Opinion for the Court filed by Circuit Judge SILBERMAN.

Circuit Judges Wald, Ginsburg, Henderson, and Garland join in Parts I, III, & IV.

Opinion Concurring in part and Concurring in the result filed by Chief Judge Edwards, with whom Circuit Judges Wald and Garland join in Part II.

Opinion Concurring in part and Concurring in the result filed by Circuit Judges Wald and Garland.

Opinion Concurring in part and Dissenting in part filed by Circuit Judge Rogers, with whom Circuit Judge Tatel joins, and Circuit Judge Wald joins in Parts II and III, and Circuit Judge Garland joins in Part III.

Dissenting opinion filed by Circuit Judge Tatel.

The District of Columbia ap-peals the district court's grant of summary judgment to plaintiffs/appellees, a group of minors, parents, and a private business, enjoining enforcement of the District's Juvenile Curfew, and holding that it violates the fundamental rights of minors and their parents and is unconstitutionally vague. A divided panel of our circuit affirmed the district court, and rehearing en banc was granted. A plurality believes that the curfew implicates no fundamental rights of minors or their parents. Even assuming the curfew does implicate such rights, we hold that it survives heightened scrutiny. And, it does not violate the First or Fourth Amendment rights of minors.

The District of Columbia Council, determining that juvenile crime and victimization in the District was a serious problem-and growing worse-unanimously adopted the Juvenile Curfew Act of 1995, which bars juveniles 16 and under from being in a public place unaccompanied by a parent or without equivalent adult supervision from 11:00 p.m. on Sunday through Thursday to 6:00 a.m. on the following day and from midnight to 6:00 a.m. on Saturday and Sunday, subject to certain enumerated defenses. See D.C. Code Ann. §§ 6-2182, 6-2183 (1996). The curfew provides that a minor (defined as "any person under the age of 17 years," but not "a judicially emancipated minor or a married minor") cannot remain in a public place or on the premises of any establishment within the District of Columbia during curfew hours. A parent or guardian commits an offense by knowingly permitting, or through insufficient control allowing, the minor to violate the curfew. Owners, operators, or employees of public establishments also violate the curfew by knowingly allowing the minor to remain on the premises, unless the minor has refused to leave and the owner or operator has so notified the police. The curfew contains eight "defenses": it is not violated if the minor is (1) accompanied by the minor's parent or guardian or any other person 21 years or older authorized by a parent to be a caretaker for the minor; (2) on an errand at the direction of the minor's parent, guardian, or caretaker, without any detour or stop; (3) in a vehicle involved in interstate travel; (4) engaged in certain employment activity, or going to or from employment, without any detour or stop; (5) involved in an emergency; (6) on the sidewalk that abuts the minor's or the next-door neighbor's residence, if the neighbor has not complained to the police; (7) in attendance at an official school, religious, or other recreational activity sponsored by the District of Columbia, a civic organization, or another similar entity that takes responsibility for the minor, or going to or from, without any detour or stop, such an activity supervised by adults; or (8) exercising First Amendment rights, including free exercise of religion, freedom of speech, and the right of assembly. If, after questioning an apparent offender to determine his age and reason for being in a public place, a police officer reasonably believes that an offense has occurred under the curfew law and that no defense exists, the minor will be detained by the police and then released into the custody of the minor's parent, guardian, or an adult acting in loco parentis. If no one claims responsibility for the minor, the minor may be taken either to his residence or placed into the custody of the Family Services Administration until 6:00 a.m. the following morning. Minors found in violation of the curfew may be ordered to perform up to 25 hours of community service for each violation, while parents violating the curfew may be fined up to $500 or required to perform community service, and may be required to attend parenting classes.

Appellees sued the District of Columbia seeking an injunction against enforcement of the curfew and a declaration that the curfew violates the minors' Fifth Amendment Due Process and Equal Protection rights to freedom of movement; violates the parents' Fifth Amendment due process rights to raise their children; violates the minors' First Amendment rights to freedom of expression and assembly; violates the minors' Fourth Amendment right to be free from unreasonable searches and seizures; and is unconstitutionally vague. The district court granted summary judgment to appellees and enjoined enforcement of the curfew. Hutchins v. District of Columbia, 942 F. Supp. 665, 668 (D.D.C. 1996). The court concluded that "it is a well-settled legal principle that the right to free movement is a fundamental right generally," and although the "[s]tate has a great interest in regulating the activities of, and providing protection for, minors," this "interest does not automatically dilute the constitutional rights of [ ] minors." Id. at 671. Thus, minors who are not in the custody of the state have a fundamental right to free movement. Since the curfew intrudes on minors' right to free movement, as well as on the parents' fundamental rights to direct their children's upbringing, it must be subjected to strict scrutiny. Accordingly, the law must be narrowly tailored to promote the District's asserted compelling interests in protecting the welfare of minors by reducing the likelihood that minors will perpetrate or become victims of crime, and by promoting parental responsibility by assisting parents in exercising reasonable supervision of minors entrusted to their care. The district court found that the statistical data produced by the District did not meet that test. The court also thought that four of the curfew's defenses-the First Amendment defense, the emergency defense, the responsible entity defense, and the sidewalk defense-were "woefully vague" and did not withstand constitutional scrutiny. Appellees' First and Fourth Amendment claims were not reached.

Appellees contend (and the district court determined) that the curfew infringes on a substantive fundamental right-the right to free movement-and as a substantive right it cannot be taken away merely through "due process."1 Of course a right to free movement is a synonym for the right to liberty; when one is put in jail it is obvious that one's right to free movement has been curtailed, but that is constitutionally permissible if the person whose liberty has been curtailed is afforded due process. But any government impingement on a substantive fundamental right to free movement would be measured under a strict scrutiny standard and would be justified only if the infringement is narrowly tailored to serve a compelling state interest. See Reno v. Flores, 507 U.S. 292, 301-02 (1993) (citing Collins v. Harker Heights, 503 U.S. 115, 125 (1992)). But does such a substantive right exist?

Although appellees cite numerous cases in support of the proposition that "the right to free movement is as old as the Republic," the cases do not support such a sweeping assertion. It is true that the right to interstate travel is wellestablished. See Saenz v. Roe, 119 S. Ct. 1518, 143 L. Ed. 2d 689 (May 17, 1999); Shapiro v. Thompson, 394 U.S. 618, 629-31 (1969). Although the precise source of this right remains somewhat obscure, see Shapiro, 394 U.S. at 629 n.8, its origins reflect a concern over state discrimination against outsiders rather than concerns over the general ability to move about. See Saenz v. Roe, 119 S. Ct. 1518, 143 L. Ed. 2d 689 (grounding at least one component of the right to interstate travel in the Privileges and Immunities Clause of the Fourteenth Amendment); United States v. Guest, 383 U.S. 745, 758 (1966) (describing the right to interstate travel as originating in the Articles of Confederation and as being a "necessary concomitant of the stronger Union the Constitution created"); Zobel v. Williams, 457 U.S. 55, 79-81 (1982) (O'Connor, J., Concurring in the judgment) (describing the right as originating in the Privileges and Immunities Clause of Art. IV); Edwards v. California, 314 U.S. 160, 173-74 (1941) (describing the right as being grounded in the Commerce Clause); Zobel, 457 U.S. at 60 n.6 (describing the right to travel cases as a particular application of equal protection analysis); Shapiro, 394 U.S. at 630 (describing the right as deriving from general principles of federalism, since the right to travel from state to state " 'occupies a position fundamental to the concept of our Federal Union' " (quoting Guest, 383 U.S. at 757-58)).

The Court has suggested on occasion that some more generalized right to movement may exist. See, e.g., Kent v. Dulles, 357 U.S. 116, 126 (1958) ("Freedom of movement is basic in our scheme of values."); Guest, 383 U.S. at 758 (proclaiming that citizens of the United States "must have the right to pass and repass through every part of [the country] without interruption, as freely as in [their] own states" (quoting Crandall v. Nevada, 73 U.S. (6 Wall.) 35, 49 (1867) (quoting The Passenger Cases, 48 U.S. (7 How.) 283, 492 (1849) (Taney, C.J., Dissenting)))); Williams v. Fears, 179 U.S. 270, 273 (1900) (indicating that the "right of locomotion," like the "right to contract," is protected by substantive due process). But those comments are only dicta-the cases involved travel across borders, not mere "locomotion."2 Indeed, the Supreme Court in Memorial Hospital v. Maricopa County, 415 U.S. 250, 255 (1974), cast strong doubt on the idea that there was a fundamental right to free movement, noting that "[e]ven a bona fide residence requirement would burden the right to travel if travel meant merely movement." In any event, the Court subsequently made clear that any right to travel involved in Kent and Aptheker was distinct from the recognized right to interstate travel, explaining that international travel is no more than an aspect of liberty that is subject to reasonable government regulation within the bounds of due process, whereas interstate travel is a fundamental right subject to a more exacting standard. See Haig v. Agee, 453 U.S. 280, 306-07 (1981) (upholding constitutionality of regulation authorizing the revocation of passport on the ground that the regulation authorized revocation only where the holder's activities in foreign countries are causing or are likely to cause serious damage to national security). Since the right to free movement would cover both interstate and international travel, Agee at least implies that the right recognized by the Court is decidedly more narrow.

