A federal prisoner at the low-security correctional institution in Allenwood, Pennsylvania brought this action pursuant to 28 U.S.C. § 1331, challenging a Congressional ban on the use of federal funds to distribute certain sexually explicit material to prisoners, along with its implementing regulation. The District Court rejected plaintiff's argument that the ban violates the First Amendment and dismissed his complaint, finding the prohibition to be reasonably related to the legitimate penological goal of prisoner rehabilitation. Because we find that the District Court erred in resolving the constitutional issue without an adequate factual basis, we will reverse and remand for further proceedings consistent with this opinion.

The Ensign Amendment, originally enacted as part of the Omnibus Consolidated Appropriations Act of 1997, prohibits the use of funds appropriated for the United States Bureau of Prisons (the "BOP") to "distribute or make available any commercially published information or material to a prisoner ... [when] such information or material is sexually explicit or features nudity." Pub.L. No. 104-208, § 614, 110 Stat. 3009-66 (1996). The amendment has been reenacted in each subsequent appropriations act, and is now codified at 28 U.S.C. § 530C(b)(6). An implementing regulation promulgated by the BOP defines the key terms of the amendment as follows: "sexually explicit" means "a pictorial depiction of actual or simulated sexual acts including sexual intercourse, oral sex, or masturbation"; "features" means that the publication in question "contains depictions of nudity or sexually explicit conduct on a routine or regular basis or promotes itself based upon such depictions in the case of individual one-time issues"; and "nudity" means "a pictorial depiction where genitalia or female breasts are exposed." 28 C.F.R. § 540.72(b).1 The definition of "features" includes an exception for material that contains nudity "illustrative of medical, educational, or anthropological content." Id. As examples of publications that do not "feature nudity," a 1996 program statement released by the BOP cites National Geographic, Our Body, Our Selves, the swimsuit issue of Sports Illustrated, and the Victoria's Secret catalog. Fed. Bureau of Prisons Program Statement 5266.07 (Nov. 1, 1996). The regulations are clearly targeted to the receipt by inmates of softcore and hardcore pornography.

Plaintiff Marc Ramirez filed suit in the Middle District of Pennsylvania in 1997, naming as defendants the United States Attorney General, the director of the BOP, and the warden of the Allenwood institution (collectively, the "government"). Alleging that magazines addressed to him were rejected as either being "sexually explicit" or "featuring nudity," Ramirez challenged the constitutionality of the Ensign Amendment and its implementing regulation on First Amendment grounds. After a series of procedural delays, the District Court finally reached the merits of Ramirez's complaint on a government motion to dismiss. Applying the familiar test for constitutional challenges to prison regulations set out in Turner v. Safley, 482 U.S. 78, 107 S. Ct. 2254, 96 L. Ed. 2d 64 (1987), the District Court determined that the amendment and regulations passed constitutional muster because they were rationally connected to the government's asserted interest in prisoner rehabilitation, prisoners still had access to a broad range of materials (including materials with sexually explicit text), accommodating the asserted right to view explicit materials would threaten the safety of correctional staff and other inmates, and no ready alternative existed that would accommodate Ramirez's asserted right at a de minimus cost to valid penological interests.

On appeal, Ramirez argues that the District Court erred in finding a rational connection between the ban on pornography and rehabilitation in the absence of any factual record, and in failing to engage in a "contextual, record-sensitive analysis" before determining the ban's overall reasonableness under Turner. The District Court had jurisdiction under 28 U.S.C. § 1331, and we exercise jurisdiction pursuant to 28 U.S.C. § 1291. We review de novo the District Court's decision to grant the government's motion to dismiss. Pryor v. National Collegiate Athletic Ass'n, 288 F.3d 548, 559 (3d Cir.2002).

In Turner v. Safley, the Supreme Court recognized an enduring tension between two conflicting principles in operation whenever a prisoner brings a constitutional challenge to a law or regulation affecting prison policy. The first principle, that "[p]rison walls do not form a barrier separating prison inmates from the protections of the Constitution," must be balanced against the practical reality that the judicial branch is ill-suited for running the country's prisons, a task committed to the particular expertise of the legislative and executive branches. Turner, 482 U.S. at 84-85, 107 S. Ct. 2254. To strike an appropriate balance between prisoners' exercise of their constitutional rights and the institutional needs of prison administrators, the Supreme Court held that a prison regulation implicating an inmate's constitutional rights must be "reasonably related to legitimate penological interests" to be valid. Id. at 89, 107 S. Ct. 2254. The Court developed a four-part test for assessing the overall reasonableness of such a regulation. As a threshold inquiry, "there must be a `valid, rational connection' between the prison regulation and the legitimate governmental interest put forward to justify it." Id. (quoting Block v. Rutherford, 468 U.S. 576, 586, 104 S. Ct. 3227, 82 L. Ed. 2d 438 (1984)). Courts must then determine "whether there are alternative means of exercising the right that remain open" to prisoners, and "[what] impact accommodation of the asserted constitutional right will have on guards and other inmates, and on the allocation of prison resources generally." Id. at 90, 107 S. Ct. 2254. Finally, a regulation's reasonableness may be evidenced by "the absence of ready alternatives" that would fully accommodate the constitutional right "at de minimus cost to valid penological interests." Id. at 90-91, 107 S. Ct. 2254. These requirements "serve as guides to a single reasonableness standard," but the first "`looms especially large' because it `tends to encompass the remaining factors, and some of its criteria are apparently necessary conditions.'" Waterman v. Farmer, 183 F.3d 208, 213-14 (3d Cir.1999) (quoting Amatel v. Reno, 156 F.3d 192, 196 (D.C.Cir.1998)).

