dissenting:
I disagree with the reasoning of, and result reached by, both the District Court and the majority. I think we are faced here with an issue of much greater import, both practically and analytically, than mere permissible prison regulation. While some measure of deference is certainly to be afforded to prison authorities, nonetheless we must make certain that we do not convert the Turner v. Safley test into a rubber stamp. Here, the policy at issue has been applied so as to target a religious group for different treatment, including a blanket denial of First Amendment rights.1 We must deal with this wholesale treatment of members of a religious group in a careful manner.
Appellants urge us to address the following question: When the prison adopts a policy and then targets members of one religion and imposes significant burdens on — even perhaps totally impedes — their religious exercise, based solely on the prisoners’ religious affiliation, does not the first prong of the Turner v. Safley test *524require that the violence of the group and the members subjected to this treatment be clearly proven in order for such group treatment to be “reasonably related” to the legitimate goal?2
Two key facts are present here that, I submit, warrant that we proceed with extra caution. First, the policy is being applied based on membership in a group, specifically, in a religious group; the individual conduct that results in the purported basis for the imposition of the restriction is not violent or threatening activity, but, rather, is religious observance essentially protected by the First Amendment. Second, the cognitive “leap” from the fact of membership in a religion to the validity of a concern about security is not an automatic one, not “common sense,” and we must require a showing of the proper fit between membership in the religious group and valid security concerns.
The confluence of these two factors should cause us to pause and consider the appropriate test, and the applicable evi-dentiary burden. We have noted that the first prong of Turner is subject to the test of a “means-end fit,” which we have described as follows:
We may conclude that the statute bears no “valid, rational connection” to rehabilitation if “the logical connection between the [statute] and the asserted goal is so remote as to render the policy arbitrary or irrational.”
Waterman v. Farmer, 183 F.3d 208, 215 (3d Cir.1999) (citing Turner v. Safley, 482 U.S. 78, 89-90, 107 S.Ct. 2254, 96 L.Ed.2d 64 (1987)). We went on to explain: “This standard is similar to rational-basis review, under which a statutory classification can be declared unconstitutional only where the relationship of the classification to its asserted goal is ‘so attenuated as to render the distinction arbitrary or irrational.’ The legislature’s judgment therefore need not be perfect, just rational.” Id. (citations omitted).
But what about the situation where the prison regulation targets a specific religious group — where it does not merely burden the exercise of religion, but, rather, effectively singles out members of a certain religious group for different treatment and denial of free exercise rights?3 I suggest that in such a situation we should require an even “closer fit” between the religious group’s classification and the state’s proffered security interests.
This unique aspect of this case has not been fully explored by the parties, but it is nonetheless troubling. Does it make any difference that the group targeted is a religion and that “core” membership is the determining factor for imposition of restrictions? Is this not more insidious than a ban on certain conduct or specific activi*525ty that happens to have an impact on one’s religious beliefs or exercise? Laws targeting religious beliefs are clearly suspect;4 and, the right to religious freedom is not to be surrendered at the prison door. See O’Lone v. Shabazz, 482 U.S. 342, 348, 107 S.Ct. 2400, 96 L.Ed.2d 282 (1987).
When applying the Turner test in a case placing harsh restrictions upon inmates with certain religious beliefs, I proffer that we should indeed require a “tight” or “closer” fit between the correctional system’s admittedly legitimate interest and an inmate’s beliefs. The Supreme Court has not indicated that the Turner test must be a rigid one and has in fact referenced with approval the concept that it would be reasonable to require a closer “fit” in certain instances, for example, where the threat to the government interest is not as great.5 I suggest that a closer fit might be required when the inmate’s interest — his religious beliefs — is so significant and the restrictions are so great.
In Waterman v. Farmer, 183 F.3d 208, 213 (3d Cir.1999), we found a rational relationship where the authorities denied prisoners the right to read pornographic materials. Testimony was presented by two different psychologists to the effect that pornographic material would thwart the effectiveness of the treatment being given to the prisoners — who were all sex offenders who had exhibited “repetitive and compulsive” behavior. The prison authorities also referred the court to a considerable body of research supporting this view. We upheld the regulation and noted that, there, we probably would not have needed the expert opinions because “ ‘common sense tells us that prisoners are more likely to develop the now-missing self-control and respect for others if prevented from poring over pictures that are themselves degrading and disrespectful.’ ” Waterman, 183 F.3d at 217 (quoting Amatel v. Reno, 156 F.3d 192, 199 (D.C.Cir.1998)).