Nor do the vagrancy cases relied on by appellees support their claim. While Justice Douglas noted in Papachristou v. City of Jacksonville, 405 U.S. 156 (1972), that "wandering or strolling" from place to place was historically part of the "amenities of life," id. at 164, the Court actually held only that the vagrancy law at issue was void for vagueness, see id. at 165-71; see also Kolender v. Lawson, 461 U.S. 352, 357-62 (1983). While vagrancy statutes certainly prohibit individuals from moving about, the constitutional infirmity in these statutes is not that they infringe on a fundamental right to free movement, but that they fail to give fair notice of conduct that is forbidden and pose a danger of arbitrary enforcement. In other words, they do not afford procedural due process.

The Supreme Court in Maricopa County specifically declined to decide whether the right to interstate travel recognized in Shapiro has its analogue in intrastate travel. The circuits are split on this question. Compare King v. New Rochelle Mun. Hous. Auth., 442 F.2d 646, 647-48 (2d Cir. 1971) (holding that a municipal resolution imposing a fiveyear residency requirement for admission to public housing burdened fundamental right to intrastate travel and stating that it would be "meaningless to distinguish between interstate and intrastate" travel) with Wardwell v. Board of Educ. of Cincinnati, 529 F.2d 625, 627-28 (6th Cir. 1976) (rejecting a fundamental right to intrastate as opposed to interstate travel) and Wright v. City of Jackson, 506 F.2d 900, 902-03 (5th Cir. 1975) (same). More pertinent to the case at hand, one circuit has recognized that traffic restrictions (although they have been easily sustained) at least implicate a substantive right of free movement. See Lutz v. City of York, 899 F.2d 255, 268 (3d Cir. 1990) (holding that ordinance outlawing "cruising," which consisted of driving repeatedly around loop of public roads, implicated substantive due process right to "move freely about one's neighborhood or town," but upholding ordinance under intermediate scrutiny test derived from First Amendment time, place, and manner doctrine); see also Townes v. City of St. Louis, 949 F. Supp. 731 (E.D. Mo. 1996) (assuming heightened scrutiny applied when resident claimed that city's placement of large flower pots across the entrance to her block infringed her fundamental right to localized travel but holding the ordinance would survive intermediate scrutiny), aff'd, 112 F.3d 514 (8th Cir. 1997). Appellees argue that restrictions of that kind, even ordinary traffic lights, impinge on this substantive free movement right. We are rather doubtful that substantive due process, those constitutional rights that stem from basic notions of ordered liberty "deeply rooted in [our] history and tradition," Washington v. Glucksberg, 117 S. Ct. 2258, 2268 (1997) (quoting Moore v. City of East Cleveland, 431 U.S. 494, 503 (1977)), can be so lightly extended. On the other hand, we recognize that a hypothetical municipal restriction on the movement of its citizens, for example, a draconian curfew, might bring into play the concept of substantive due process.

Be that as it may, there is an important caveat to bear in mind when considering potential extensions of substantive due process, which "has at times been a treacherous field," Michael H. v. Gerald D., 491 U.S. 110, 122 (1988) (plurality) (quoting Moore, 431 U.S. at 502). The Supreme Court has warned us that our analysis must begin with a careful description of the asserted right for the more general is the right's description, i.e., the free movement of people, the easier is the extension of substantive due process. See Reno v. Flores, 507 U.S. at 302; see also Michael H., 491 U.S. at 127 n.6 (proper level of generality at which to describe the right is "the most specific level at which a relevant tradition protecting, or denying protection to, the asserted right can be identified") (opinion of Scalia, J., joined by Rehnquist, C.J.). And the "doctrine of judicial self-restraint requires us to exercise the utmost care whenever we are asked to break new ground in this field." Reno v. Flores, 507 U.S. at 302 (quoting Collins v. Harker Heights, 503 U.S. 115, 125 (1992)). For that reason we must ask not whether Americans enjoy a general right of free movement, but rather whatever are the scope and dimensions of such a right (if it exists), do minors have such a substantive right? Do they have the right to freely wander the streets-even at night? See id. (defining the asserted right, not as freedom from physical restraint, but as "the alleged right of a child who has no available parent, close relative, or legal guardian, and for whom the government is responsible, to be placed in the custody of a willing and able private custodian rather than a governmentoperated or government-selected child care institution").

We think that juveniles do not have a fundamental right to be on the streets at night without adult supervision. The Supreme Court has already rejected the idea that juveniles have a right to "come and go at will" because "juveniles, unlike adults, are always in some form of custody," id. (quoting Schall v. Martin, 467 U.S. 253, 265 (1984)), and we see no reason why the asserted right here would fare any better. That the rights of juveniles are not necessarily coextensive with those of adults is undisputed, and "unemancipated minors lack some of the most fundamental rights of self-determination-including even the right of liberty in its narrow sense, i.e., the right to come and go at will." Vernonia Sch. Dist. 47J v. Acton, 515 U.S. 646 (1995). While appellees claim that this reasoning obscures the difference between parental custody and governmental custody, appellees necessarily concede that juveniles are always in some form of custody. Not only is it anomalous to say that juveniles have a right to be unsupervised when they are always in some form of custody, but the recognition of such a right would fly in the face of the state's well-established powers of parens patriae in preserving and promoting the welfare of children. The state's authority over children's activities is unquestionably broader than that over like actions of adults. See Prince v. Massachusetts, 321 U.S. 158, 169 (1944) (observing that the state's power to prohibit street preaching by "children not accompanied by an older person hardly seems open to question"). And it would be inconsistent to find a fundamental right here, when the Court has concluded that the state may intrude upon the "freedom" of juveniles in a variety of similar circumstances without implicating fundamental rights, see id., 321 U.S. at 166-67, 168-69 (citing compulsory school attendance and child labor laws), and can do so in far more intrusive ways than is contemplated here, see, e.g., Flores, 507 U.S. at 301-03 (upholding on rational basis review detention of deportable juveniles for release generally only to their parents, close relatives, or legal guardians); Schall, 467 U.S. at 263-64 (upholding pretrial detention of juvenile delinquents after a finding of "serious risk" on the ground that it served a legitimate, nonpunitive regulatory purpose); Prince, 321 U.S. at 169-70 (upholding law prohibiting children from selling magazines on the street, even when accompanied by parent or guardian, against claim that the law violated child's freedom of religion); Ginsberg v. New York, 390 U.S. 629, 637-643 (1968) (upholding on rational basis review a ban on sale of material to minors that would not be considered "obscene" for adults).

Neither does the asserted right here have deep roots in our "history and tradition." As the District noted, juvenile curfews were not uncommon early in our history, see Note, Curfew Ordinances and the Control of Nocturnal Juvenile Crime, 107 U. Pa. L. Rev. 66, 66-69 n.5 (1958), nor are they uncommon now, see Debra Livingston, Police Discretion and the Quality of Life in Public Places: Courts, Communities, and the New Policing, 97 Colum. L. Rev. 551, 555 & n.11 (1997) (discussing research demonstrating that the use of curfews to control delinquency and reduce juvenile victimization is the norm in major American cities) (citing William Ruefle & Kenneth M. Reynolds, Curfews and Delinquency in Major American Cities, 41 Crime & Delinq. 347, 353 (1995)). That juvenile curfews are common is, of course, not conclusive in determining whether they comport with due process, but the historical prevalence of such laws is "plainly worth considering" in determining whether the practice " 'offends some principle of Justice so deeply rooted in the traditions and conscience of our people as to be ranked as fundamental.' " Schall, 467 U.S. at 268 (quoting Snyder v. Massachusetts, 291 U.S. 97, 105 (1934)). In sum, neither history nor precedent supports the existence of a fundamental right for juveniles to be in a public place without adult supervision during curfew hours, and we decline to recognize one here.3

Even if juveniles themselves lack a fundamental right of movement, appellees claim that parents have a fundamental, substantive due process right to direct and control their children's upbringing and that such a right is abridged by the curfew. Whether children under the age of 17 are to be free to be abroad at night is presumptively a matter for their parents to determine, as part and parcel of that upbringing. (Appellees suggest that this concept extends to permitting a child of any age-even four-to be on the street in the middle of the night.) This parental fundamental right alone, it is argued, obliges us to Judge the D.C. curfew by heightened scrutiny. We disagree, not because we think that no such fundamental right exists in any dimension, but rather because we think it not implicated by the curfew.

