Judge WINTER dissents in a separate opinion.
CARDAMONE, Circuit Judge.Plaintiffs Janet Ramos and her sons, Angel and Richard, residents of the Town of Vernon, Connecticut, brought suit *172against the Town of Vernon and its police chief to challenge the constitutionality of Vernon’s juvenile curfew ordinance. Among other constitutional claims, plaintiffs allege that the ordinance violates minors’ equal protection rights. The case was tried in the United States District Court for the District of Connecticut before Judge Alan H. Nevas, who ruled against plaintiffs and denied their request for declaratory relief and their motion for a preliminary injunction, see Ramos ex rel. Ramos v. Town of Vernon, 48 F.Supp.2d 176 (D.Conn.1999), in a judgment dated December 27, 2000. From this judgment, plaintiffs appeal.
Juvenile curfews have existed throughout our Nation’s history, and we do not question the Town of Vernon’s authority to have such an ordinance under some circumstances. But it is not the case that what the town wills is permissible as of course. Instead, to be upheld as a valid exercise of state power, the curfew’s enactment must have been done right. The constitutionality of a curfew is determined by balancing the recognized interests the state has in protecting children and fighting crime against the constitutional right of all citizens, including juveniles, to move about freely. Here, Vernon’s curfew interferes with juveniles’ freedom of movement, that is, their right with parental consent to walk the streets, move about at will, meet in public with friends, and leave their houses when they please. This right to free movement is a vital component of life in an open society, both for juveniles and adults.
After consideration of these interests, we think the present Town of Vernon ordinance infringes on the equal protection rights of minors. Hence, we declare it unconstitutional and reverse the district court’s judgment, and we remand to the district court.
BACKGROUND
A. Vernon’s Curfew Ordinance
On August 2, 1994 the Vernon Town Council adopted its first curfew ordinance. That ordinance makes it unlawful for any person under 18 years of age “to remain, idle, wander, stroll or play in any public place or establishment in the Town during curfew hours.” From Sunday through Thursday the curfew is in effect from 11:00 p.m. until 5:00 a.m. the next day. On Friday and Saturday nights, the curfew begins at 12:01 a.m. and ends at 5:00 a.m. the next day.
The ordinance includes several exceptions permitting a minor to be out in public during curfew hours, if: (1) “accompanied by a parent, guardian, custodian or other adult person having custody or control of such minor”; (2) “on an emergency errand”; (3) engaged in a “specific business or activity directed or permitted by his parent, guardian, or other adult person having the care and custody of the minor”; or (4) the minor’s presence “is connected with or required by some legitimate employment, trade, profession or occupation.” The ordinance also excepts from its strictures minors attending, with parental permission, a “special function or event sponsored by any religious, school, club or other organization.” On December 15, 1998, while this suit was pending, another exception was added to the ordinance to exclude from coverage any minor “exercising his/her first amendment rights.”
At the trial testing the curfew’s constitutionality, two former members of the town council testified that prior to enacting the ordinance, the town’s elected officials had begun to notice groups of young people gathering in certain parts of town. In June 1994, two months before the August *173enactment of the ordinance, a 16-year-old Vernon resident was murdered. Further, the returns from a 1994 youth survey distributed to school students in Vernon indicated they were concerned about gangs, guns and violence. With these circumstances in mind, the town council enacted the curfew ordinance to “protect minors from each other and from other persons on the streets during nocturnal hours,” to “promote parental responsibility for and supervision of minors” and to “protect the general public from nocturnal mischief and crime committed by minors.”
The punishment for violating the curfew depends on the age of the offender. A minor 16 or older may be cited and fined up to $50 for a first offense, $75 for the second offense and $90 for all subsequent offenses. Failure to respond to a curfew ticket has resulted in a minor’s arrest. A minor under the age of 16 is not subject to a fíne, but the ordinance directs police to issue a warning, send the minor home and report the incident. If a minor under the age of 16 refuses to cooperate or is a repeat curfew violator, he or she may be taken to the police station and held there until retrieved by a parent or adult acting in loco parentis.
Moreover, the ordinance makes it unlawful for a “parent, guardian or other adult person having custody or control of any minor under the age of sixteen ... to suffer or permit or by inefficient control to allow” such a minor to violate the curfew. An adult in violation of this provision may be cited and fined, but adults are only in violation if a minor is under the age of 16.
B. Plaintiffs’ Claims
Plaintiff Angel Ramos was found in violation of the curfew ordinance on numerous occasions because he was out past curfew hours with general permission from his mother to be out, but without permission to be on a specific errand, or out pursuant to any of the enumerated exceptions in the ordinance. Plaintiff Richard Ramos claims that his rights under the Fourteenth Amendment are chilled due to the ordinance and that he must run home when out past curfew, even with permission, for fear of being found in violation. In their complaint plaintiffs assert the ordinance is unconstitutionally vague and violates the First, Fourth and Fourteenth Amendments to the United States Constitution. Specifically, they contend that several of the exceptions in the ordinance are too vague to provide notice of prohibited conduct or to guide enforcement efforts. Richard and Angel Ramos further allege the curfew infringes on minors’ rights to free speech and association under the First Amendment and to equal protection under the Fourteenth Amendment. In addition, they aver enforcement of the ordinance violates their Fourth Amendment guarantee to be free from unlawful searches and seizures. Finally, Janet Ramos argues that the curfew’s restrictions interfere with parents’ due process right to raise their children as they see fit, because she as well as several other parents want the freedom to allow their juvenile children to be out late at night under less stringent restrictions than provided for in the town’s ordinance.
Plaintiffs also declare in their complaint that the Vernon ordinance violates several provisions of the Connecticut Constitution. These claims are comparable to the federal claims plaintiffs make, but we need not concern ourselves with them because they are not before us on appeal.
C. Proceedings Below
Because plaintiffs moved for a preliminary injunction against enforcement of the curfew, the district court combined a hearing on that motion with a trial on the merits. The four-day bench trial featured testimony from Janet, Angel and Richard *174Ramos, police officials, other parents in Vernon, former town officials and experts for both sides.