However, this reasoning does not apply here. There, we approved of a ban on certain literature based on specific objective criteria demonstrably consistent with legitimate penological objectives; here, we *526are faced with a round-up of all members of a purportedly violent religion so that they can be subjected to religious “detox” in the name of security. I submit that if prison authorities are to be permitted to target and categorize a certain religion so as to severely circumscribe First Amendment rights, based solely on membership in the religion, we should require that the first prong of the Turner v. Safley test be satisfied only by, at a minimum, a close fit between the targeted religion and problem sought to be avoided, here, to “minimize the occurrence of assaults on staff and inmates,” and evidentiary requirements that leave no room for doubt. To require any less is to permit — perhaps encourage-profiling: that is, the arbitrary attribution of certain characteristics to a group and, therefore, members of that group, resulting in denial of rights and different, disadvantaged treatment.
Further, it would be one thing if the prisons were only “profiling” security threat groups that are clearly violent “gangs;” but, here, the District Court assumed that the Five Percent Nation was a religion.6 The evidence before the District Court was woefully lacking that membership in the Five Percent Nation carried with it a set of beliefs that each member acts upon to promote violence and disorder.7 The District Court relied on a report prepared by Ronald Holvey, an eighteen-year veteran employee of the New Jersey correctional system. His credentials consist largely of on-the-job training,8 and his report includes no proof of what, I suggest, is required — namely, that membership equates to an active commitment *527to violence. Instead, the report is anecdotal, recounting, as the Appellees even note in their brief, “twelve violent or threateningly violent incidents involving a member or members of the Five Percent Nation” during a seven year period. Appellees’ Brief, p. 14. There is no proof of violent gang activity involving FPN members in New Jersey prisons, and none of the incidents links the conduct to the members’ religious beliefs.9 Mr. Holvey cites absolutely no statistics with respect to crimes by Five Percent Nation members, as compared to crimes by other groups. In fact, in pointing out that one in seven prison inmates is a member of the Five Percent Nation, the paucity of violent incidents purported to be linked to FPN members actually casts doubt on the violent nature of the group. No showing was made that there was a greater proportion of violence by FPN members than by groups of other kinds, such as Christians, Jews, or Muslims. The evidence is probative only of the assertion that there are several members of the FPN that have committed violent or unruly acts.
Nowhere in their brief do Appellees counter, let alone point to evidence that would meet, Appellants’ statement that the Five Percent Nation’s “teaching does not in any way advocate or encourage violence or disorderliness.”10 App. Br. at 22. Rather, the District Court and the majority allude to the findings of other courts to the effect that the Five Percent Nation fosters violence. Those courts based their rulings on evidence before them, involving the facts presented to them. The District Court here should demand no less, but only the Holvey report was presented and relied upon. There is no basis for the District Court to take judicial notice of the evidence before other courts.
Further, most of the decisions referenced by the District Court and the majority, in addition to being non-precedential, are either distinguishable or not relevant.11 *528In the most persuasive and well-reasoned opinion cited, Self-Allah v. Annucci, No. 97-CV-607(H), 1999 WL 299310, *2 (W.D.N.Y.1999), the issue is quite different — whether the court would enjoin the prison authorities’ confiscation of the prisoners’ copies of The Five Percenter. The injunction was denied, based upon extensive testimony and evidence presented with respect to the violent propensity of the Five Percenters in the New York system. Many witnesses testified (including Mr. Holvey) in favor of the relationship between the gang violence in the system and the group. The Court concluded:
Plaintiff has demonstrated that Five Percenterism, in its pure, uncorrupted form, represents a system of beliefs which, outside the prison context, does not advocate or promote violence. However, the testimony presented by defendants showed a clear relationship between Five Percenter literature and prison gang activities.
Id. at *9. The majority is misguided to suggest that this supports their conclusion “that inmates belonging to the Five Percent Nation present a serious security threat.” Maj. at 517 (emphasis added). There is a very important difference between a threat posed by “belonging” to a religion and that allegedly posed by the circulation of the group’s literature.