In the early twenties, the Supreme Court held unconstitutional a state statute that prohibited the teaching of subjects in foreign languages and the teaching of foreign languages to children before the eighth grade (even in a private school), see Meyer v. Nebraska, 262 U.S. 390 (1923), and a statute that required children 8 to 16 to attend a public school, see Pierce v. Society of Sisters, 268 U.S. 510 (1925). Although these cases could be thought to rest on the Court's perception that the statutes had an irrational basis, see Meyer, 262 U.S. at 628 (concluding that the statute as applied was "arbitrary and without reasonable relation to any end within the competency of the state"), in Pierce the Court did observe that "[t]he 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," id. at 533. And by 1944 in Prince, the Court said that "[i]t is cardinal with us that the custody, care and nurture of the child reside first in the parents, whose primary function and freedom include preparation for obligations the state can neither supply nor hinder." Prince, 321 U.S. at 166 (citing Pierce and Meyer) (emphasis added). Although the Court in Prince held that the state could ban children from selling magazines on the street, even when accompanied by a parent and despite the religious nature of the publications, it did so after balancing the state's interest against the parents' rights. See id. at 165-70. That approach might suggest a more searching inquiry than rational basis review. (This was long prior to the doctrinal development of the formal tests that are now part of modern substantive due process, and, therefore, the Court did not speak in terms of strict scrutiny or rational basis.) But the Court emphasized that the state's interest in guarding the welfare of children-even against the wishes of a parent-was particularly powerful to ward off the "evils ... [of] public places" and the "possible harms arising from other activities subject to all the diverse influences of the street." Id. at 168. By so reasoning, the Court distinguished between the "private realm of family life," id. at 166, and those activities subject to the evils of public places, applying something very close to rational basis review for laws restricting the latter. See also Wisconsin v. Yoder, 406 U.S. 205, 215, 231 (1972) (high school attendance law unconstitutionally infringed on parents' rights to direct the religious upbringing and education of their children; only those interests of the "highest order" can overcome those parental rights).

We glean from these cases, then, that insofar as a parent can be thought to have a fundamental right, as against the state,in the upbringing of his or her children, that right is focused on the parents' control of the home and the parents' interest in controlling, if he or she wishes, the formal education of children. It does not extend to a parent's right to unilaterally determine when and if children will be on the streets-certainly at night. That is not among the "intimate family decisions" encompassed by such a right. Schleifer v. City of Charlottesville, 159 F.3d 843, 853 (4th Cir. 1998), cert. denied, 119 S. Ct. 1252 (1999).

Even if the curfew implicated fundamental rights of children or their parents, it would survive heightened scrutiny. Assuming such rights are implicated, we must first decide whether, as the district court held, strict scrutiny applies or whether, as Judge Rogers concluded, see Hutchins v. District of Columbia, 144 F.3d 798, 809 (D.C. Cir. 1998), vacated and reh'g en banc granted, 156 F.3d 1267 (D.C. Cir. 1998), intermediate scrutiny is called for. We think the latter. Considering children's rights first, we agree that constitutional rights do not instantaneously appear only when juveniles reach the age of majority. See Planned Parenthood v. Danforth, 428 U.S. 52, 74 (1976). Still, children's rights are not coextensive with those of adults. See Prince, 321 U.S. at 169; see also Bellotti v. Baird, 443 U.S. 622, 633-39 (1979) (plurality opinion). So "although children generally are protected by the same constitutional guarantees ... as are adults, the State is entitled to adjust its legal system to account for children's vulnerability" by exercising broader authority over their activities. Bellotti, 443 U.S. at 635. This means, at minimum, that a lesser degree of scrutiny is appropriate when evaluating restrictions on minors' activities where their unique vulnerability, immaturity, and need for parental guidance warrant increased state oversight. See Carey v. Population Servs. Int'l, 431 U.S. 678, 693 n.15 (1977) (plurality opinion); Bellotti, 443 U.S. at 634. The reasoning of Bellotti, Prince, and Carey necessarily suggests that something less than strict scrutiny-intermediate scrutiny-would be appropriate here. Not only can juveniles be thought to be more vulnerable to harm during curfew hours than adults, but they are less able to make mature decisions in the face of peer pressure, and are more in need of parental supervision during curfew hours. See Schleifer, 159 F.3d at 847 (applying intermediate scrutiny, reasoning that the "qualified rights" of juveniles should be subject to something more than rational basis and something less than strict scrutiny review). Compare Nunez v. City of San Diego, 114 F.3d 935, 946 (9th Cir. 1997) (rejecting lesser degree of scrutiny for equal protection challenge to juvenile curfew but noting that strict scrutiny in the context of minors "may allow greater burdens on minors than would be permissible on adults").

To withstand intermediate scrutiny, the curfew must be "substantially related" (rather than narrowly tailored) to the achievement of "important" (rather than compelling) government interests.4 See Craig v. Boren, 429 U.S. 190, 197 (1976); see also Mississippi Univ. for Women v. Hogan, 458 U.S. 718, 724 (1982). The asserted government interest here is to protect the welfare of minors by reducing the likelihood that minors will perpetrate or become victims of crime and by promoting parental responsibility. The District presented reams of evidence depicting the devastating impact of juvenile crime and victimization in the District-the juvenile violent crime arrest rate for juveniles ages 10 to 17 was higher than that in any state and was more than three times the national average, see Kids Count Data Book: State Profiles of Child Well-Being (Annie E. Casey Foundation, Baltimore, Md.) 1995, the District had the highest violent death rate for teens ages 15 to 19, which was four times the national average, and the District was ranked dead last, almost three times worse than the worst state, in children's overall well-being. See id. This was the abysmal situation confronting the District when it voted to adopt the curfew law. Statistics showed the situation worsening. See Office of Corporation Counsel Juvenile Section Statistical Report By Priority Charge, Fiscal Years 1987-1995 (showing dramatic increase in juvenile arrests for, inter alia, aggravated assault, murder, and carrying a dangerous weapon). Given this picture of juvenile crime and victimization, there can be no serious dispute that protecting the welfare of minors by reducing juvenile crime and victimization is an important government interest. See Turner Broadcasting System, Inc. v. FCC, 512 U.S. 622, 669 (1994) (government must demonstrate that its asserted interests are real and not merely conjectural).

Whether the curfew is "substantially related" to the achievement of that interest is the more difficult question here. Neither the Supreme Court nor the lower federal courts has expounded upon-explained in doctrinal termsthe phrase "substantial relationship." That test obviously calls for a more searching inquiry than rational basis (the minimum standard for judging equal protection claims), yet a more deferential one than strict scrutiny's narrow tailoring component. In judging the closeness of the relationship between the means chosen (the curfew), and the government's interest, we see three interrelated concepts: the factual premises upon which the legislature based its decision, the logical connection the remedy has to those premises, and the scope of the remedy employed.

The plaintiffs in this case criticize the District's legislative decision on all three grounds. Thus, appellees argue: that the District improperly relied on statistical evidence from other cities showing the effectiveness of similar curfew laws in reducing juvenile crime and victimization because the other cities are not sufficiently comparable; that testimony as to the effectiveness of the curfew in the District itself (during the first three months) was unreliable; that the District's juvenile arrest statistics (the most fundamental factual premise for the need for a curfew) were flawed because they included 17 year olds not covered by the curfew; that the District's statistics did not adequately establish that the District's problem centered on juvenile crime and victimization during curfew hours; and that the District did not produce data showing that crimes committed by and against juveniles occurred in "public," i.e., outside of the home where juveniles will presumably be during curfew hours.

Of course, in considering the District Council's decision, we must bear in mind that we are not reviewing a district court's or an agency's findings of historical fact which is a more structured kind of decision than a legislative judgment. And even in the context of review of agency rulemaking, we are obliged to give great leeway to predictive judgments based on a matter within the agency's sphere of expertise. See Fresno Mobile Radio, Inc. v. FCC, 165 F.3d 965, 971 (D.C. Cir. 1999). To be sure, in two cases applying intermediate scrutiny in the context of quasi-suspect classes, the Supreme Court closely and skeptically examined statistical social science data purporting to justify differential treatment of men and women. See Craig, 429 U.S. at 199-204; Hogan, 458 U.S. at 23-31. But we think the key to understanding the Supreme Court's close analysis in those cases is the Court's observation in Craig "that proving broad sociological propositions by statistics is a dubious business, and one that inevitably is in tension with the normative philosophy that underlies the Equal Protection Clause." Craig, 429 U.S. at 203. We think by that the Court implied that it is particularly troubling when legislation provides for differential treatment between suspect (or quasi-suspect) classes and others. There is really no dispute in this case comparable to the hotly contested and sensitive question as to the differences between men and women. Plaintiffs do not dispute that the difference between adults and minors generally justifies a government's differential treatment of minors; they dispute only this particular differential treatment because of its interference with their "fundamental" right to free movement.

Bearing in mind, then, that we are reviewing a legislative decision, we turn to appellees' specific objections to the District's decisionmaking. Taking first the District's diagnosis of its own situation, we ask whether it was impermissible for the Council to rely on arrest statistics that included 17 year olds and victimization statistics that covered 15 to 19 year olds. Appellees claim that including 17 year olds' arrests will necessarily overstate the magnitude of juvenile crime-at least as the District has defined juveniles. But the District brought to our attention more data showing that arrests for youths under 17 have been increasing steadily.

In any event, the District is not obliged to prove a precise fit between the nature of the problem and the legislative remedy-just a substantial relation. The District can hardly be faulted for determining not to include 17 year olds in the curfew; obviously that would be more intrusive and create more of an enforcement problem. And even if minors under 17 are less likely to commit crimes than 17 year olds, common sense tells us that younger children will surely be more vulnerable.