After the trial concluded, the district court dismissed as moot the case brought by Angel Ramos because he had turned 18 and therefore was no longer subject to the curfew. With respect to the state constitutional claims, the district court certified six questions to the Connecticut Supreme Court, which later ruled that the Connecticut Constitution provides no independent grounds on which to invalidate the ordinance. See Ramos v. Town of Vernon, 254 Conn. 799, 761 A.2d 705 (2000) (rejecting all plaintiffs’ claims under the Connecticut Constitution). In answering the certified questions, the Connecticut Supreme Court avoided reviewing the federal constitutional claims, considering only whether the Connecticut Constitution provided any extra protections that required striking down the curfew ordinance. Id. at 818-20, 761 A.2d 705. Upon receipt of the Connecticut Supreme Court’s response to the certified questions, the district court entered final judgment in favor of the defendants. Plaintiffs then filed this appeal.1
DISCUSSION
I Standard of Review
Challenges to the constitutionality of a local ordinance are subject to de novo review. See, e.g., Myers v. County of Orange, 157 F.3d 66, 74 (2d Cir.1998). We accept the district court’s findings of fact, however, unless they are clearly erroneous. See United States v. El-Hage, 213 F.3d 74, 79 (2d Cir.2000) (per curiam).
II Equal Protection
Although all of plaintiffs’ federal constitutional claims are before us on this appeal, we focus particularly on Richard Ramos’ claim that the curfew ordinance violates the Equal Protection Clause of the Fourteenth Amendment. The Fourteenth Amendment to the United States Constitution guarantees that “[n]o state shall ... deny to any person within its jurisdiction the equal protection of the laws.” This means the state must treat similarly situated individuals similarly, in the absence of an adequate reason to distinguish between them. As with all equal protection claims, we begin our analysis by ascertaining the appropriate level of scrutiny. See Dunn v. Blumstein, 405 U.S. 330, 335, 92 S.Ct. 995, 31 L.Ed.2d 274 (1972).
A. Levels of Scrutiny Defined
When a legislative enactment has been challenged on equal protection grounds, *175one standard of review is rational basis review, which requires that the law be rationally related to a legitimate government interest. See Allied Stores of Ohio, Inc. v. Bowers, 358 U.S. 522, 528-29, 79 S.Ct. 437, 3 L.Ed.2d 480 (1959); Burke Mountain Acad., Inc. v. United States, 715 F.2d 779, 783 (2d Cir.1983). A law will survive this level of scrutiny unless the plaintiff proves that the law’s class-based distinctions are wholly irrational. See, e.g., Hodel v. Indiana, 452 U.S. 314, 331-32, 101 S.Ct. 2376, 69 L.Ed.2d 40 (1981); see also Lewis v. Thompson, 252 F.3d 567, 582 (2d Cir.2001) (explaining that rational speculation, as opposed to facts, can justify class-based distinction under rational basis test). But see Romer v. Evans, 517 U.S. 620, 116 S.Ct. 1620, 134 L.Ed.2d 855 (1996) (applying less deferential form of rational basis review); Able v. United States, 155 F.3d 628, 634 (2d Cir.1998) (noting same); City of Cleburne, 473 U.S. at 456, 458, 105 S.Ct. 3249 (Marshall, J., concurring in the judgment in part and dissenting in part) (noting majority’s analysis of the equal protection claim of developmentally disabled more searching than traditional rational basis test).
A heightened level of review — strict scrutiny — applies when legislation discriminates on the basis of a person’s membership in a suspect class or when it burdens a group’s exercise of a fundamental right. Plyler v. Doe, 457 U.S. 202, 216-17, 102 S.Ct. 2382, 72 L.Ed.2d 786 (1982); see, e.g., City of Richmond v. J.A. Croson Co., 488 U.S. 469, 493-94, 109 S.Ct. 706, 102 L.Ed.2d 854 (1989) (applying strict scrutiny to program that discriminated on basis of race, a suspect classification); Harper v. Va. Bd. of Elections, 383 U.S. 663, 668-70, 86 S.Ct. 1079, 16 L.Ed.2d 169 (1966) ($1.50 poll tax unconstitutional because, though facially neutral, it discriminated on basis of wealth in allocation of right to vote); Skinner v. Okla, ex rel. Williamson, 316 U.S. 535, 541-42, 62 S.Ct. 1110, 86 L.Ed. 1655 (1942) (invalidating sterilization law because it impinged on the fundamental right to marriage and procreation of some criminals, but not others). To satisfy strict scrutiny, the government must show the law is narrowly tailored to achieve a compelling governmental interest. Adarand Constructors, Inc. v. Pena, 515 U.S. 200, 227, 115 S.Ct. 2097, 132 L.Ed.2d 158 (1995).
More recently, the Supreme Court has developed an intermediate level of scrutiny that lies “[bjetween [the] extremes of rational basis review and strict scrutiny.” Clark v. Jeter, 486 U.S. 456, 461, 108 S.Ct. 1910, 100 L.Ed.2d 465 (1988). Intermediate scrutiny typically is used to review laws that employ quasi-suspect classifications, United States v. Coleman, 166 F.3d 428, 431 (2d Cir.1999) (per curiam), such as gender, Craig v. Boren, 429 U.S. 190, 197, 97 S.Ct. 451, 50 L.Ed.2d 397 (1976), or legitimacy, Mills v. Habluetzel, 456 U.S. 91, 98-99, 102 S.Ct. 1549, 71 L.Ed.2d 770 (1982). On occasion intermediate scrutiny has been applied to review a law that affects “an important, though not constitutional, right.” Coleman, 166 F.3d at 431; cf. Plyler, 457 U.S. at 223, 102 S.Ct. 2382 (applying, without labeling it as such, an intermediate form of scrutiny to review of a law that implicated right to education). Under intermediate scrutiny, the government must show that the challenged legislative enactment is substantially related to an important governmental interest. Wengler v. Druggists Mut. Ins. Co., 446 U.S. 142, 150, 100 S.Ct. 1540, 64 L.Ed.2d 107 (1980); Craig, 429 U.S. at 197, 97 S.Ct. 451.
B. Choice of Intermediate Scrutiny in District Court
Plaintiffs press for heightened scrutiny on the grounds that the curfew *176ordinance, which employs an age-based classification, implicates a fundamental right. More specifically, plaintiffs argue that the curfew ordinance impinges on the exercise of the right to free movement and “to travel freely within the state and within one’s own community.” Defendants— avoiding the question of how to define the relevant interest and conceding that minors may have a right to freedom of movement under some circumstances — focus instead on whether strict or intermediate scrutiny should be the standard applied to the curfew ordinance. The district court in its review used intermediate scrutiny and defendants do not take issue with that ruling on this appeal.2 We have also concluded that intermediate scrutiny is the appropriate standard, though our reasons and conclusion differ from those of the district court. To explain how we reached this conclusion we turn to a discussion of the rights implicated by the Vernon ordinance and the class that is burdened.