In the only other court of appeals case involving a similar threat group policy, In re Long Term Admin. Segregation of Inmates Designated as Five Percenters, 174 F.3d 464 (4th Cir.1999) [hereinafter Five Percenters], the Court of Appeals for the Fourth Circuit noted its jurisprudence as requiring “some minimally rational relationship,” id. at 468, and emphasized the Supreme Court’s jurisprudence calling for deference, especially when dealing with state correctional institutions and the preservation of order therein. Id. at 469 (citing Turner, 482 U.S. at 85, 107 S.Ct. 2254, and Bell v. Wolfish, 441 U.S. 520, 547, 99 S.Ct. 1861, 60 L.Ed.2d 447 (1979)). The court noted “ample evidence in the records” supporting the reasonableness of the conclusion that they posed a threat to prison security, including incidents in 1992, 1993, and three specific incidents in 1995, and referenced at least one incident report in which it was stated that “ ‘these five inmates acted as a group,’ that they ‘felt as if they were acting in a manner acceptable to the[ir] religious beliefs,’ and that they ‘spoke of more violence to come.’ ” Id. at 466. The prisoners in that case were not denied their religious literature (this claim had been settled), and the court gave great deference to the prison system’s decision as “manifestly a rational action.” Id. at 470. The court went on to explain:
The question is not whether [the South Carolina Department of Corrections (“SCDC”) director’s] conclusion was indisputably correct, but whether his conclusion was rational and therefore entitled to deference. Confronted with multiple reports of an identifiable group whose members not only threatened but had actually committed serious, violent acts in the SCDC system and else*529where, [SCDC Director’s] decision to designate the Five Percenters as an STG was manifestly a rational action.
Id. at 470 (citation omitted). It explained: “Allowing prison officials to act only after a demonstration of individual dangerousness would deprive them of the all-important option of prevention. The threat of violence here was a group threat, and prison administrators were entitled to address it in those terms.” Id. at 466.
While I think the test of “minimally” rational relationship has not been employed by our court and chips away at Turner, I cannot argue with the result reached in that case.12 The group had been identified based on clear, repeated past group violent conduct, attributable to its set of beliefs. But we have no such evidentiary record here. Accordingly, not only is the purported relationship more tenuous but the genuineness of the security threat is more remote as well. Additionally, the restrictions there did not include denial of literature, which is at the heart of the FPN’s exercise. In the case before us, the implications are much more far-reaching and the evidence much less relevant and convincing. There has been no showing of the “means-end fit” to satisfy the wholesale denial of religious freedom and exercise.
If the inquiry does not satisfy the first prong of Turner, which we have explained received the greatest weight, the prison’s action must fail. See Shaw v. Murphy, 532 U.S. 223, 121 S.Ct. 1475, 1479, 149 L.Ed.2d 420 (2001) (“If the connection between the regulation and the asserted goal is ‘arbitrary or irrational,’ then the regulation fails, irrespective of whether the other factors tilt in its favor.”). Therefore, we need not reach the other prongs. However, I cannot help but note my disagreement with the ease with which the majority dispenses with the second and third prong as well. The answer to the “alternate means” prong is really self-evident— the prison authorities’ course of “treatment” is designed to cause the FPN adherent to give up his faith, not permit him to practice it. This is much different than the facts before the Fourth Circuit in Five Percenters, where the court found that because the prisoners could still pray, fast and study religious materials, the “alternate means” test was satisfied, and the proven violence satisfied the “no ready alternatives” prong. In the course of this treatment, the FPN member is barred from the teachings, which are at the heart of the Five Percent Nation religious experience. Furthermore, to be released from close custody he must promise to never again affiliate with FPN. Thus, the desired result of the treatment is to eradicate the belief. It is difficult to see how, realistically, there are “alternate means” here.
In connection with the “impact of the accommodation on others” prong, here, the premise that potential violence cannot be accommodated assumes the very violence that I suggest has not been shown. I would suggest that the absence of a showing of the violent connection as discussed above undermines the findings of the District Court and the majority with respect to this prong as well.