Appellees also claim that the District's data is flawed because it failed to establish that the District had a problem with juvenile crime and victimization during curfew hours. The material presented to the Council on this point consisted of a chart prepared by the Metropolitan Police Department which showed that most juvenile arrests took place during curfew hours. Echoing the district court, appellees argue that this evidence is "woefully deficient," Hutchins, 942 F. Supp. 2d at 677, because the source data, from which the chart was compiled, appears to conflict with the chart. While the data is admittedly less than crystal clear, any discrepancies appear to be minor.5 The bottom line is that the District's statistics indicate that more than 50% of juvenile arrests took place during curfew hours. The Fifth Circuit, in evaluating an almost identical curfew, concluded that the curfew would pass even strict scrutiny, notwithstanding that "the city was unable to provide precise data concerning the number of juveniles who commit crimes during curfew hours, or the number of juvenile victims of crimes committed during the curfew." Qutb v. Strauss, 11 F.3d 488, 493 (5th Cir. 1993). That serious crimes such as murder, rape, and aggravated assault, committed by groups of all ages, were more likely to occur during curfew hours was sufficient to demonstrate a "fit" between the curfew ordinance and the compelling state interest. See id. Similarly, that the District did not produce data showing where juvenile crime and victimization occurred (i.e., that it occurred primarily outside of the home) is not problematic. That a substantial percentage of violent juvenile victimizations (approximately 33%) occurred on the streets adequately supports the relationship between the government's interest and the imposition of the curfew.

Nevertheless, appellees argue that the District was obliged to confine the curfew to high-crime areas of the city. We flatly disagree. To have done so would have opened the Council to charges of racial discrimination. Indeed, it would have faced attacks on that decision similar to those directed to the "broad sociological propositions" the Supreme Court disapproved of in Craig.

Appellees' claim that the District was not entitled to rely on curfew experiences in other cities strikes us as particularly weak. Of course no city is exactly comparable to any other, but it would be folly for any city not to look at experiences of other cities. And in drawing Conclusions from those experiences, legislatures are not obliged to insist on scientific methodology. See City of Renton v. Playtime Theatres, Inc., 475 U.S. 41, 51-52 (1986) (holding that under intermediate scrutiny in the First Amendment context, a city may rely on evidence generated by other cities "so long as whatever evidence the city relies upon is reasonably believed to be relevant to the problem that the city addresses"); see also Craig, 420 U.S. at 201-04 (noting that state had relied on statistical evidence from other jurisdictions and, although criticizing state's proof on many grounds, not disapproving of such evidence per se). The Fourth Circuit in Schleifer noted that Charlottesville, in adopting its own juvenile curfew, had relied on a showing that Lexington, Kentucky had a successful juvenile curfew. Although the court there recognized that there was testimony that curfews may be more effective in smaller cities (suggesting that Lexington and Charlottesville may have similar experiences), the court also emphasized that the judgment about the potential efficacy of a curfew "is a political debate, not a judicial one." Schleifer, 159 F.3d at 850. In any event, the District had its own indications that the curfew was effective in the District of Columbia-the Deputy Chief of the Metropolitan Police Department testified before the D.C. Council that in its first three months the curfew had resulted in fewer juveniles on the streets during curfew hours, and thus a "reduction of the number of juvenile late night arrests," noting a 34% decrease in arrests of juveniles under 17 years old. Appellees question the relevance of this testimony because the District did not demonstrate that this drop in juvenile arrests was attributable to the curfew as opposed to some other factor. We think that objection calls for an absurd preciseness in legislative decisionmaking which would make it virtually impossible for any city to adopt any curfew.

Finally, we note that the eight defenses to the curfew strengthen the relationship between the curfew and its goal of reducing juvenile crime and victimization by narrowing the scope of the curfew.6 That is, the defenses (the constitutionality of which we take up below) help ensure that the ordinance does not sweep all of a minor's activities into its ambit but instead focuses on those nocturnal activities most likely to result in crime or victimization.

Assuming, as we do in this section of the opinion, that the fundamental rights of parents are implicated by curfews,7 we also conclude that this curfew passes intermediate scrutiny because it is carefully fashioned much more to enhance parental authority than to challenge it. If the parents' interests were in conflict with the state's interests, we would be faced with a more difficult balancing of sharply competing claims. See generally Bellotti, 443 U.S. at 637-39 & n.18 (noting that limitations on children's rights can be justified by the state's attempt to support parental authority). Thus, in Ginsberg, the Supreme Court observed that a ban on selling magazines to minors-magazines that would not be Judged constitutionally obscene if sold to adults-did not substantially conflict with parental authority because a parent could always buy those sorts of magazines for their children. See Ginsberg, 390 U.S. at 639. It could be said in that case that the ban nevertheless interfered with a parent's desire to allow his or her children independence to purchase magazines without parental supervision, but the Court did not consider that theoretical impingement on parental authority worth mentioning; it saw the statute as essentially supporting parental authority. The same dynamic is true here. The curfew's defenses allow the parents almost total discretion over their children's activities during curfew hours. There are no restrictions whatsoever on a juvenile's activities if the juvenile is accompanied by a parent, guardian, or an adult over the age of 21 authorized by the parent to supervise the juvenile. See D.C. Code § 6-2183(b)(1)(A); id. at § 6-2182(8). Parents can allow their children to run errands, which gives the parents great flexibility in exercising their authority. Contrary to appellees' view, we do not see how the curfew would preclude parents from allowing their children to walk the dog or go to the store. Id. at § 6-2183(b)(1)(B). Juveniles may attend any "official school, religious, or other recreational activity sponsored by the District of Columbia, a civic organization, or another similar entity that takes responsibility for the minor" as well as to travel to and from such activities. Id. at § 6-2183(b)(1)(G). Although the extent to which this "civic organization" defense would cover events at the Kennedy Center, lectures at the Smithsonian, church group activities, athletic events, early morning sports practice, high school band practice, and the like, can wait for the test of concrete cases raising those questions, the defense certainly gives parents a good deal of discretion over their children's activities. Together with the defenses provided for employment and emergencies, see id. at §§ 6-2183(b)(1)(D)-(E), parents retain ample authority to exercise parental control. Since the curfew generously accommodates parental rights, preserving parental discretion to direct the upbringing of their children, it does not unconstitutionally infringe on such rights. See Schleifer, 159 F.3d at 853 (concluding that parents' fundamental rights were not implicated by curfew, and then stating that exceptions to the curfew would accommodate the rights of parents); Qutb, 11 F.3d at 496 (same); Bykofsky v. Borough of Middletown, 401 F. Supp. 1242, 1264 (M.D. Pa. 1975) (same), aff'd, 535 F.2d 1245 (3d Cir.), cert. denied, 429 U.S. 964 (1976); compare Nunez, 114 F.3d at 946 (striking down curfew as violation of parental rights based on broad sweep of ordinance and limited exceptions). We think under applicable precedent the curfew facilitates rather than usurps parental authority.

Appellees' remaining attacks on the curfew fall away. They contend that the district court correctly concluded that four of the curfew's defenses-the First Amendment activity defense, the responsible entity defense, the sidewalk defense, and the emergency defense-are "woefully vague and undefined," and that these defenses therefore do not withstand constitutional scrutiny. Hutchins, 942 F. Supp. at 679. Insofar as appellees contend that there is too much imprecision in the articulation of these defenses, they are really undermining their claim that parental rights are impinged upon. For the very flexibility that the administration of the curfew contemplates enhances parental control.8 In any event, as the District noted, the Constitution does not require "unattainable feats of statutory clarity." United States v. Maude, 481 F.2d 1062, 1068 (D.C. Cir. 1973). Rather, a statutory provision is sufficiently definite to satisfy due process requirements so long as a person of ordinary intelligence would have a reasonable opportunity to know what is prohibited. See Grayned v. City of Rockford, 408 U.S. 104, 108 (1972); Connally v. General Constr. Co., 269 U.S. 385, 391 (1926). That "the fertile legal 'imagination can conjure up hypothetical cases in which the meaning' " of disputed terms could be questioned does not render the provision unconstitutionally vague. Terry v. Reno, 101 F.3d 1412, 1421 (D.C. Cir. 1996) (quoting Grayned, 408 U.S. at 110 n.15 (quoting American Communications Ass'n v. Douds, 339 U.S. 382, 412 (1950))).