Ill Our Rationale for Intermediate Scrutiny
A. Right to Free Movement
We begin our analysis with plaintiffs’ assertion that Vernon’s curfew ordinance implicates the constitutional right to free movement and intrastate travel. The right to intrastate travel, or what we sometimes will refer to as the right to free movement, has been recognized in this Circuit. King v. New Rochelle Mun. Hous. Auth., 442 F.2d 646, 648 (2d Cir.1971); see also Spencer v. Casavilla, 903 F.2d 171, 174 (2d Cir.1990) (observing that this Circuit “has held that the Constitution ... protects the right to travel freely within a single state”); cf. Kolender v. Lawson, 461 U.S. 352, 358, 103 S.Ct. 1855, 75 L.Ed.2d 903 (1983) (recognizing “the constitutional right to freedom of movement”). Because the curfew limits the constitutional right to free movement within the Town of Vernon at certain hours of the night, we assume that were this ordinance applied to adults, it would be subject to strict scrutiny. Analysis in this case is more complicated because Vernon’s ordinance targets juveniles and we have not yet determined whether children, like adults, possess the right to intrastate travel or, if they do, how such right is impacted by their age.3
B. Differing Analytical Approaches to Minors’ Rights
Courts addressing the constitutionality of juvenile curfew ordinances have incorporated the plaintiffs’ status as minors into the equal protection framework in three different ways. The first approach defines the relevant interest so narrowly that it is not deemed a constitutional right and heightened scrutiny does not come into play. Under this methodology, the characteristic that defines the plaintiffs’ class— youth — divests them of a right they would otherwise hold. See, e.g., Hutchins v. District of Columbia, 188 F.3d 531, 538-39 (D.C.Cir.1999) (en banc) (plurality opinion). The second approach recognizes that children, like adults, have a constitutional right to free movement, but then reduces the level of scrutiny to compensate for children’s special vulnerabilities. Courts that favor this approach have ruled that the unique interrelationship between minors and the state renders strict scrutiny *177inappropriate. See, e.g., Hutchins, 188 F.3d at 563 (Rogers, J., concurring in part and dissenting in part [hereinafter Rogers, J.]); Schleifer ex rel. Schleifer v. City of Charlottesville, 159 F.3d 843, 847 (4th Cir.1998).
The third approach assumes that once a constitutional right has been recognized, its exercise by minors should be protected by strict scrutiny, just as it is for adults. Rather than using children’s status to divest them of rights or to weaken the formal protections of those rights, courts taking this third approach factor in the unique attributes of minors in determining whether the government has a compelling interest justifying restrictions on minors’ freedoms. See, e.g., Schleifer, 159 F.3d at 863 (Michael, J., dissenting); Nunez ex rel. Nunez v. City of San Diego, 114 F.3d 935, 946 (9th Cir.1997); see also Qutb v. Strauss, 11 F.3d 488, 492 (5th Cir.1993) (assuming, without deciding, that a fundamental right is implicated and applying strict scrutiny).
1. Rational Basis Review Rejected. An example of the first approach is in the plurality opinion in Hutchins v. District of Columbia. It held that a juvenile curfew, like the one in this case, did not impinge upon the exercise of the fundamental rights of minors. In so holding, Hutchins defined the relevant interest as juveniles’ “right to be on the streets at night without adult supervision.” 188 F.3d at 538. Importantly, this formulation of the right to free movement incorporates two controversial elements: the class characteristic of plaintiffs — their age — and the specific manner in which they might exercise their freedom of movement' — at night without supervision. Id. at 557 (Rogers, J.). For the following reasons, we do not accept this approach.
The latter element — the manner in which the right would be exercised — is one that defines the interest too narrowly at the outset. If the Hutchins formulation of the interest were correct, then juveniles’ constitutional rights would appear in the morning and disappear at night. But daylight and darkness are not related to whether a constitutional right exists. Instead, they are relevant to the strength of the government’s interest in regulating the manner in which minors exercise their rights. If nighttime is more dangerous, the government’s interest in protecting minors is stronger. If not, then nighttime has no constitutional significance. The same reasoning follows with respect to adult supervision: the presence or absence of supervision is relevant to the government’s interest in protecting minors from danger, but the right to free movement itself does not magically appear and disappear with an adult’s presence. Cf. King, 442 F.2d at 648-49 (adopting broad view of right in question, then analyzing specifics under strict scrutiny).
The more difficult issue related to defining the right is whether juveniles are defi-nitionally excluded altogether from the right to intrastate travel. While the characteristic that defines the burdened class is not typically relevant to a constitutional rights inquiry, juveniles occupy a unique position in our constitutional scheme. Bellota v. Baird, 443 U.S. 622, 633, 99 S.Ct. 3035, 61 L.Ed.2d 797 (1979) (plurality opinion). “[Although children generally are protected by the same constitutional guarantees against governmental deprivations as are adults, the State is entitled to adjust its legal system to account for children’s vulnerability and their [other] needs.... ” Id. at 635, 99 S.Ct. 3035. Accordingly, the Supreme Court has protected children from governmental infringement of their constitutional rights in a number of instances. See Breed v. Jones, 421 U.S. 519, 541, 95 S.Ct. 1779, 44 L.Ed.2d 346 (1975) *178(double-jeopardy); Goss v. Lopez, 419 U.S. 565, 574, 95 S.Ct. 729, 42 L.Ed.2d 725 (1975) (due process prior to deprivation of certain property interests); In re Gault, 387 U.S. 1, 41, 55, 87 S.Ct. 1428, 18 L.Ed.2d 527 (1967) (right to counsel and privilege against self-ineiimination).
Even in eases where the Court has upheld the infringement of such rights, the decisions do not summarily state the minor has no interest worth protecting, but rather they reason that the government’s interest and the special status of minors justify the incursion. See McKeiver v. Pennsylvania, 403 U.S. 528, 545, 91 S.Ct. 1976, 29 L.Ed.2d 647 (1971) (holding that juvenile court need not provide jury trial); cf. Erznoznik v. City of Jacksonville, 422 U.S. 205, 212-14, 95 S.Ct. 2268, 45 L.Ed.2d 125 (1975) (citing cases where incursions on minors’ First Amendment rights upheld).