To date, the United States Court of Appeals for the D.C. Circuit is the only federal appellate court to have considered the merits of a First Amendment challenge to the Ensign Amendment and its implementing regulation. In Amatel v. Reno, 156 F.3d 192 (D.C.Cir.1998), that court rejected the challenge, finding the restriction on the distribution of sexually explicit material to be reasonably related to the asserted penological interest of prisoner rehabilitation. See 156 F.3d at 202-03. After identifying prisoner rehabilitation as the legitimate penological interest advanced by the government, the court defined that interest broadly. It reasoned that the government's power to inculcate values in contexts such as public education transferred readily to the context of prison administration, implicitly identifying the promotion of "respect for authority and traditional values" as a legitimate rehabilitative purpose in and of itself. Id. (internal citation omitted). Having done this, it found that "Congress might well [have] perceive [d] pornography as tending generally to thwart the character growth of its consumers," and that, as a matter of common sense, "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." Id. at 199.

The Amatel court did not see the need for an evidentiary record, holding that its own common sense was sufficient to verify the rational connection between the Ensign Amendment's proscriptions and the asserted rehabilitative goal. Id. It did, however, cite a body of scholarly research to support the reasonableness of the proposition that pornography leads to male objectification of women, and that certain types of pornography can lead to male aggression and desensitize viewers to violence and rape. See id. at 199-200. The court determined that none of the three other Turner factors undermined the overall reasonableness of the Ensign Amendment and its implementing regulation.

Our own court has considered the constitutionality of a restriction similar to the Ensign Amendment, albeit in a different context than the one here. In Waterman v. Farmer, 183 F.3d 208 (3d Cir.1999), we upheld a New Jersey statute that restricted prisoners' access to pornographic materials at a facility for sex offenders who exhibited "repetitive and compulsive" behavior. After identifying the legitimate penological interest at stake as the rehabilitation of the state's "most dangerous and compulsive sex offenders," we evaluated the connection between the statute and that interest in light of an evidentiary record that included two expert affidavits from the facility itself. Those experts testified that sex offenders' exposure to pornography would thwart specific rehabilitative strategies and treatments administered by prison staff. Id. at 215-16. In reversing a district court that had found the prisoners' experts "more reasonable" than the government's, we cited Amatel for the basic proposition that "as long as [a] statute is rational, it clears [Turner ]'s first hurdle." Id. at 217. At least within the specific context of the rehabilitation of recidivist sex offenders, we also approved the Amatel court's use of common sense with regard to whether a ban on pornography might encourage the development of self-control and respect for others. See id. After examining the other Turner factors, we upheld the New Jersey statute as being reasonably related to the legitimate penological interest of sex-offender rehabilitation.

We addressed whether the requisite rational connection between a prison restriction and a legitimate penological interest can be found on the basis of "common sense" alone in Wolf v. Ashcroft, 297 F.3d 305 (3d Cir.2002). In Wolf, we reversed a district court's decision upholding a restriction on the showing of R-rated and NC-17-rated movies in federal prisons. The district court in that case found that no evidentiary record was necessary because we had endorsed Amatel' s "common sense" approach in Waterman, and summarily concluded that the restriction was "neutral and reasonable" under Turner. We found the district court's opinion deficient because it never stated or described the relevant penological interest (the government had asserted three distinct interests: prison security, crime deterrence, and rehabilitation). Id. at 308. We also noted that while a court "need not necessarily engage in a detailed discussion" of the connection between a prison policy and that interest, a "brief, conclusory statement" is insufficient for evaluating the application of Turner' s first prong. Id. Finally, we rejected the government's contention that such a connection could always be found without an evidentiary hearing:

While the connection may be a matter of common sense in certain instances, such that a ruling on this issue based only on the pleadings may be appropriate, there may be situations in which the connection is not so apparent and does require some factual development. Whether the requisite connection may be found solely on the basis of "common sense" will depend on the nature of the right, the nature of the interest asserted, the nature of the prohibition, and the obviousness of its connection to the proffered interest. The showing required will vary depending on how close the court perceives the connection to be.