Appellees claim that the First Amendment defense9 is impermissibly vague because juveniles would need to be "constitutional scholars" to know what activities were forbidden and that police officers untrained in the intricacies of the First Amendment will, in their unguided discretion, enforce the curfew unconstitutionally. But the defense simply ensures that the curfew will not be applied to protected expression; it is no more vague than the First Amendment itself. As the Fourth Circuit noted in upholding a nearly identical exception against a vagueness challenge, it is perfectly clear that some activities, such as religious worship and political protests, would be protected under the defense, and that other activities, such as rollerblading, would not. See Schleifer, 159 F.3d at 854. That there may be marginal cases between these two poles can be addressed as they arise, but such cases do not render the provision void for vagueness.10

The responsible entity defense,11 according to the appellees, is impermissibly vague because it does not define the term "a civic organization, or another similar entity that takes responsibility for the minor." While "civic organization" and "entity that takes responsibility for the minor" are admittedly imprecise terms, any ambiguity is not of constitutional magnitude. As the District points out, the defense by its own terms applies to activities sponsored by schools, religious organizations, or the District of Columbia. In this context, the addition of "civic organization, or another similar entity" simply includes within the defense the general class of organizations that may be thought analogous to schools, religious organizations, or governmental entities. Compare Hynes v. Mayor of Oradell, 425 U.S. 610, 621 (1976) (finding the term "civic" vague when striking down ordinances that required permits for door-to-door solicitation but that exempted civic organizations) with Schleifer, 159 F.3d at 854 (noting that Hynes does not stand for the broad proposition that "civic" is per se vague, and noting that the ordinary meaning of the term as used in the curfew law was not vague).

Appellees contend that the sidewalk defense12 is unconstitutionally vague because it "improperly delegates standardless discretion to neighbors." This argument is also without merit. The defense provides clear parameters as to what conduct is prohibited. It is irrelevant, for purposes of evaluating vagueness, that a neighbor has the "discretion" to call the police if a juvenile remains on the neighbor's sidewalk during curfew hours-the discretion exercised in this situation is analogous to that exercised by property owners under trespass laws.

Appellees also challenge the "emergency" defense,13 despite the detailed definition of emergency provided in the statute. It is argued that "emergency" is unconstitutionally vague because it is unclear whether the need to walk the dog or to go buy typing paper the night before a homework assignment is due constitutes an emergency under the curfew law. Again, this argument borders on the frivolous. Mere "speculative musings" about the possible meaning of a term do not render it unconstitutionally vague; to do so would make the drafting of laws an impossible task. Schleifer, 159 F.3d at 854.

Appellees argued before the district court that the curfew also violated their First and Fourth Amendment rights, but because the district court found the curfew unconstitutional on equal protection and due process grounds, it did not reach these additional constitutional claims. We exercise our discretion to resolve these purely legal claims in the interest of judicial economy. See Committee of 100 on the Federal City v. Hodel, 777 F.2d 711, 718-19 (D.C. Cir. 1985).

The curfew "possesses the potential to suppress First Amendment rights," according to appellees, and this defect is not cured by the curfew's defense for First Amendment activities. This argument is self-defeating because we cannot hold a statute facially unconstitutional (appellees' challenge is a facial one) based on a mere possibility that the statute might be unconstitutional in particular applications. See City Council of Los Angeles v. Taxpayers for Vincent, 466 U.S. 789, 797 (1984).14 In any event, the curfew does not itself regulate or proscribe expression, and thus would only be subject to scrutiny under the First Amendment if it regulated "conduct that has an expressive element," or if it "impose[d] a disproportionate burden upon those engaged in protected First Amendment activity." Arcara v. Cloud Books, Inc., 478 U.S. 697, 703-04 (1986). The curfew regulates the activity of juveniles during nighttime hours; it does not, by its terms, regulate expressive conduct. See Spence v. Washington, 418 U.S. 405, 410-11 (1974) (to be expressive, conduct must intend to convey a particular message, and the likelihood of that message being understood by others must be great). Nor can the curfew, on its face, be said to burden disproportionately those engaged in expressive conduct-the curfew covers all activities and provides a specific defense for juveniles engaged in First Amendment activities. Appellees suggest, however, that the curfew-even with the defense-will significantly deter juveniles from engaging in First Amendment activities in the first instance. But appellees have not provided a convincing argument as to why this might be so. Given that the First Amendment defense by definition provides full protection, any residual deterrent caused by the curfew would pose at most an incidental burden on juveniles' expressive activity or rights of association.

Finally, appellees argue that the curfew violates the Fourth Amendment because it allows a police officer to arrest an individual without probable cause. The curfew provides that a police officer may not make an arrest "unless the officer reasonably believes that an offense has occurred." § 6-2183(c)(1). This formulation, however, is precisely how the Supreme Court has defined probable cause, see Ker v. California, 374 U.S. 23, 34 (1963), and the curfew therefore conforms to the requirements of the Fourth Amendment.

* * * *

For these reasons, we conclude that the curfew law is constitutional. Accordingly, we reverse the district court's grant of summary judgment in favor of appellees and remand for the district court to enter summary judgment for the District of Columbia.

So ordered.

Harry T. Edwards, Chief Judge, Concurring in part and Concurring in the result, with whom Circuit Judges Wald and Garland join in Part II:

In my view, the disputed curfew law implicates significant rights of both minors and parents and, accordingly, is subject to no less than so-called "intermediate scrutiny." I therefore do not join Part II of the opinion for the court, which rests on the proposition that the curfew law does not implicate the fundamental rights of minors or their parents.1 However, generally for the reasons cited in Part III.A of the opinion, I agree that the law survives intermediate scrutiny with respect to the rights of minors. I also agree that, in the final analysis, the law survives intermediate scrutiny with respect to parents' rights as well. Accordingly, I concur in Parts I, III.A, and IV, and I concur in the result reached in Part III.B. I do not join the analysis underlying Part III.B, because I start from a very different premise. In my view, parental rights are implicated in this case and they are truly significant-indeed, these rights are at the core of our society's moral and constitutional fiber. I have more than a little difficulty in finding that the curfew law passes constitutional muster as against the claim of parents.

Part II of the opinion for the court suggests that the fundamental rights accorded to parents are limited to "the parents' control of the home and the parents' interest in controlling, if he or she wishes, the formal education of children." This section of the opinion concludes that this right "does not extend to a parent's right to unilaterally determine when and if children will be on the streetscertainly at night." It goes on to hold that the curfew law does not implicate any fundamental rights of parents, because limitations on where one's child may be at night are "not among the 'intimate family decisions' encompassed by such a right." In Part III, the opinion holds, alternatively, that, "even if the curfew implicated fundamental rights of ... parents," the curfew law survives intermediate scrutiny. The opinion acknowledges in a footnote that "a substantially broader formulation" of parental rights than that discussed in Part II.B is assumed for the purposes of Part III.B. However, the opinion never specifically defines what fundamental parental rights are at issue here. Some explication is necessary, I think.

Certainly it should be clear that parents' rights cannot be limited to only those activities that are within the home or involve the formal education of one's child-such a formulation is much too narrow. I do not agree with the suggestion in Part II.B of the opinion for the court that parents' rights are limited solely to "intimate family decisions," unless "intimate" is meant to include more than just what goes on within the confines of the home and with regard to the child's education. As numerous Supreme Court decisions make clear, a parent's stake in the rearing of his or her child surely extends beyond the front door of the family residence and even beyond the school classroom.

Over fifty years ago, the Supreme Court broadly stated that "[i]t is cardinal with us that the custody, care and nurture of the child reside first in 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 (1944); accord Reno v. ACLU, 521 U.S. 844, 865 n.31 (1997). More recently, the Court has recognized that the parental right to raise children in the manner that the parents see fit is deeply entrenched:

"The history and culture of Western civilization reflect a strong tradition of parental concern for the nurture and upbringing of their children. This primary role of the parents in the upbringing of their children is now established beyond debate as an enduring American tradition."

Wisconsin v. Yoder, 406 U.S. 205, 232 (1972); see also Stanley v. Illinois, 405 U.S. 645, 651 (1972) ("It is plain that the interest of a parent in the companionship, care, custody, and management of his or her children 'come[s] to this Court with a momentum for respect lacking when appeal is made to liberties which derive merely from shifting economic arrangements.' " (quoting Kovacs v. Cooper, 336 U.S. 77, 95 (1949) (Frankfurter, J., Concurring))); Pierce v. Society of Sisters, 268 U.S. 510, 534-35 (1925) (striking down state law requiring children to attend public schools as "interfer[ing] with the liberty of parents and guardians to direct the upbringing and education of children under their control").

To be sure, there are circumstances, as I discuss below, under which the state's interests may trump the rights of parents. To say, however, as Part II.B of the opinion for the court suggests, that a curfew law that regulates and restricts minors' activities outside the home during the nighttime hours does not even implicate the broad fundamental rights of parents is to disregard the teachings of decades of Supreme Court case law. The Court has never limited its definition of parental rights to include only the right to supervise activities that take place literally inside the home or literally inside the classroom. Indeed, such a limitation is implausible.