Simply denying the existence of a constitutional right is too blunt an instrument to resolve the question of juvenile rights to freedom of movement. Instead, once such a right has been acknowledged, the equal protection framework allows for a more discerning inquiry to accommodate competing interests. Therefore, we prefer to admit minors to the protected zone and then engage in a balancing of constitutional rights and children’s vulnerabilities. This seems particularly appropriate in the case at hand, given that “[t]he rights of locomotion, freedom of movement, to go where one pleases, and to use the public streets in a way that does not interfere with the personal liberty of others,” do not turn solely on the circumstances of childhood itself. Bykofsky v. Borough of Middletown, 401 F.Supp. 1242, 1254 (M.D.Pa.1975), aff'd, 535 F.2d 1245 (3d Cir.1976) (unpublished table decision).
2. Strict Scrutiny Rejected. Having rejected the first approach — -and thus rational basis review — we must decide between the second and third approaches, which employ intermediate scrutiny and strict scrutiny, respectively. Whether we review the ordinance using strict or intermediate scrutiny, we face the inherent tension between the government’s compelling interest in protecting minors, on the one hand, and children’s interests in freely moving about their own community, on the other. See Hutchins, 188 F.3d at 555 (Rogers, J.). As we shall see, strict scrutiny poses an additional problem that arises from the unusual interplay between rights-based and class-based equal protection doctrines. Adopting intermediate review, in our view, best rationalizes this tension.
As discussed, the level of scrutiny in constitutional rights cases typically is determined by the right, not the class, affected. Thus, for example, the state is prohibited from restricting the interstate travel of people with brown hair to the same extent that it is prohibited from restricting the travel of women. In equal protection cases involving legislative burdens on protected classes, class membership almost always is used, if at all, to justify raising the level of scrutiny. Hence, lowering the level of review on account of children’s minority status is a relatively unusual step, cf. Turner v. Safley, 482 U.S. 78, 89-90, 107 S.Ct. 2254, 96 L.Ed.2d 64 (1987) (applying lower level of scrutiny to prison regulations that impinge upon prisoners’ constitutional rights); Able, 155 F.3d at 633-34 (explaining courts afford great deference to laws that infringe on the rights of military personnel and noting this is atypical), in light of the usual operation of the rights-based and class-based equal protection doctrines. Nevertheless, we believe that these adjustments are not only logical, but necessary in the circumstances of this case. For the following reasons, we decline to apply strict scrutiny and hence *179reject the third approach to analyzing juvenile curfew ordinances.
Strict scrutiny, when applicable, is a highly restrictive test that embodies a constitutional preference for “blindness.” More specifically, in the context of the rights-based equal protection doctrine, it reflects the notion that some rights are so important that they should be afforded to individuals in a manner blind to all group classifications, absent the most compelling reasons to do otherwise. See, e.g., Harper, 383 U.S. at 670, 86 S.Ct. 1079 (holding “the right to vote is too precious, too fundamental” to be conditioned on ability to pay even small fee). Similarly, in the context of suspect classes, strict scrutiny views some classifications, such as race, as so pernicious that society should be blind to them in all but rare situations. See Adarand Constructors, Inc., 515 U.S. at 228-29, 115 S.Ct. 2097; see also Fullilove v. Klutznick, 448 U.S. 448, 532, 100 S.Ct. 2758, 65 L.Ed.2d 902 (1980) (Stewart, J., dissenting) (By “making race a relevant criterion” the government teaches that people should “view themselves and others in terms of their racial characteristics.”). Accordingly, even when a legislative scheme employs a classification with the best intentions (e.g., “benign” racial discrimination), it is usually only permitted if it helps diminish the continuing relevancy of the classification. See Adarand Constructors, Inc., 515 U.S. at 237, 115 S.Ct. 2097 (explaining that government can use suspect classification to remedy “both the practice and the lingering effects of racial discrimination”).
In contrast, when blindness to a classification is not the desired end, but there are still constitutional concerns akin to those that justify strict scrutiny, the Supreme Court has adopted the more flexible, yet still searching, intermediate form of review. For example, in the context of the class-based equal protection framework, the Court has explicitly repudiated complete blindness with regard to gender-based laws, reasoning that, although such laws elicit some suspicion, the physical differences between the sexes are relevant and enduring. See United States v. Virginia, 518 U.S. 515, 533, 116 S.Ct. 2264, 135 L.Ed.2d 735 (1996); Michael M. v. Superior Court, 450 U.S. 464, 480, 101 S.Ct. 1200, 67 L.Ed.2d 437 (1981) (plurality opinion); see also Zachary Potter & C.J. Summers, Reconsidering Epistemology and Ontology in Status Identity Discourse, 17 Harv. BlaekLetter L.J. 113, 165-66 (2001) (collecting cases where sex-based rule exempted from anti-discrimination law and explaining manner in which they support recognition of gender differences). Similarly, in rare circumstances, blindness to classes is not always a goal in the allocation of constitutional rights. For example, for obvious reasons, the Supreme Court applies a deferential standard of review to laws that implicate the constitutional rights of prisoners. Turner, 482 U.S. at 89, 107 S.Ct. 2254. In such cases, strict scrutiny is not appropriate, even though significant constitutional interests are at stake.
Given that the inherent differences between children and adults, both mental and physical, remain cause for concern, the Supreme Court has indicated that youth-blindness is not a goal in the allocation of constitutional rights. See Bellotti, 443 U.S. at 635, 99 S.Ct. 3035 (discussing cases that implicate fundamental rights and holding that juveniles “constitutionally may be treated differently from adults”); id. at 634, 99 S.Ct. 3035 (explaining that children’s immaturity, vulnerability, and place within a family justify the different constitutional treatment); -Laurence Tribe, American Constitutional Law § 16-31, at 1589 (2d ed.1988) (recognizing that strict scrutiny is generally incompatible with so*180ciety’s desire to give children a “special place”). Youth-blindness is not a constitutional goal because, even with regard to fundamental rights, failing to take children’s particular attributes into account in many contexts, such as marriage, would be irresponsible. See Hutchins, 188 F.3d at 559 (Rogers, J.) (observing that children, below a certain age, do not have the developmental prerequisites for the bundle of rights and responsibilities that accompany marriage). Hence, strict scrutiny would appear to be too restrictive a test to address government actions that implicate children’s constitutional rights. Cf. Patryk J. Chudy, Note, Doctrinal Reconstruction: Reconciling Conflicting Standards in Adjudicating Juvenile Curfew Challenges, 85 Cornell L.Rev. 518, 556-57 (2000) (“The decisions that [have] articulated strict scrutiny as the appropriate standard contain several textual references indicating some decrease in the standard of review.”). Consequently, we choose the second of the three approaches described above and apply intermediate scrutiny.
