rules, are subject to federal regulation. Since some of the prison's gang activities rules purportedly have a discriminatory effect on white inmates, plaintiffs argue those rules violate 28 CFR 42.104 and must be enjoined.
As the district court correctly found, however, plaintiffs' broad assumption that the entire Pontiac prison system is benefited by receipt of NIC funds, is not supportable absent proof of such a benefit. In Grove City College v. Bell, 465 U.S. 555, 104 S.Ct. 1211, 79 L.Ed.2d 516 (1984), the Supreme Court found that regulations promulgated under Title IX (which was patterned on Title VI) were subject to certain program-specific limitations. Thus, only a program which actually benefited from the receipt of federal funds administered under Title IX was subject to the Title IX regulations prohibiting discrimination. In Grove City, the Supreme Court found that while federal student aid, first received by students and then paid to Grove, benefited Grove and, in effect, constituted federal aid to Grove, only the college's financial aid program-the program actually affected by the student loans-was subject to federal regulation. The entire college did not become subject to Title IX regulations merely because one program was benefited by federally funded student aid. In this instance, •the district court, relying on Grove City, found that the entire Pontiac prison system was not subject to Title VI regulations merely because Pontiac received funds earmarked for one specific program-the population forecasting and prisoner classification program-absent evidence that those funds somehow affected or benefited other Pontiac programs.
On appeal, plaintiffs contend Grove City is inapplicable to this case because unlike the plaintiffs in Grove City, they seek only injunctive relief and do not seek termination of federal funding.
Both Title IX and Title VI contain parallel limiting provisions which state:
Compliance with any requirement adopted pursuant to {20 USC 1681(a) (Title IX) or 42 USC 2000(d) (Title VI) J may be effected (1) by the termination of or refusal to grant or to continue assistance under such program or activity to any recipient as to whom there has been an express finding on record ... but such termination or refusal shall be limited to the particular political entity or part thereof or other recipient as to whom such a finding has been made and, shall be limited in effect to the particular program, or part thereof in which such non-compliance has been found, or (2) by any other means authorized by law.
(Emphasis added.) Plaintiffs contend that in Grove City, the plaintiffs sought termination of funding as described in subpara-graph (1) above. The present suit, on the other hand, is based on a request for in-junctive relief, which must be brought under subparagraph (2). Since only subpara-graph (1) contains the program-specific limitations, plaintiffs argue they need not show a nexus between the federal funds and the administration's policies in order to obtain an injunction.
We now hold that the type of remedy sought is not relevant to the issue of whether the regulations promulgated under Title VI should be applied in a program-specific manner. We think plaintiffs' construction of the federal regulations, which is based on a distinction in remedies, makes no sense. If, for example, an institution receives federal funds for two programs but only one is discriminatorily administered and a plaintiff seeks to enjoin such discriminatory behavior, injunctive relief would have to be applied to both programs under plaintiffs' view of the regulations, since injunctive relief is not program-specific. The Supreme Court, in Grove City, has held that if an institution receives federal funds which only affects certain of its other non-funded programs, only those non-funded programs so benefited or otherwise affected are subject to regulation. Thus, Grove City applies whether the remedy requested for a violation of a Title VI *1276regulation is termination of funding or in-junctive relief.
While it is possible that the federal funds received by Pontiac and earmarked for the forecasting models may directly benefit or relate to the implementation of gang regulations and protective custody procedures, plaintiffs presented no evidence of such a connection. In the absence of such evidence, the conclusion that funds earmarked for prisoner classification and population forecasting will indirectly benefit the entire Pontiac system, may be reached only by “ignoring Title (Vi’s) program-specific language.” We therefore hold that the district court correctly found that plaintiffs’ regulatory claim also has little likelihood of success on the merits.
C. The Protective Order Claim
As a final matter, plaintiffs contend that the district court hampered their ability to present evidence at the preliminary injunction hearing because the court granted only a limited protective order.6 During the course of the hearing in district court, plaintiffs requested that the courtroom be completely closed and that no testimony “be released or published without prior court approval” in order to encourage two white inmates to testify about incidents of homosexual rape by gang members.