Surely a nighttime curfew law implicates parents' rights to control the "care," "nurture," "upbringing," "management," and "rearing" of their children, even if the law-by definition-regulates activity that takes place outside the home and school. The fact that some of the aforecited Supreme Court cases involve parents' rights to control the education of their children is not surprising, but neither is it evidence that the Court meant to imply that parents have no rights to control other aspects of their children's lives. Thus, when the Court explained in Ginsberg v. New York, 390 U.S. 629, 639 (1968), that "constitutional interpretation has consistently recognized that the parents' claim to authority in their own household to direct the rearing of their children is basic in the structure of our society," no one could reasonably believe that the Court meant to limit parents' authority to only child-rearing that takes place literally within the physical confines of "their own household." Such a view would come as a stunning surprise to countless parents throughout our history who have imposed restrictions on their children's dating habits, driving, movie selections, part-time jobs, and places to visit, and who have permitted, paid for, and supported their children's activities in sports programs, summer camps, tutorial counseling, college selection, and scores of other such activities, all arising outside of the family residence and school classroom. To ignore this reality is to ignore the Supreme Court's admonition in Yoder that the "primary role of the parents in the upbringing of their children is now established beyond debate as an enduring American tradition." 406 U.S. at 232.

There is no doubt that, in certain instances, the state may lawfully regulate the activity of children without regard to parental preferences. Indeed, the Supreme Court has noted that "the state has a wide range of power for limiting parental freedom and authority in things affecting the child's welfare," Prince, 321 U.S. at 167, and has permitted parental rights to be circumscribed to accommodate the Government's legitimate interest in the "moral, emotional, mental, and physical welfare of the minor," Stanley, 405 U.S. at 652 (internal quotation marks omitted). However, when the Government does intervene in the rearing of children without regard to parents' preferences, "it is usually in response to some significant breakdown within the family unit or in the complete absence of parental caretaking," Action for Children's Television v. FCC, 58 F.3d 654, 679 (D.C. Cir. 1995) (Edwards, C.J., Dissenting), or to enforce a norm that is critical to the health, safety, or welfare of minors. The difficult question, then, is how to accommodate both the state's interests and parents' rights where there has been no specific finding of a breakdown within an identified family unit and there is no indisputable threat to the health, safety, or welfare of minors.

It would be unreasonable to require the state to make a particularized showing that every child will benefit from a specific law enacted to protect the welfare of minors. For example, not every child will gain precisely equal benefits from child labor laws or education laws, but there is no doubt that the state may reasonably regulate education, see Yoder, 406 U.S. at 213, and that it may regulate and even prohibit child labor, see Prince, 321 U.S. at 166. Rather, the case law suggests that if there is a significant and important goal to be achieved that generally enhances the health, safety, or welfare of unemancipated minors, the state may pass legislation to achieve that goal, so long as the legislation does not unduly tread on parents' rights to raise their children.

There are three obvious categories of cases in which the state may pass legislation that is aimed at protecting children: (1) laws in which parents' rights are not accommodated, because accommodating parents' interests would defeat the entire purpose of the legislation, e.g., preventing parents from retaining custody of children they have abused; (2) laws in which parents' rights are not implicated at all, e.g., preventing convicted sex offenders from working in places where they would have substantial contact with children; and (3) laws in which parents' rights are implicated, but are accommodated.

This case involves the third category, i.e., accommodation. A good example of the "accommodation" category is found in the area of education. It is by now well-established that a state may enact compulsory education requirements; however, it is equally clear that the state must accommodate parents' rights to raise their children by allowing a child to attend private, rather than public school, see Pierce, or by allowing parents to teach their children at home, see Yoder. In other words, as long as certain standards are met, parents may educate their children as they see fit.

As the opinion for the court acknowledges, the Court in Prince appeared to engage in a more searching inquiry than mere rational basis review, although that case was decided before the Court had adopted the labels of strict scrutiny, intermediate scrutiny, or rational basis to characterize the appropriate standard of review. See Prince, 321 U.S. at 16570 & nn.15-16 (balancing the parental interest with the state interest and looking to child labor statistics for support). In my view, Prince and other such cases indicate that there must be a substantial relationship between the objectives of a law that limits parents' rights and the protection of children. Such a law must also reasonably accommodate parents' rights to raise their children as they see fit.

In this case, I have no real doubt that, as the opinion for the court shows, the curfew law is substantially related to the protection of minors from the dangers of juvenile crime. The difficult question here is whether the curfew law, in seeking to protect children, adequately accommodates parents' rights to determine what activities are necessary to their children's upbringing and growth. In my view, the D.C. law adequately accommodates parents' rights, because, although parents' decision making is not unfettered, the law allows parents great discretion in how to manage the activities of their children.

First, as the opinion for the court notes, § 6-2183(b)(1)(A) allows a minor to travel anywhere with a parent or other adult. In addition, subsection (B) allows minors to run "errands" for their parents, and I read this to include any task a parent may assign a child, including walking the family dog, running to the store for milk, and checking on an elderly family member. Furthermore, subsection (D) allows a minor to travel to and from work, and subsection (E) allows a minor to be out during curfew hours if necessitated by an emergency. Finally, subsection (G) allows a minor to attend any "official school, religious, or other recreational activity sponsored by the District of Columbia, a civic organization, or another similar entity that takes responsibility for the minor," or travel to or return from "an official school, religious, or other recreational activity supervised by adults and sponsored by the District of Columbia, a civic organization, or another similar entity that takes responsibility for the minor" during curfew hours. I read this exception to allow a minor to attend a movie at a local theater or musical concert at the Kennedy Center. Theaters are adult supervised, because an adult must be in charge of the premises while it is open, and may remove a patron if his or her behavior is inappropriate. Furthermore, business owners are generally responsible for the welfare of patrons on their premises, at least in the sense that owners must protect against obvious dangers. In short, when read broadly-as it should be to accommodate the significant parental rights implicated by the law-the law's list of exceptions leaves great room for the exercise of parental control.

In a different context, I have had much to say about the distinction between governmental regulations that facilitate parental rights as distinguished from those that impermissibly preempt parental rights. See Action for Children's Television v. FCC, 58 F.3d at 678-82 (Edwards, C.J., Dissenting). So I will not belabor the point further here. Suffice it to say, in my view, this case involves a situation in which the Government's interests are clear, as is the connection between the objectives of the law and the protection of minors. In fact, this is one of those unique cases in which the governmental regulations both serve to protect minors and, also, to facilitate parents' control over the activities of their children. See id. at 682 ("It would be hard to object to some sort of regulation of indecency in broadcast as well as other media were it narrowly tailored to facilitate parental supervision of children's exposure to indecent material."). No responsible parent would willingly send a child into danger. A law designed to curb the possibility of danger, while at the same time affording parents wide freedom to direct their children's activities, is one that passes constitutional muster. Although parental rights have been implicated by the curfew law, they have not been impermissibly infringed.

I therefore concur in the Conclusion that the curfew is constitutional, but only because I find that the curfew law is substantially related to the protection of children and that the rights of parents have been adequately accommodated.

Wald and Garland, Circuit Judges, Concurring in part and Concurring in the result:

For the reasons stated in the Fourth Circuit's opinion in Schleifer v. City of Charlottesville, 159 F.3d 843, 846-47 (4th Cir. 1998), as well as those expressed in Part II of Chief Judge Edwards' opinion and Part III of Judge Rogers' opinion, we conclude that the District of Columbia's Juvenile Curfew Act implicates the constitutional rights of children and their parents, and that intermediate scrutiny is the appropriate level of review. For the reasons stated in Part III of the Opinion of the Court, we conclude that the Curfew Act passes that scrutiny, and for the reasons stated in Part IV agree that it is otherwise constitutional as well.

Rogers, Circuit Judge, with whom Circuit Judge Tatel, joins, Concurring in part and Dissenting in part, and with whom Circuit Judge Wald joins in Parts II and III, and Circuit Judge Garland joins in Part III:

All members of the court agree that a test at least as rigorous as intermediate scrutiny would be proper for evaluating burdens on minors' fundamental right to freedom of movement. To the extent that the court hedges on the breadth of the right to free movement, however, the court mistakenly concludes that the right, if it exists at all, does not protect minors here.1 Were the plurality to define the right without regard to age, inasmuch as the Constitution applies to people of all ages, and consider age only in determining that minors can less successfully resist the interests of the government in their welfare, then it could avoid departing from traditional analysis of fundamental rights and suggesting that adults may lack a right to freedom of movement.

Even when the court assumes that the curfew burdens a fundamental right to movement, it fails to conform its application of intermediate scrutiny to Supreme Court instruction and example demonstrating that the proper judicial role requires attention to the evidence on which the legislature relies in intruding upon a fundamental right. When properly applied, intermediate scrutiny reveals that key elements of the curfew-age and time-are insufficiently tailored to address the problem of juvenile crime and victimization that confronted the legislature. By ignoring evidence that almost half of juvenile crime is committed by persons not covered by the curfew, and that most of that crime occurs at hours not within the curfew, the legislature has failed to demonstrate, on this record, the requisite fit between the problem and the chosen solution.

Enticed by the apparent success of curfews in other cities, the District of Columbia transplanted a Dallas, Texas ordinance without apparent determination that circumstances here warranted exactly the same solution. The Council of the District of Columbia had an accurate understanding that juvenile crime and victimization are serious problems, but, so far as the record shows, no accurate basis for concluding that nocturnal crime in certain public areas by youths under 17 was a sufficiently serious part of this problem to warrant severely limiting the rights of thousands of minors who were neither criminals nor likely victims of crime. The rhetoric supporting the curfew therefore does not fit the reality of what the curfew does. Consequently, the court's labored effort to construct a rationale for the curfew, attempting to avoid the inconveniences created by flawed and deficient information before the legislature, see, e.g., Op. at 18, eviscerates the distinction between intermediate scrutiny, which requires that justifications for complex, but burdensome, policy choices emanate from the legislature and that burdens be tailored to specific ends, and the less rigorous rationalbasis scrutiny, where the court defers to legislative policy choices with far less concern for serious evidentiary defects or loose tailoring.