C. Intermediate Scrutiny
It bears repeating that to satisfy intermediate scrutiny, the state must show that the challenged classification serves “important governmental objectives and that the discriminatory means employed [are] substantially related to the achievement of those objectives.” Wengler, 446 U.S. at 150,100 S.Ct. 1540.
In the context of minors’ rights, an important governmental objective would, at the very least, address the vulnerabilities particular to minors. See Bellotti, 443 U.S. at 635, 99 S.Ct. 3035 (stating general rule that children “are protected by the same constitutional guarantees against governmental deprivations as are adults”); Planned Parenthood of Cent. Mo. v. Danforth, 428 U.S. 52, 75, 96 S.Ct. 2831, 49 L.Ed.2d 788 (1976) (analyzing whether infringement on minors’ right to privacy could be justified by a “significant state interest ... that is not present in the case of an adult ” (emphasis added)). Hence, if minors have the capacity to exercise a right and face the same risks and benefits as adults when doing so, a reviewing court should be skeptical of the argument that minors alone need protection.
Identifying the true beneficiaries of a restriction of this sort is particularly important in assessing both the legitimacy of the government’s objectives and the relationship of those objectives to the means employed to achieve them. Cf. Orr v. Orr, 440 U.S. 268, 282-83, 99 S.Ct. 1102, 59 L.Ed.2d 306 (1979) (performing such analysis for sex-based classification purportedly benefitting vulnerable women). If the direct and primary beneficiaries are children, then the constraint on liberty is more likely to pass constitutional muster. See McKeiver, 403 U.S. at 552, 91 S.Ct. 1976 (White, J., concurring) (discussing ways juvenile justice system operates to benefit of minors). Conversely, when evidence suggests that a curfew targeting juveniles was passed for the benefit of others in the community, that law’s constitutionality is more suspect. For example, testimony indicating that the restrictions were passed because adult residents are uncomfortable with the lifestyles of some juveniles tends to undermine the legitimacy of the restrictions. Minors may on occasion look different from and act differently than adults, but these differences are ordinarily not appropriate reasons for enacting a curfew law that burdens a particular class of people. See Romer, 517 U.S. at 633, 116 S.Ct. 1620 (“Respect for [the] principle [of equal protection of the laws] explains why laws singling out a certain class of citizens for disfavored legal status or general hardships are rare.”).
*181It may be tempting to label the entire class of minors in a certain way; such an approach would make the case easier to analyze. But in applying the intermediate scrutiny standard of review, we avoid relying on stereotypes and assumptions about young people. See In re Gault, 387 U.S. at 29-30, 87 S.Ct. 1428 (“So wide a gulf between the State’s treatment of the adult and of the child [in the criminal justice system] requires a bridge sturdier than mere verbiage, and reasons more persuasive than cliché can provide.”). Moreover, generalizations about youth do not uniformly suggest lowering the level of scrutiny. For instance, some might consider minors, as a group, to be immature and dependent and use these considerations to justify, a juvenile curfew ordinance. However these very same considerations — immaturity and dependency — also impede minors’ relative ability, as a class, to articulate or mount an effective defense against such a restriction. Juveniles also lack the right to vote. Without an independent yoke in legislative decisionmaking, minors must rely on others to ensure adequate protection of their rights. This consideration places youth outside “those political processes ordinarily ... relied upon to protect minorities.” See United States v. Carolene Prods. Co., 304 U.S. 144, 153 n. 4, 58 S.Ct. 778, 82 L.Ed. 1234 (1938).
We mention these considerations of political powerlessness only to illustrate the unreliability of assumptions and generalizations as justification for laws infringing on the constitutional rights of minors. We do not conclude that youth are a suspect class; there is no need even to consider this prong of the equal protection framework in this case.4
For all of these reasons, we believe intermediate scrutiny is sufficiently skeptical and probing to provide the rigorous protection that constitutional rights deserve. At the same time, it is flexible enough to accommodate legislation carefully drafted to account for “children’s vulnerability and their needs.” Bellotti, 443 U.S. at 635, 99 S.Ct. 3035. Consequently, it is the level of review we apply to the town ordinance in this case.
IV Review of Vernon’s Curfew Ordinance
A. Strength of Vernon’s Interest
We pass finally to an assessment of the ordinance. The first consideration is whether it furthers an important interest of the Town of Vernon. The ordinance has three stated goals — protecting minors from harm at night, protecting the general population from nighttime juvenile crime, and promoting responsible parenting. We recognize the government has an important interest in protecting all its citizens from crime. Schall v. Martin, 467 U.S. 253, 264, 104 S.Ct. 2403, 81 L.Ed.2d 207 (1984). Further, we acknowledge this interest may take on added strength in light of attributes particular to children. See id. at 264-65, 104 S.Ct. 2403; Nunez, 114 F.3d at 946-47; Qutb, 11 F.3d at 492; Eclipse Enters., Inc. v. Gulotta, 134 F.3d 63, 67 & n. 1 (2d Cir.1997); see also Hodgson v. Minnesota, 497 U.S. 417, 444, 110 S.Ct. 2926, 111 L.Ed.2d 344 (1990) (“The State has a strong and legitimate interest in the welfare of its young citizens, whose *182immaturity, inexperience, and lack of judgment may sometimes impair their ability to exercise their rights wisely.”).