We have in this country a rich tradition of protecting individual rights, including the rights of prisoners. We have explained that “the Supreme Court has made clear that ‘convicted prisoners do not forfeit all constitutional protections by reason of their conviction and confinement in prison’ ” DeHart v. Horn, 227 F.3d 47, 50 (3d Cir.2000) (en banc) (quoting Bell v. Wolf*530ish, 441 U.S. 520, 545, 99 S.Ct. 1861, 60 L.Ed.2d 447 (1979)). Furthermore, “ ‘Inmates clearly retain protections afforded by the First Amendment, ... including its directive that no law shall prohibit the free exercise of religion.’ ” Id. (quoting O’Lone v. Shabazz, 482 U.S. 342, 348, 107 S.Ct. 2400, 96 L.Ed.2d 282 (1987)). At the same time, we have a fast-developing body of law to the effect that, while inmates do not shed their constitutional protections at the jailhouse door, nonetheless “a prison inmate ‘retains [only] those rights that are not inconsistent with his status as a prisoner or with the legitimate penological objectives of the corrections system,’ ” id. at 51 (quoting Pell v. Procunier, 417 U.S. 817, 822, 94 S.Ct. 2800, 41 L.Ed.2d 495 (1974)), and “the constitutional rights that prisoners possess are more limited in scope than the constitutional rights held by individuals in society at large.” Shaw v. Murphy, 532 U.S. 223, 229, 121 S.Ct. 1475, 149 L.Ed.2d 420 (2001).
We have, at times, overreacted in response to perceived characteristics of groups thought to be dangerous to our security or way of life and condemned individuals based on group membership. See, e.g., Garner v. Bd. of Pub. Works, 341 U.S. 716, 71 S.Ct. 909, 95 L.Ed. 1317 (1951) (upholding requirement that all city employees must disclose membership to the Communist party and swear an oath of loyalty); Korematsu v. United States, 323 U.S. 214, 65 S.Ct. 193, 89 L.Ed. 194 (1944) (affirming the constitutionality of “excluding” people of Japanese descent from the West Coast during World War II). Only later, when we have viewed these reactions with some perspective, have we acknowledged that the wholesale treatment of certain groups was not consistent with the basic tenets of our democracy. Here, similarly, it seems as though there is a rush to brand the Five Percent Nation as a “violent” religious sect. But, who is next? Would it be the Sunni Muslims, whose tenets, Appellants argue, are similar? Would it be the Nation of Islam, viewed by some as racist? While these may be inmates, and prisoners, they are nonetheless people. We should therefore be concerned, and be careful in labeling and judging them based solely on membership in a religious group.
If membership in such groups can objectively be shown, upon close scrutiny, to be equated to posing a real threat of violence in the prison setting, then treatment of such group members in wholesale fashion, even though it deprives them of their constitutional rights, would be consistent with legitimate penological objectives, and would be permissible. Otherwise, such discriminatory treatment treads impermis-sibly on their constitutional rights.
I would reverse the District Court’s ruling and deny defendant’s motion for summary judgment for lack of a showing that the first prong of the Turner v. Safley test has been satisfied.13
. While providing a detailed description of the procedures provided by the STG policy, the majority does not reference the extensive restrictions imposed on Appellants. According to the STG policy, restrictions on inmates in Phase 1 of the program include: strip-searches each time they leave or return to their cells; a total of five hours per week out of their cells; a shower or shave only every third day; only a single, non-contact visit each month; only one monitored phone call per week; prohibition on correspondence with any other inmate, including incarcerated family members; all meals eaten in cells; and, no access to regular prison programs. App. al A138-42, 148. Further, the Policy instructs that "[tjhere is a 'Zero Tolerance’ level for Security Threat Group activity within the Department’s correctional facility(ies).” App. at A152. Examples of such activities include: "Possession of Security Threat Group literature such as lessons, membership lists, and artwork; Possession of Security Threat Group paraphernalia such as beads, artwork, medallions, and clothing articles; ... Participation in Security Threat Group related . .. meetings, gatherings, ... and events.... ” App. at A152.