Accordingly, because the court accords less respect to minors than is constitutionally required, and more deference to the D.C. Council than is constitutionally warranted, I respectfully Dissent from its holding that the curfew survives intermediate scrutiny.2

Claims invoking fundamental rights have been a source of institutional diffidence for Article III courts, which are reluctant to venture where "guideposts for responsible decisionmaking ... are scarce and open-ended." Collins v. City of Harker Heights, 503 U.S. 115, 125 (1992). Yet though the terrain may be unchartered, the Constitution's guarantees of "liberty" and "due process" are entrusted, along with countless others, to independent oversight by the judiciary. See, e.g., Roberts v. United States Jaycees, 468 U.S. 609, 618-20 (1984). Courts must carefully define the contested right, employing sufficient specificity to ground the right in a concrete application and sufficient generality to connect the right to its animating principles. See, e.g., Washington v. Glucksberg, 117 S. Ct. 2258, 2268 (1997); Griswold v. Connecticut, 381 U.S. 479, 481-85 (1965).

The parties differ as to how abstractly the court should define the right that plaintiffs invoke. Appellees-plaintiffs assert a broad right, regardless of age, to "freedom of movement," while appellant-defendant denies that juveniles have a fundamental right "to wander in public places at night without adult supervision." The United States, as amicus curiae, similarly opposes juveniles' alleged right "to roam the streets unsupervised" during curfew hours.

The plurality initially vacillates between reviewing a broad and narrow right, but ultimately views this case as raising only a narrow question. The opinion first suggests that plaintiffs invoke a right to "liberty," Op. at 6, but then proceeds as if this case has nothing to do with whether "Americans" in general have a right to "free movement" because it relates only to juveniles' claimed right to be free from adult supervision at night. See Op. at 9-10. The plurality seems to assume that the general right to free movement is entirely distinct from a right of (1) minors to (2) unsupervised movement (3) at night. This distinction between the right and a particular manifestation of it is an unhelpful means of weighing a state burden on an asserted liberty interest. Rather, by confronting the broader claim the court can develop meaningful standards to guide its review of the subsidiary claim that is directly at issue.

At first glance, the plurality's narrow construction of the contested right seems sensible. This country lacks a tradition of tolerance for the nocturnal wanderlust of minors, and the plurality's recognition of this uncontested fact avoids the more searching analysis that fundamental rights review entails.3 But, on closer inspection, the plurality's narrow statement of what is at issue relies on a suspect methodology.

First, defining a right as the mirror-image of a particular burden (i.e., the right to do the specific thing that a challenged rule prevents) tips the scales against recognizing the right. Safeguarding the abstract ideals of the Constitution frequently entails protecting conduct that many citizens find deeply offensive. See, e.g., Texas v. Johnson, 491 U.S. 397 (1989) (flag burning); Cohen v. California, 403 U.S. 15 (1971) (wearing jacket with "Fuck the Draft" in courthouse corridor); Brandenburg v. Ohio, 395 U.S. 444 (1969) (Ku Klux Klan rally). Hence, rights must be defined in a manner that will protect disfavored conduct while not needlessly constraining legislative and executive discretion. See, e.g., Kennedy v. Mendoza-Martinez, 372 U.S. 144, 160 (1963). By defining minors' rights in the narrowest sense possible, the plurality separates conduct that is discomforting to many adults from principles that animate due process doctrine. Disfavored conduct will rarely resist state regulation of its own force absent intervention of a more abstract guiding principle. By using the ostensibly neutral process of defining a right to transform a case about freedom of movement into one about nocturnal rambling, the plurality in effect ignores the role that abstract rights play in shaping constitutional discourse.

Second, the plurality's decision to define the asserted right narrowly confuses the ultimate question of balancing state interests against individual interests with the question of how to define an individual's interest with sufficient care to ensure that judicial review is not a hollow exercise of deference to conventional wisdom. The plurality has relied on the District of Columbia's strong defense of the curfew to hold that there is nothing to defend against-that there is no principle against which the curfew need be tested. See Op. at 10. The difficult issue in this case involves reconciling two conflicting interests: individual freedom to walk on public streets without fear of police intervention, see, e.g., Gomez v. Turner, 672 F.2d 134, 143 n.18 (D.C. Cir. 1982), and the authority of the state to act in the best interest of minors, see, e.g., Bellotti v. Baird, 443 U.S. 622, 633-34 (1979) (plurality opinion). This issue arises only if one recognizes a right at a sufficient degree of abstraction to connect with precedent in analogous areas. The plurality avoids this question by citing clear governmental interests-controlling the aimless wandering of minors in areas where harm can befall them-to eliminate any possibility that a contrary right may exist. Yet the fact that a state may have good reasons to treat the movement of minors differently from that of adults does not therefore mean that minors lack a right to movement; it means only that the right may in some circumstances be insufficient to overcome a particular burden. See Mississippi Univ. for Women v. Hogan, 458 U.S. 718, 724 n.9 (1982). Consequently, age should not be an element of the right at issue because the state interests that are relevant at the balancing stage of analysis do not aid the distinct inquiry at the definitional stage.

Third, construing rights narrowly displaces delicate value judgments, but does not avoid them. The admirable aim of narrowly defining a right is to "rein in the subjective elements that are necessarily present in due-process judicial review." Glucksberg, 117 S. Ct. at 2268. Broadly defined rights are prone to manipulation, and afford courts ample discretion when applying general principles to concrete fact patterns. Rights defined too narrowly, however, suffer from the opposite problem: the more specific the definition of a right, the more its vitality can become a question of judicial preference or unwarranted deference to legislative discretion because the court lacks external standards to guide its analysis. By asking a broader question, such as 'does a curfew impermissibly interfere with a generally applicable right of movement,' the court can gain access to standards and precedents to structure and guide its analysis. There may never be an objective answer to a claim involving the balance between individual rights and state interests, cf. Moore v. City of East Cleveland, 431 U.S. 494, 502-03 (1977) (plurality opinion), but some ways of framing the claim make the ensuing analysis more principled than others. See Poe v. Ullman, 367 U.S. 497, 541-45 (1961) (Harlan, J., Dissenting).

The plurality's methodology also obscures another, still deeper, value judgment. Here, the plurality defines the asserted right narrowly; in another case, the court might define a right more broadly, because the plurality does not articulate a standard to guide the process of defining rights. The court's choice about how abstractly to define a right may easily become influenced by its view of the underlying conduct at issue. Favored conduct will be integrated with similar cases that have protected analogous rights, while disfavored conduct will be relegated to unprotected isolation. Compare Franz v. United States, 707 F.2d 582, 595 (D.C. Cir. 1983) (recognizing "freedom of a parent and child to maintain, cultivate, and mold their ongoing relationship") with Dronenburg v. Zech, 741 F.2d 1388, 1395 (D.C. Cir. 1984) (rejecting right to "homosexual conduct in the Navy"). Although this subjectivity plagues any attempt to find an appropriate level of generality at which to define a right, it is more disconcerting where the court professes to act out of concern for judicial restraint. See Op. at 9.

Fourth, narrowly focusing on the movement rights of minors-as opposed to a right of movement generally-needlessly entangles equal protection and due process analysis by defining a fundamental right with reference to the class of people asserting it. Usually, due process challenges involve generally applicable rights, while equal protection challenges involve burdens that fall disproportionally on classes that share a disfavored trait. Here, appellees-plaintiffs have raised both types of claim under the Fifth Amendment. However, because they do not allege that youth is a suspect classification,4 their Fifth Amendment claims turn on the same question: whether the rights at issue are fundamental, such that burdens on minors' movement warrant heightened judicial scrutiny. Cf. Bearden v. Georgia, 461 U.S. 660, 66667 (1983). The plurality recognizes this overlap, see Op. at 5 n.1, but blurs the tests: by incorporating a class component (youth) into the definition of the right, the plurality avoids answering the difficult question of whether youth is an acceptable criteria for narrowing the scope of an otherwise applicable right (i.e., a right that would shield adults from a similar curfew), and instead assumes no rights are applicable.5

Finally, the plurality's reductionist reasoning relies on a methodology that the Supreme Court has repudiated. See Op. at 9. In Michael H. v. Gerald D., 491 U.S. 110 (1989), Justice Scalia foreshadowed the court's approach by suggesting that fundamental rights must be defined at "the most specific level at which a relevant tradition protecting, or denying protection to, the asserted right can be identified." Id. at 127 n.6. Higher "level[s] of generality" were to be avoided. Id. However, only Chief Justice Rehnquist joined this portion of Justice Scalia's opinion; Justices O'Connor and Kennedy, who joined the remainder of Justice Scalia's opinion, pointedly refused to concur in his Discussion of how to define fundamental rights. See id. at 132 (O'Connor, J., Concurring in part, joined by Kennedy, J.). Likewise, Justices Brennan, Marshall, and Blackmun rejected Justice Scalia's analysis, noting that it relied on a vision of the Constitution as a "stagnant, archaic, hidebound document steeped in the prejudices and superstitions of a time long past." See id. at 141-42 (Brennan, J., Dissenting).6 In the ten years since Michael H. was decided, Justice Scalia's approach to defining fundamental rights has never garnered a majority on the Supreme Court;7 yet a plurality of this court now embraces it, inviting the subjectivity that the plurality seeks to avoid.