Plaintiffs concede the first two goals of the ordinance — protecting minors from harm and protecting the community from nighttime juvenile crime — are important governmental interests. A third goal of the curfew ordinance — promoting parental responsibility — is one defendants do not elaborate upon. While we agree that the goal of encouraging parental responsibility is an admirable one, we cannot help but observe the irony of the supposition that responsible parental decisionmaking may be promoted by the government removing decisionmaking authority from responsible parents and exercising that authority itself. Defendants do not elaborate upon this interest, and we note that the state has the burden of justifying a restriction under intermediate scrutiny. For all of the reasons set forth infra under “B. Sub-stantiality of Relationship,” the relationship defendants presumably assert between the goal of promoting responsible parenting and the criminalization of late night youth activity fails to meet the burden. However, since the dissent elaborates the town’s arguments for them, we address them here.
We recognize that whatever right to free movement a minor may have, he or she may not assert it in the face of parental prohibition against it. See Reno v. Flores, 507 U.S. 292, 302, 113 S.Ct. 1439, 123 L.Ed.2d 1 (1993); Vervonia Sch. Dist. 47J v. Acton, 515 U.S. 646, 654, 115 S.Ct. 2386, 132 L.Ed.2d 564 (1995). However, it does not follow from this that a minor’s right to travel may exist only if that minor has a parent’s specific permission to be on a particular errand to a certain place at some definite time. We therefore understand the right that the plaintiffs in this case seek to vindicate to be the right of minors to move about freely if their parents do not forbid them from doing so.
We disagree with the dissent’s assertion that juveniles out in violation of the letter of the curfew are either out against the express wishes of their parents or, if not, the parents are negligent in granting their children such general permission. The dissent suggests that activity-specific permission, provided for in the ordinance, is sufficient to protect any right of mobility of minors, but that permission to be out and about at certain hours is per se “neglectful” or not “fit” parenting, and the municipal ordinance furnishes authority to guide juvenile activity in place of parental decisionmaking. Surely a parent can decide that a child must generally be home between midnight and five in the morning. But a parent can also decide that a child has general permission to be outside the home late at night. Although we ourselves may not adhere to such a philosophy of parenting, a parent may ascribe value to granting a child freedom to move about the neighborhood' — even if fettered somewhat by ordinary parental admonitions such as “don’t drink alcohol,” “don’t get into trouble,” “drive safely.” Deciding so may be a parent’s way of preparing a child for adult life.
A parent has discretion to decide whether to allow his or her child to be out late at night, for it is the parents “whose primary function and freedom include preparation for obligations the state can neither supply nor hinder.” Prince v. Massachusetts, 321 U.S. 158, 166, 64 S.Ct. 438, 88 L.Ed. 645 (1944); see also Pierce v. Society of Sisters, 268 U.S. 510, 535, 45 S.Ct. 571, 69 L.Ed. 1070 (1925) (“The child is not the mere creature of the State; those who nurture him and direct his destiny have the right, coupled with the high duty, to recognize and prepare him for additional obligations.”).
*183The state may, of course, intervene to prevent child neglect, provide for education, and assume responsibility when parents fail. See Flores, 507 U.S. at 302, 113 S.Ct. 1439; see also Conn. Gen.Stat. § 17a-101g (2001). But where, as here, there has been no suggestion of specific dangers arising out of the kind of parenting in question, and there has been no adjudication that the parent is unfit, we presume that the parent acts in the best interest of her children. Cf. Troxel v. Granville, 530 U.S. 57, 68, 120 S.Ct. 2054, 147 L.Ed.2d 49 (2000) (noting the importance of fact that a court had adjudicated parent unfit, because fit parents are presumed to act in the best interests of their children); see also Parham v. J. R., 442 U.S. 584, 602, 99 S.Ct. 2493, 61 L.Ed.2d 101 (1979).
Indeed, here even the Town of Vernon does not assert that a parent’s bestowal of general permission for children to be out late at night is per se unfit parenting. Perhaps this is implied by the intent behind the curfew ordinance of encouraging responsible parenting, as the dissent asserts. However, absent a threshold showing by the town of possible harm to children sufficient to override parental due process rights, or absent adjudication of parental unfitness, we cannot sit in judgment of a parental philosophy allowing late night activity, for “between parents and judges, the parents should be the ones to choose whether to expose their children to certain people or ideas.” Troxel, 530 U.S. at 63, 120 S.Ct. 2054; see also Flores, 507 U.S. at 301-02, 113 S.Ct. 1439 (acknowledging that governmental interference with fundamental rights requires adequate justification by the state).
If a parent decides not to limit a child’s mobility, and if this decision is not “unfit” parenting warranting state intervention, then the child has a right to free movement. The government may limit this right, but not without showing a substantial relationship of the limitation to an important objective.5
B. Substantiality of Relationship
To determine whether there is an adequate justification for the curfew ordinance we examine whether the burdens imposed by it are substantially related to the government’s interests as just outlined. See Miss. Univ. for Women v. Hogan, 458 U.S. 718, 724-25, 102 S.Ct. 3331, 73 L.Ed.2d 1090 (1982). Although the government need not produce evidence to a scientific certainty of a substantial relationship, see, e.g., Ginsberg v. New York, 390 U.S. 629, 642-43, 88 S.Ct. 1274, 20 L.Ed.2d 195 (1968), it carries the burden of proving such a relationship. The Supreme Court has explained that “[t]he purpose of requiring [proof of] that close relationship is to assure that the validity of a classification is determined through reasoned anal*184ysis rather than through the mechanical application of traditional, often inaccurate, assumptions.” Hogan, 458 U.S. at 725-26, 102 S.Ct. 3331 (emphasis added). We agree with the D.C. Circuit’s statement that “[i]n judging the closeness of the relationship between the means chosen (the curfew), and the government’s interest,” three interrelated concepts must be considered: the factual premises which prompted the legislative enactment, the logical connection between the remedy and those factual premises, and the breadth of the remedy chosen. Hutchins, 188 F.3d at 542. We discuss each factor.
1. Factual Premises
Unfortunately, we do not have the benefit of a documentary record of the town council’s discussions regarding the need for a curfew or the research that went into deciding the scope of the curfew. Instead, we must rely on the goals written into the curfew ordinance itself and the testimony of individuals who played a role in its adoption. Drawing from those sources, the factual premises considered by the Vernon Town Council can be broken down into two types. First, groups of young people had been seen gathering on the streets. Second, police and some members of the community perceived an increase in gang activity locally. Part of this perception was based upon the murder of a 16-year-old Vernon resident. See Ramos, 48 F.Supp.2d at 185.