. One could quibble with whether the restrictions are on all members, because the close custody only applies to those determined to be "core members.” However, given that mere possession of materials about FPN (the sole basis for Fraise’s designation) raises an individual to "core member” status and the fact that these individuals will only be released from close custody upon repudiation of the religion, together with the weak case against Appellants, points to the conclusion that all FPN believers who either read or express in any fashion the teachings of the Five Percent Nation, are clearly at risk and subject to restrictive custody. Interestingly, it has been noted that Five Percenters read and learn, rather than pray, as their religious observance, and this goes to the essence of what is being denied here. See Self-Allah v. Annucci, No. 97-CV-607(H), 1999 WL 299310, * 2 (W.D.N.Y.1999) ("Five Percenters are obligated to study and learn the lessons of the Five Percent Nation of Islam.”).
. In order to be released from the STGMU, inmates must sign a "Letter of Intention” expressing their intention "to renounce formally and in spirit affiliation with all Security Threat Groups.” App. at A443.
. See Church of Lukumi Babalu Aye, Inc. v. Hialeah, 508 U.S. 520, 533, 113 S.Ct. 2217, 124 L.Ed.2d 472 (1993) ("[A] law targeting religious beliefs as such is never permissible....”); McDaniel v. Paty, 435 U.S. 618, 626, 98 S.Ct. 1322, 55 L.Ed.2d 593 (1978) (“The Free Exercise Clause categorically prohibits government from regulating, prohibiting, or rewarding religious beliefs as such."); Torcaso v. Watkins, 367 U.S. 488, 495, 81 S.Ct. 1680, 6 L.Ed.2d 982 (1961) ("We repeat and again reaffirm that neither a State nor the Federal Government can constitutionally force a person 'to profess a belief or disbelief in any religion.’ ”); Cantwell v. Connecticut, 310 U.S. 296, 304, 60 S.Ct. 900, 84 L.Ed. 1213 (1940) ("In every case the power to regulate must be so exercised as not, in attaining a permissible end, unduly to infringe the protected freedom.”).
. In Abbott, the Supreme Court noted that where "the nature of the asserted governmental interest is such as to require a lesser degree of case-by-case discretion, a closer fit between the regulation and the purpose it serves may safely be required.” Thornburgh v. Abbott, 490 U.S. 401, 411-12, 109 S.Ct. 1874, 104 L.Ed.2d 459 (1989) (discussing and overruling Procunier v. Martinez, 416 U.S. 396, 94 S.Ct. 1800, 40 L.Ed.2d 224 (1974)). The Supreme Court explained that the rejection of the regulation in Martinez was based on the Court's "recognition that the regulated activity centrally at issue in that case — outgoing personal correspondence from prisoners — did not, by its very nature, pose a serious threat to prison order and security.” Id. at 411, 109 S.Ct. 1874. The Court clarified: "We do not believe that Martinez should, or need, be read as subjecting the decisions of prison officials to a strict 'least restrictive means' test.” Abbott, 490 U.S. at 411, 109 S.Ct. 1874. The Court overruled Martinez as far as it suggested a legal distinction between incoming correspondence from prisoners and incoming correspondence from nonprisoners. Id. at 413-14, 109 S.Ct. 1874.
. This case would present different issues had the District Court not assumed that the Five Percent Nation was a religion. The Court would have been required to determine whether FPN would be considered a religion, and therefore accorded the protections provided by the Free Exercise Clause of the First Amendment. As this issue is not before us, and was not before the District Court, we need not decide whether the FPN would satisfy these requirements, but only stress that non-traditional belief systems found to be religious in nature will be afforded the same protections as traditional ones. See Africa v. Pennsylvania, 662 F.2d 1025, 1031-32 (3d Cir.1981) (setting forth three indicia to be used in determining whether a ''religion'' is at issue).
. While the FPN tenets may be racial in tone, racism is not the same as violence. See McCabe v. Arave, 827 F.2d 634, 638 (9th Cir.1987) ("[Pjrison authorities have no legitimate penological interest in excluding religious books from the prison library merely because they contain racist views. Courts have repeatedly held that prisons may not ban all religious literature that reflects racism.”); Stefanow v. McPadden, 103 F.3d 1466, 1472-73 (9th Cir.1996) (applying Turner and observing that "[mjerely 'advocating racial purity’ is insufficient to justify confisca-ti[ng]” religious material, and upholding the confiscation of the book Christianities Ancient Enemy because it directly advocates violence by issuing an explicit "call to arms for white Christians to fight back in 'a war for survival’ ”); Murphy v. Missouri Dep’t of Corrs., 814 F.2d 1252, 1257 (8th Cir.1987) (holding that restriction of inmate access to racist religious materials "must be limited to those materials that advocate violence or that are so racially inflammatory as to be reasonably likely to cause violence at the prison”); Aikens v. Jenkins, 534 F.2d 751, 756-57 (7th Cir.1976) (striking down a regulation banning all racist periodicals in prison because the regulation "is not narrow enough to reach only that material which encourages violence, and invites prison officials to apply their own personal prejudices and opinions as standards”).