Plaintiffs had earlier sought a protective order limiting the number of guards in the courtroom and a directive that those guards and defendants’ personnel be prohibited from repeating the testimony of inmate witnesses. In addition, plaintiffs also requested that the record be sealed and that inmates witnesses’ names not be used in the record on appeal.
The district judge, recognizing that he was faced with difficult choices regarding public and media access to the courtroom and prior restraint of media coverage, balanced open access to courtrooms and First Amendment rights with the danger of retaliation against inmate witnesses. The court apparently found there was insufficient reason to compel either media restrictions7 or closure of the courtroom to the public.
We cannot say that the district judge erred in denying plaintiffs’ request to place restrictions on media coverage and access to the civil proceedings. See, Press-Enterprise Co. v. Superior Court of Cal., 478 U.S. 1, 106 S.Ct. 2735, 2743, 92 L.Ed.2d 1 (1986) (closure of criminal proceedings permitted only where such closure is essential to preserve “higher values” and is narrowly tailored); In the Matter of Continental Illinois Securities Litigation, 732 F.2d 1302, 1308 (7th Cir.1984) (policy reasons for granting public (and press) access to criminal proceedings apply to civil cases as well). Further, “Although the. bar against prior restraints is not absolute, Nebraska Press Ass’n. v. Stuart, 427 U.S. 539, 590, 96 S.Ct. 2791, 2817, 49 L.Ed.2d 683 (1976) (Brennan, J. concurring in judgment), the Court has repeatedly emphasized that the prior censorship of expression can be justified only by the most compelling governmental interest.” Brown v. Glines, 444 U.S. 348, 364, 100 S.Ct. 594, 611, 62 L.Ed.2d 540 (1980). New York Times Co. v. United States, 403 U.S. 713, 91 S.Ct. 2140, 29 L.Ed.2d 822 (1971).
Moreover, problems of constitutional dimension relating to prior restraint or restrictions on access to the courtroom could have been avoided without sacrificing the desired anonymity of the witnesses. For example, plaintiffs could have, but did not, request that the court order counsel to refrain from asking inmate witnesses to identify themselves in open court.8
*1277Although the district court granted plaintiffs’ request to limit the number of prison guards in the courtroom and to order those guards and defendants’ personnel to refrain from discussing testimony, the district judge denied plaintiffs’ request to seal the record of inmate testimony. Given that the parameters of denying access to the transcript of a witness’ testimony is a matter which remains unclear in both criminal and civil proceedings, compare, Gannett Co., Inc. v. DePasquale, 443 U.S. 368, 99 S.Ct. 2898, 61 L.Ed.2d 608 (1979) with Press-Enterprise Co. v. Superior Court of Cal., supra, and Courier-Journal v. Marshall, 828 F.2d 361, (6th Cir.1987), Bank of America Nat. Trust v. Hotel Rittenhouse, 800 F.2d 339 (3d Cir.1986) with Wilson v. American Motors Corp., 759 F.2d 1568 (11th Cir.1985), the district court did not commit reversible error in denying this request.
The district court also denied plaintiffs’ request to delete each inmates’ name and to instead use a letter designation in the record on appeal.9 It is in this area that the district court could have provided some additional protection to the testifying inmates, particularly with regard to the substitution of letters for inmate names in the hearing transcript.
On balance, under the circumstances of this case, we believe that the minimal protections provided the witnesses did not prejudice the plaintiffs’ presentation of their case. Without severely implicating First Amendment rights or the public’s right to access to the court, the'district court could have provided greater protection to the plaintiffs’ witnesses. The district court’s failure to enter a more comprehensive protective order does not require a new trial. Even considering the proffered testimony of the plaintiffs’ two witnesses, who refused to appear because of the lack of protection, the plaintiffs still failed to demonstrate a likelihood of success on the merits on both their Fourteenth Amendment claim and their federal regulatory claim. In the future, district courts should be sensitive to the obvious dangers faced by witnesses called to testify in cases such as this.