With respect to the first type of evidence, two former town officials testified. Steven Wakefield, a defense witness, was deputy mayor of the Town of Vernon from 1991 until 1995. He stated that between 1992 and 1993 “[t]here seemed to be a general increase of people, primarily younger types. They were starting to gather in certain areas in the Rockville section of Vernon ... pretty much days and nights, and sometimes they became more visible. This was in warmer weather.”
Wakefield further testified that he had become aware of this situation both from his own observations and through “calls [to City Hall] from people who told us they had concerns to the point of being scared of traversing some of these areas.” Wake-field also reported that some townspeople had attended town council meetings and lodged complaints about “juvenile activity.” The proof at trial presented by defendants did not disclose the number of people who had voiced these concerns or the precise nature of their concerns. For instance, information such as the time of day or ages of the offending individuals was not offered to support the factual premises.
Thomasina Clemons, a member of the Vernon Town Council prior to and at the time of the enactment of the curfew ordinance, also testified for the defense. While going to and from work, she had observed “[a]n increase in the number of people on the street and [an] increase in the number of younger people” on certain streets in Vernon. When asked to describe the changes brought about by the curfew ordinance, Clemons stated that pri- or to the adoption of the curfew, she had stopped taking walks on Sunday mornings because she “saw people who look[ed] like skinheads ... [that] were new to the environment and then later, [she] started seeing other young people.” Clemons readily volunteered, however, that she “didn’t see them doing anything but ... sort of being there.” As for post-curfew improvements, Clemons said that she had no knowledge or personal opinion on that subject.
In June 1994 two months prior to the adoption of the curfew ordinance, a 16-year-old Vernon resident was murdered in his home. Although the victim was a former gang member, a report prepared by defendants’ expert indicates that the crime *185was probably a result of a robbery and not gang-related. Wakefield testified he had been shocked by the murder and related
I had never anticipated or heard of anything like that. I lived in Vernon since I was five years old. It was the first time I became aware of anything like that in our backyard. It just seemed like this thing was starting to escalate and get to [the] point where it was non-controllable.
Plaintiffs and defendants both presented statistics at trial that they urged support their respective positions. Although the defense purported to show the curfew was successfully reducing crime, the defense expert admitted, “I would feel uncomfortable saying that the curfew directly decreases crime simply because I didn’t conduct an analysis, because data wasn’t available to me, and I don’t want to overstep the data that I had.” Plaintiffs’ expert, drawing on the Town of Vernon’s Police Department Quarterly Reports from July 1993 to June 1998, concluded that Vernon experienced a larger decline in crime before the curfew took effect than after, and that Vernon’s crime decline did not correspond to the curfew’s enforcement. Plaintiffs’ expert additionally examined the 410 reports of curfew stops involving youths ages 16 and 17 provided by Vernon police. Analyzing these reports by time of stop, age, sex, race, residence, other criminal activity and circumstances of the juvenile, they found negligible juvenile crime and no instances of juvenile endangerment in connection with the stops.
A problem for both plaintiffs’ and defendants’ expert analyses is that the available sample of curfew citations and warnings was much smaller than the total number actually stopped for violations because state law prohibited the release of the records of children under 16. In addition, the records of arrests maintained by the Vernon Police Department revealed the total number of juvenile arrests, but did not reveal the time when the crimes were committed. Further, the police did not compile statistics on the age of victims of crime in Vernon. Moreover, the curfew was enacted at the same time that several other law enforcement changes were made in Vernon, including the hiring of a new police chief, the addition of new police officers to the force and the implementation of new community programs. In light of all of these changes, the district court correctly declined to attribute any changes in arrests or crime levels to the curfew ordinance. See Ramos, 48 F.Supp.2d at 185.
Another factor leading to the ordinance’s adoption was the survey. In April and May 1994 the Town of Vernon surveyed students in middle school and high school to determine the strengths and weaknesses of existing services for youth and families in Vernon. The study of the Vernon schoolchildren reveals that many of them were concerned about guns and violence. Ramos, 48 F.Supp.2d at 185.
2. Proof Fails to Support Aims of Curfew
When reviewing a law under the lens of intermediate scrutiny, “the Equal Protection Clause requires more than the mere incantation of a proper state purpose.” Trimble v. Gordon, 430 U.S. 762, 769, 97 S.Ct. 1459, 52 L.Ed.2d 31 (1977). Accordingly, it is not enough for defendants to recite interests that have been used to support curfew ordinances in other municipalities. Instead, defendants must show that this ordinance, which restricts constitutional rights, is the product of “reasoned analysis.” Hogan, 458 U.S. at 725-26, 102 S.Ct. 3331.
*186Although the Town of Vernon’s curfew aims to reduce juvenile crime and victimization at night, defendants produced nothing to show that any consideration was given to the nocturnal aspect of the curfew ordinance. Wakefield’s personal observations, both pre- and post-curfew, were primarily at hours that were not covered by the curfew. Clemons also admitted that her evening observations of groupings of people on the streets occurred between the evening hours of 6:00 p.m. and 9:00 p.m., a period of time in which the curfew was not in effect. The defendants have the burden of proof under the intermediate scrutiny standard, and they failed to present any persuasive reason for the curfew hours chosen by the town. In fact, there is a disconnect between the proof of purportedly problem hours and the curfew hours set out in the ordinance.
Similarly, the curfew, by its terms, keeps the under-18 set off the streets at night, but no effort seems to have been made by the town to ensure that the population targeted by the ordinance represented that part of the population causing trouble or that was being victimized (or that was even in particular danger of being victimized). For all we know, gang members and intimidating idlers might have been mostly over 18 years old. After all, though Clemons thought that the loiterers she observed were teens, she guessed that they were anywhere from 15 to 19 years of age. Similarly, the age of victims and the vulnerable does not appear to have been specifically examined beyond the general assumption that children are more vulnerable than adults. Although assumptions about children may suffice to establish the significance of the government’s interests and may even sustain the validity of a legislative enactment under a lower level of scrutiny, assumptions will not carry the government’s burden of showing the presence of the “requisite direct, substantial relationship,” Hogan, 458 U.S. at 725, 102 S.Ct. 3331, between the factual premises that motivated the enactment of a curfew and its terms.