.Holvey has eighteen years of experience in corrections employment, including service as a corrections officer. Appellees note that he belongs to several national or regional law enforcement or intelligence organizations, including the National Major Gang Task Force, and that he has assisted several states and organizations, including the Federal Bureau of Investigation, with the process of identifying security threat groups and members as well as training. Holvey Deposition, App. at A249-50, 254-58.
.The majority characterizes the Holvey Report as reporting a “string of incidents” and as citing "numerous instances of actual or planned violence involving Five Percenters in New Jersey correctional facilities between August 1990 and July 1997.” Maj. at 516. In reality, a careful scrutiny of the report and corresponding attachments reveals very little evidence of planned or actual violence by FPN members, let alone by the Five Percent Nation as a group. In his report, Holvey lists twelve “Specific Violent Acts/Intended Acts of Violence/Specific Illegal or Prohibited Acts.” App. at A342-43. Of this twelve, three do not even involve the New Jersey correctional facilities, App. at A402-03, 404-11, and two more relate to New Jersey youth facilities— one involving a gathering of approximately 50 FPN and NETA members, and the other consisting of a fight including some believed to be FPN members. App. at SA41-46. Of the remaining seven incidents there were two gatherings, App. at SA24-29, A412, and three incidents involving a single FPN member. App. at SA5-20, SA48-56. There was only one report of a series of altercations allegedly involving more than one FPN member along with several Sunni Muslims. App. at SA37-39. And one letter allegedly from an FPN member threatening violence against prison guards. SA21. None of these incidents reflects activity atypical of aggressive behavior one would anticipate in a prison setting. The incidents cited in the report do not demonstrate the FPN’s violent tendencies as a group or gang in the New Jersey prison system. The report conflates incidents from other places, as well as violence by other inmates where a FPN member may have been tangentially involved, with incidents involving the FPN as a group.
. Moreover, it should be noted that Appel-lees in their brief really fail to address the underlying “disconnect” that I perceive, but urge instead that the threat group policy is related to a legitimate goal. Appellants concede this, but argue that the designation of the Five Percent Nation is not so related because there is no valid connection in the New Jersey prison system between the Five Percent Nation and security concerns.
. See Allah v. Beyer, 1994 WL 549614, at *1-3 (D.N.J. Mar.29, 1994) (upholding the trans*528fer of FPN member where there was specific evidence that the inmate took a leadership role in planning a violent uprising in the prison); Box v. Petsock, 697 F.Supp. 821, 831 (M.D.Pa.1987) (considering petitioner's religious affiliation in the context of an ineffective assistance of counsel claim); Abed v. Comm'r of Corrs., 43 Conn.App. 176, 682 A.2d 558 (1996) (holding that petitioner, not an FPN member, did not have a liberty interest in good-time credit); Allah v. Dep’t of Corrs., 326 N.J.Super. 543, 742 A.2d 162, 165 (1999) ("We accept the argument of the DOC that it neither 'targeted' a religion nor classified religious beliefs as a security threat group, but merely designated an association of inmates based on its history of violence as a security threat group."); Buford v. Goord, 258 A.D.2d 761, 686 N.Y.S.2d 121, 122 (N.Y.App.Div.1999) (upholding a ban on FPN literature).
. In Shaw v. Murphy, Justice Thomas referred to rebutting the presumption of rationality. However, that language was dicta and I suggest did not lower the standard of Turner, nor did it require the burden of proof to fall on the plaintiff.
. I see no need to address Appellants’ equal protection or due process claims. I agree with the majority that we apply the same Turner analysis to Appellants' equal protection claim that we did to their First Amendment claims, and my concerns with the majority's Turner analysis carries over to the equal protection analysis as well. On the other hand, I will not provide my own reasoning regarding Appellants' due process claim, as I agree with that provided by the majority.