Id. at 308-09. On remand, we directed the district court to "describe the interest served, consider whether the connection between the policy and interest is obvious or attenuated — and, thus, to what extent some foundation or evidentiary showing is necessary — and, in light of this determination, evaluate what the government has offered." Id. at 309.

Turning to the appeal before us, we find that the District Court erred in evaluating the Ensign Amendment and its implementing regulation under Turner' s first prong on a motion to dismiss, without any analysis or inquiry into the interests involved and the connection between those interests and the restriction at issue. First, although the District Court correctly identified rehabilitation as a legitimate penological interest, see O'Lone v. Estate of Shabazz, 482 U.S. 342, 348, 107 S. Ct. 2400, 96 L. Ed. 2d 282 (1987), it did so without adequately describing the specific rehabilitative goal or goals furthered by the restriction on sexually explicit materials. Second, even though the connection between the amendment and the rehabilitation of federal sex offenders may be obvious under Waterman, that connection becomes attenuated upon consideration of the entire population of BOP inmates, such that a factual record becomes necessary for determining the rationality of the amendment's overall connection to rehabilitative interests. On remand, therefore, the District Court must first identify with particularity the specific rehabilitative goals advanced by the government to justify the restriction at issue, and then give the parties the opportunity to adduce evidence sufficient to enable a determination as to whether the connection between these goals and the restriction is rational under Turner.

While the obvious end of rehabilitation is the prevention of further lawbreaking once offenders are released from prison, the scope of the interest itself has never been defined by the Supreme Court. See Amatel, 156 F.3d at 209 ("Unlike its interest in institutional security, the contours of the government's interest in rehabilitation are quite amorphous and ill-defined.") (Wald, J., dissenting). Certainly falling within the legitimate bounds of the interest are prison policies designed to target the specific behavioral patterns that led to a prisoner's incarceration in the first place, or behavioral patterns emerging during incarceration that present a threat of lawbreaking activity other than that for which the prisoner was confined. To say, however, that rehabilitation legitimately includes the promotion of "values," broadly defined, with no particularized identification of an existing harm towards which the rehabilitative efforts are addressed, would essentially be to acknowledge that prisoners' First Amendment rights are subject to the pleasure of their custodians. See, e.g., id. at 210 (arguing that under such a broad definition of rehabilitation, lawmakers could constitutionally engage in viewpoint discrimination by proscribing texts expressing disfavored positions) (Wald, J., dissenting). To the extent that the Amatel majority defines rehabilitation in this way, we disagree with its reasoning. See id., 156 F.3d at 209 ("[T]o proceed on some vague assertion of an interest in `rehabilitation' without the need to define the term or to show a connection between the proscribed activity and the chosen definition ... runs an overwhelming risk of overregulation.") (Wald, J., dissenting). While the actual right to view materials subject to the Ensign Amendment's proscriptions might be significantly narrow in this case, courts may not abdicate their responsibility to scrutinize carefully the government's reasons for infringing that right.2 See Thornburgh, 490 U.S. at 414, 109 S. Ct. 1874 ("[Turner' s] reasonableness standard is not toothless.") (internal quotation omitted); Amatel, 156 F.3d at 206, 211 ("[M]ore precisely, [the standard] is not a license for lawmakers, any more than prison wardens, to shortchange the constitutional rights that the Supreme Court has insisted prisoners continue to possess.... If rehabilitation is to be deemed a legitimate penological interest, the term must be given some shape, at least when it collides with fundamental liberties.") (Wald, J., dissenting). As a preliminary step in determining the extent to which evidence is required under Wolf where the penological interest advanced by the government is rehabilitation, therefore, a district court must describe with particularity the specific rehabilitative goal or goals relied upon by the government to justify the challenged regulation. See Wolf, 297 F.3d at 308 (rejecting "conclusory" statements that make it difficult to determine what connection a court sees between the advanced penological interest and a prison restriction).

We may gather from the District Court's reliance upon the scholarly works discussed in Amatel that, at the very least, it believed the government's specific rehabilitative goals to include the prevention of sex crimes and violence against women. See Amatel, 156 F.3d at 199-200. Were the Ensign Amendment's scope limited to federal prisoners who have committed sex crimes or violence against women, the means-end connection would be sufficiently obvious such that the first prong of Turner could be resolved on the basis of common sense. In Waterman, we found the prohibition against sexually explicit material to be clearly connected to the rehabilitation of recidivist sex offenders whose demonstrated inability to control their sexual impulses had led to their incarceration at the facility in question. See Waterman, 183 F.3d at 217 (noting that restrictions on pornography foster the "deferring of sexual gratification, [ ] sublimation of sexual impulses, [and] channeling of sexual expression into long-term relationship of caring and affection" related to the "now-missing self-control and respect for others") (quoting Amatel, 156 F.3d at 199). However, we do not find the connection between the Ensign Amendment and the government's rehabilitative interest to remain obvious upon consideration of the entire federal inmate population, including those prisoners not incarcerated for sex-related crimes. In this case, therefore, we believe Wolf necessitates the development of a factual record. See Wolf, 297 F.3d at 309 (requiring an evidentiary showing roughly corresponding to the degree to which the required means-end connection is "attenuated").