Likewise, the concerns expressed in the 1994 youth survey do not translate into a mandate for the instant curfew. Significantly, the survey results do not identify any hours as particularly dangerous, and they do not indicate that the schoolchildren themselves are the source of the problem or any more likely than adults to be victims. We see no direct connection therefore between the survey results and the curfew. Finally, the murder of the Vernon teen provides scant support for the existing Vernon curfew ordinance. That murder, as shocking as it was, occurred inside the victim’s home in the afternoon, not on the streets at night. The circumstances surrounding that particular crime (ie., time and place) are therefore outside the scope of Vernon’s curfew ordinance.
Because defendants have failed to demonstrate that Vernon’s curfew ordinance is substantially related to an important governmental interest, we hold that it is unconstitutional as applied. We do not intend by our holding to rule that the Equal Protection Clause prohibits the enactment of a juvenile curfew ordinance. Nor do we think communities need to bide their time waiting for unspeakable tragedies to befall them before responding with legislation. But equal protection demands that the municipality “carefully stud[y] the contours of the problem it [is] seeking to address and legislate[ ] in accordance with its findings.” Buzzetti v. City of N.Y., 140 F.3d 134, 142 (2d Cir.1998). Such careful study of the problem and the requisite findings are lacking in the case before us. In other words, there is a conspicuous lack of relationship between the contours of the problem identified by the Vernon Town Council and the curfew ordinance enacted in re*187sponse. Even one of defendants’ own expert witnesses acknowledged the randomness of the ordinance by stating that “[t]he adoption of the curfew itself probably could be considered a knee jerk reaction.” Consistent with this characterization, the expert report submitted by defendants suggests that “[t]he murder may have been the deciding factor for the town council to pass a juvenile curfew.” Moreover, were we to hold that the curfew’s validity could be measured by its success in accomplishing the goals that prompted its enactment in the first place, it does not pass that test, given the equivocal nature of the evidence. In addition to the equal protection claims here addressed, we have before us Janet Ramos’ assertion that her due process right to parent has been violated by the ordinance. Because in our view the curfew ordinance is unconstitutional under the Equal Protection Clause, we need not reach or rule upon plaintiffs’ other constitutional challenges.
CONCLUSION
The essence of the present case is not whether teenagers have a constitutional right to “idly hang[ ] out” as the dissent suggests, but whether a town may pass an ordinance arbitrarily targeting a minor’s right to mobility in what amounts to a “knee jerk reaction” to events insufficiently related or relevant to those burdened by the ordinance. The Town of Vernon’s curfew ordinance is unconstitutional because it infringes on the plaintiffs’ rights under equal protection. This is not to say that parents have the right to direct their children to violate legitimately enacted curfew ordinances, or that minors may flout a city’s legitimately passed safety rules in the name of the Constitution. Rather, if a municipality wishes to single out minors as a group to curtail a constitutional freedom, which the minors have absent parental prohibition, then the municipality must satisfy constitutional requirements by tying their policies to the special traits, vulnerabilities, and needs of minors. Here, we give the town’s actions intermediate scrutiny, and based on the particular set of facts reflected in the record before us, the town’s ordinance does not pass muster, insofar as it bars juveniles from being on the streets with parental consent during curfew hours.
Accordingly, we reverse and direct that judgment be entered declaring that the rights of plaintiffs have been violated by the Town of Vernon’s curfew ordinance and that an appropriate permanent injunction against the Town of Vernon be issued.
. The present case has never been explicitly characterized as either facial or as-applied. Rather, plaintiffs’ complaint without specificity alleges the ways the ordinance has infringed on their rights in their specific circumstances, and then asks for relief. While some of the claims plaintiffs raise are logically analyzed as facial challenges, e.g., the challenges for overbreadth and vagueness, the equal protection claim is more logically viewed “as-applied" given the statements in the complaint. Even if a facial challenge was intended, a facial challenge in the context of the present equal protection claim would logically include within it an as-applied challenge, and thus we cannot ignore the constitutional violation simply because the words "as-applied" were not used. Cf. City of Cleburne v. Cleburne Living Ctr., Inc., 473 U.S. 432, 478, 105 S.Ct. 3249, 87 L.Ed.2d 313 (1985) (Marshall, J., concurring in part and dissenting in part) (“The formal label under which an equal protection claim is reviewed is less important than careful identification of the interest at stake and the extent to which society recognizes the classification as an invidious one.”); see also Bornholdt v. Brady, 869 F.2d 57, 68 (2d Cir.1989) (entertaining argument where district court's attention was directed to facts pertinent to the argument and only legal questions were at issue).
. The dissent argues that rational basis review should instead be used in the instant case. For the reasons stated later, we disagree and emphasize that defendants themselves did not challenge on appeal the district court’s application of intermediate scrutiny.
. The right we evaluate in this case is narrower than an adult’s right to free movement, however. It is a minor's right to move about freely with parental consent.
. We note, however, that the Supreme Court has never considered the issue. Although courts typically assume that no age cohort is a suspect class, see, e.g., Hutchins, 188 F.3d at 536 n. 1 (plurality opinion), old age has been the burdened class in all of the cases in which the Supreme Court has concluded that age is not a suspect class, see Gregory v. Ashcroft, 501 U.S. 452, 470, 111 S.Ct. 2395, 115 L.Ed.2d 410 (1991); Mass. Bd. of Ret. v. Murgia, 427 U.S. 307, 313-14, 96 S.Ct. 2562, 49 L.Ed.2d 520 (1976).
. While “children are at all times in the custody of either their parents or the state,” see Flores, 507 U.S. at 302, 113 S.Ct. 1439, we cannot accept the assumption, implicit in the dissent’s argument, that the state’s right to control children is coextensive with that of fit parents. The dissent cites case law to support the assertion that the state may control children as it wishes, and therefore the ordinance here is only invalid if it interferes with parents' rights to control the upbringing of their children. The cases dissent cites, however, are ones arising in special circumstances, such. as that of children in a public school, where we accept that the state necessarily has a greater degree of control over children in its role in loco parentis, see Vernonia Sch. Dist. 47J, 515 U.S. at 654-55, 115 S.Ct. 2386, or juvenile aliens in state custody, see Flores, 507 U.S. at 302-04, 113 S.Ct. 1439. Apart from these and perhaps other specific contexts, the fact that children are in the "custody of the state” in some metaphysical sense does not mean that the state may arbitrarily exert physical control the way that parents can without adequate justification.