D. Conclusion
In holding, as we do today, that plaintiffs are not entitled to a preliminary injunction under either the Equal Protection Clause or Title VI federal regulations, we in no way condone either gang activity or membership at Pontiac or any other penal institution. We share the district court’s sense of outrage that the necessity of recognizing gang membership has led to a crisis situation at Pontiac. We understand that the administration must presently, out of practical consideration, allow prisoners to affiliate with gangs. We do not, however, be-liqve that the control exerted by gangs, as a result of gang membership, should be enhanced by the administration’s failure to prevent incidents of physical harm or by the fact that the administration feels threatened enough to discuss certain policy decisions with gang leadership. Such lapses of both security and judgment should not be tolerated. Prisoners should never be permitted to freely roam Pontiac, administer beatings to other prisoners, or “aid” the administration in ferreting out wrongdoers. Moreover, we, unlike the Pontiac administration, do not believe displaying gang insignia on walls or letting inmates control prison job assignments, (to name just a few gang activities) need be permitted. At this time it is apparently impossible to prohibit gang affiliation — but we believe that the. extreme manifestations of such affiliations should be prohibited and such prohibitions enforced.
We are aware that under the present state of the law relating to prisoner’s civil rights, the administration’s ability to impose more stringent disciplinary measures may be hampered. The administration faces two unpleasant choices. It may either impose more severe restrictions and *1278face the uncertainty of lawsuits for civil rights and constitutional violations; or, it may tacitly accept some of the gang-related problems, picking and choosing when to impose restrictions. Although the former option may expose the administration to uncertain results, the latter option unfortunately, as we have seen, causes the administration to lose control quickly thus, endangering the lives and safety of inmates. We strongly urge the administration to take firmer control and to seek to ultimately eliminate gang affiliation by such reasonable methods as it may develop.
For the reasons stated, the district court’s protective order and order denying the plaintiffs’ motion for a preliminary injunction are
Affirmed.
. The claims of Joliet Correctional Center inmates were all dismissed on the basis of res judicata and collateral estoppel with the exception of certain claims by one individual.
. The district court specifically found that the high number of white inmates in protective custody was not a random mathematical occurrence. The average number of white inmates in protective custody is triple that of the number of black inmates.
. The major gangs in Pontiac are extensions of several Chicago street gangs including the Disci-pies, Vice Lords, Cobra Stones and the El Rukns.
. In Griffin High School v. Illinois High School Ass’n., 822 F.2d 671 (7th Cir.1987), we explained:
The usual standard of review for a statute or regulation challenged on equal protection grounds is the rational basis test. Under this test, legislation is presumed to be valid, and will be sustained as long as the classification drawn by the statute is rationally related to a legitimate state interest. [Citations omitted.] However, the Constitution requires heightened judicial scrutiny [strict scrutiny] ... when a law classifies so as to burden a suspect class....
822 F.2d at 674. The traditional indicia of a "suspect class”, as defined by the Supreme Court, include a class "saddled with such disabilities, or subjected to such a history of purposeful unequal treatment, or relegated to such a position of political powerlessness as to command extraordinary protection from the majori-tarian political process." San Antonio Independent School Dist. v. Rodriquez, 411 U.S. 1, 28, 93 S.Ct. 1278, 1294, 36 L.Ed.2d 16 (1973). For example, classifications based on race, McLaughlin v. Florida, 379 U.S. 184, 85 S.Ct. 283, 13 L.Ed.2d 222 (1964) or ancestry, Oyama v. California, 332 U.S. 633, 68 S.Ct. 269, 92 L.Ed. 249 (1948) are inherently suspect and subject to strict scrutiny.
. Indeed, in their prayer for relief, plaintiffs did not request monetary compensation but only sought a declaration that the administration’s actions violated the regulations in question and an order enjoining further violations.
. The problems in conducting the hearing were substantially compounded because the hearing was held (by necessity) at the Pontiac Correctional Center.
. The single media representative in the courtroom was asked (but not required) to refrain from publishing the names of inmate witnesses. The media representative apparently complied with this request.
.In order to make their record and to avoid publicly identifying the inmate witnesses, the parties and counsel could have been provided necessary names and backgrounds of those witnesses in writing, in chambers or at a side-bar conference.
. There is some confusion on this point. The record of the proceedings shows that the court declined to enter an order granting plaintiffs’ request to seal the record and to use initials; but, in its order denying the motion for a preliminary injunction, the court indicated that it had granted those requests.