By no means do we wish to suggest that the only legitimate target of the Ensign Amendment is the class of convicted federal sex offenders. We recognize that the government has wide latitude in pursuing legitimate rehabilitative goals; courts may not substitute their own judgment in place of that of the legislative or executive branches where the position advanced by the government is not "irrational or unreasonable" but simply "less reasonable" than that of the prisoner-plaintiffs. See Waterman, 183 F.3d at 216. In the absence of a factual record, however, we cannot ignore the possibility that the proscription rationally applies to such a small percentage of the BOP inmate population that its connection to the government's rehabilitative interest "is so remote as to render [it] arbitary or irrational." Turner, 482 U.S. at 89-90, 107 S. Ct. 2254; Waterman, 183 F.3d at 213 (holding that the Turner test subsumes traditional overbreadth and vagueness analyses). Determining whether there is a rational link between sexually explicit material and the harms toward which the government's overall rehabilitative efforts are directed requires more than a conclusory assertion that the "consumption of [sexually explicit] publications [ ] implicitly elevate[s] the value of the viewer's immediate sexual gratification over the values of respect and consideration for others" and a generalized statement that sexual self-control is relevant to the rehabilitation of the entire class of federal prisoners.3 Amatel, 156 F.3d at 199.

As to whether an evidentiary basis is required for the remaining three Turner prongs, we repeat our observation that "we have historically viewed these inquiries as being fact-intensive... [requiring] `a contextual, record-sensitive analysis.'" Wolf, 297 F.3d at 310 (quoting DeHart v. Horn, 227 F.3d 47, 59 n. 8 (3d Cir.2000) (en banc)). Where the link between the regulation at issue and the legitimate government interest is sufficiently obvious, no evidence may be necessary to evaluate the other Turner prongs. See, e.g., Waterman, 183 F.3d at 217; but see Wolf, 297 F.3d at 310 (observing that the first prong does not subsume the rest of the inquiry). In this case, however, we agree with Ramirez that the third and fourth Turner factors cannot be adequately assessed in the absence of an evidentiary foundation.4

The third and fourth Turner factors require consideration as to whether accommodating the asserted right would have an adverse impact "on guards and other inmates[ ] and on the allocation of prison resources," as well as a determination as to whether alternatives exist that can accommodate the right "at de minimus costs to valid penological interests." Turner, 482 U.S. at 90-91, 107 S. Ct. 2254. The District Court's apparent factual conclusion that accommodation "would increase the risks of sexual crimes and misconduct within the prison walls," is speculative and unsupported. The existence of a possible "ripple effect" on the rehabilitation of prisoners legitimately targeted by the Ensign Amendment could reasonably be disputed; certainly relevant to this inquiry is whether those prisoners are housed separately from inmates whose rehabilitation would not be affected. For the same reason, it does not follow from our decision in Waterman that limited distribution can never be conducted at de minimus costs to valid penological interests. See Waterman, 183 F.3d at 219 (finding the third and fourth Turner prongs satisfied because the facility in question was insufficiently staffed to conduct case-by-case reviews and prisoners were "more than likely" to pass materials among one another); cf. Amatel, 156 F.3d at 213 (arguing that a return to the case-by-case review embodied in the previously BOP regulation might not constitute an additional administrative burden because prison officials are required under the Ensign Amendment to examine each publication and determine whether it is "sexually explicit or features nudity") (Ward, J., dissenting). Contrary to the decision in Amatel, we believe this to be a case in which factual development is necessary for evaluating the Ensign Amendment and its implementing regulation under Turner. See Wolf, 297 F.3d at 310 ("[C]ourts of appeals ordinarily remand to the trial court where the Turner factors cannot be assessed because of an undeveloped record.") (citing Doe v. Delie, 257 F.3d 309, 317 (3d Cir.2001)).

For the reasons discussed above, we find that the District Court erred in determining that the Ensign Amendment and its implementing regulation were reasonably related to the legitimate government interest of rehabilitation without an adequate factual basis for so doing.5 Accordingly, we will reverse the judgment of the District Court entered on February 28, 2002 and remand with instructions to conduct an appropriate proceeding before reevaluating the amendment and regulation under Turner.