From this analysis it follows that the contested right should be defined more abstractly in two ways: first without regard to age, and second without regard to the manner in which it is exercised. This section discusses the former issue, the next section discusses the latter. In neither section is it necessary to define a "right to liberty," Op. at 6, but neither is it necessary to disconnect the rights of minors at night from those of citizens in general, see Op. at 10.

The plurality defines a right that is coherent only in cases involving minors, as the age of the claimant is an element of the definition. Apparently, the plurality views freedom of movement as a privilege earned-if at all-by ritual passage into adulthood. Yet "[c]onstitutional rights do not mature and come into being magically only when one attains the state-defined age of majority. Minors, as well as adults, are protected by the Constitution and possess constitutional rights." Danforth, 428 U.S. at 74. The question here is whether "fundamental" rights, like "constitutional" rights more generally, apply to minors.8

There is no doubt that minors possess rights that are "fundamental,"9 including First Amendment10 and due process rights,11 as well as the right of equal protection to similarly situated children.12 Likewise, minors bear some of the burdens that accompany rights.13 The more difficult question is how to define the scope of these fundamental rights in view of the fact that "[t]he state's authority over children's activities is broader than over like actions of adults." Prince v. Massachusetts, 321 U.S. 158, 168 (1944); see also id. at 169. The Supreme Court has confronted this dilemma in various circumstances, in each case attempting to tailor concepts from adult jurisprudence to fit claims by juveniles. For example, minors "are entitled to a significant measure of First Amendment protection," Erznoznik v. City of Jacksonville, 422 U.S. 205, 212 (1975), but the "First Amendment rights of minors are not 'co-extensive with those of adults.' " Id. at 214 n.11 (quoting Tinker, 393 U.S. at 515 (Stewart, J., Concurring)). Similarly, in the due process context, "certain basic constitutional protections enjoyed by adults accused of crimes also apply to juveniles.... But the Constitution does not mandate elimination of all differences in the treatment of juveniles." Schall v. Martin, 467 U.S. 253, 263 (1984); see also McKeiver v. Pennsylvania, 403 U.S. 528 (1971).

The most reasonable reading of these cases is that minors and adults share many fundamental rights, but that the protective force of some of these rights is contracted or diluted when applied to minors. To the extent that a right defines a boundary to state authority, age is generally not a meaningful credential for access to the protected zone, "magically" conferring admission on a given birthday. There may be good reasons for making the boundaries of a right more malleable for minors than adults-states have stronger countervailing interests and minority status renders minors less competent to resist state intervention14-but not for denying the existence of the right altogether, at least not where minors are capable of exercising the right. Of course, where the rationale for a right raises questions about its suitability for minors, minors might not possess the right at all, as opposed to having a less robust version of it. For example, although there is a fundamental right to marriage, see, e.g., Turner v. Safley, 482 U.S. 78, 95 (1987), it might not apply below a certain relatively mature age. (The age of consent for marriage in the District of Columbia is 16. See D.C. Code § 30-103.) The developmental prerequisites for walking down a public street, however, are substantially lower than for the bundle of rights and responsibilities that attend marriage.

In a relative sense, a right that is "fundamental" for adults in their relationship with the state is equally fundamental, if not equally forceful, for minors because it defines the few areas of activity warranting especially careful tailoring of intrusive state means to worthy state ends. Minors, like adults, are able to enjoy the fruits of free movement and to chafe under its restriction, and thus there is little reason to link the fundamentally of the right to the age of the claimant. The cases on which the court relies to contract the scope of minors' rights are inapposite to curfews because they arise in unique contexts, such as challenges to school regulations and disciplinary procedures, involving state interests associated with the educational environment warranting enhanced control over minors' behavior. See, e.g., Vernonia School Dist. 47J v. Acton, 515 U.S. 646, 656 (1995); Hazelwood School Dist. v. Kuhlmeier, 484 U.S. 260, 266-67 (1988); Bethel School Dist. No. 403 v. Fraser, 478 U.S. 675, 685 (1986); Ingraham v. Wright, 430 U.S. 651, 681 (1977). Just as adults may have more freedom as civilians than as prison inmates or members of the armed forces,15 minors' rights vary depending on whether they are at home, on the streets, or in school.

The plurality assumes that minors cannot claim a right to be "unsupervised" because they are always in "some form of custody." Op. at 10. This characterization misses the point. Minors subject to the curfew are by definition unaccompanied by a responsible adult. To say that they are in some metaphysical bond of "custody" begs the question of whose custody they are in, and the extent to which certain personal prerogatives are immune from custodial restraint, at least by a government custodian. At a minimum, unaccompanied minors are not under direct government control, and thus theories of custody announced in a case dealing with incarcerated juvenile delinquents are unhelpful in assessing the burdens imposed by a curfew. See Op. at 10, citing Schall v. Martin, 467 U.S. 253, 265 (1984). The Supreme Court appeared to recognize as much in Prince, which relied on a balancing of state and parental interests rather than an undifferentiated notion of custody to regulate the activities of minors in public streets. See Prince, 321 U.S. at 164-71.16

For the reasons discussed, the conduct at issue should be more generally defined to encompass the activity of movement rather than how particular minors engage in it. The plurality's limited definition of the contested right appears to flow from an unarticulated perception of what minors might be doing while "freely wander[ing] ... at night." Op. at 10. How minors exercise, and whether they abuse, their right to movement is relevant in weighing the constitutionality of a contrary state burden, but should not be part of the definition of the right itself. Plaintiffs in this case contend that the curfew prevents them from using public streets as a means of conveyance from one place to another. See Complaint ¶¶ ¶¶ 3, 4, 6, 7, 11, 12, 13, 14, 15. They do not seek to linger in any one location, or to access any particular area, such as a park, that the District of Columbia might have a special reason to close. Rather, they protest a blanket restriction on their movement. Whether they plan to "wander" Op. at 10,-or amble, stroll, sashay, or saunter-is irrelevant; the only question under the Constitution is whether the District's action burdens a fundamental right to be on and to use public streets. When one chooses to walk, how one does so, where one goes, and what one does once there are factors relevant to reviewing burdens on the right, but not to defining the right itself. Therefore, the question before the court should be defined as whether there is a fundamental right to walk in public without thereby subjecting oneself to police custody; in short, a right to free movement.

The Supreme Court's jurisprudence on the right to "move" encompasses several distinct concepts. The discrete components include the right to relocate from state to state, the right to cross state borders for purposes other than relocation, the right to cross national borders, and the right to intrastate or localized movement. These rights are "fundamental" under established doctrine.17 As early as the Articles of Confederation, state citizens "possessed the fundamental right, inherent in citizens of all free governments, peacefully to dwell within the limits of their respective states, to move at will from place to place therein, and to have free ingress thereto and egress therefrom." United States v. Wheeler, 254 U.S. 281, 293 (1920) (emphasis added).

To date, however, the Court has not expressly held that there is a fundamental right to intrastate movement, possibly because it has not been seriously contested.18 While most of the cases discussing the "right to travel" or "right to free movement" have involved an interstate or international component, language in the decisions suggests that the right extends to purely local movement, see, e.g., Kolender, 461 U.S. at 358; Papachristou, 405 U.S. at 164; Kent, 357 U.S. at 126; Wheeler, 254 U.S. at 293; Bell v. Maryland, 378 U.S. 226, 255 (1964) (Douglas, J., Concurring), and at least two circuits have expressly agreed. See Lutz v. City of York, 899 F.2d 255, 268 (3d Cir. 1990);19 King v. New Rochelle Mun. Housing Auth., 442 F.2d 646, 648 (2d Cir. 1971).20 This circuit has also recognized the value of free movement, noting that the ability to "walk the streets, without explanations or formal papers, is surely among the cherished liberties that distinguish this nation from so many others." Gomez v. Turner, 672 F.2d 134, 143 n.18 (D.C. Cir. 1982); see also Waters v. Barry, 711 F. Supp. 1125, 1134 (D.D.C. 1989). Thus, simply being on a public street, without some further incidence of misfeasance, is usually not a crime. Cf. Shuttlesworth v. City of Birmingham, 382 U.S. 87, 96 (1965) (Douglas J., Concurring).

The importance of intrastate mobility is apparent from its utility and the implications of its denial. As Justice Douglas explained:

"Freedom of movement, at home and abroad, is important for job and business opportunities-for cultural, political, and social activities-for all the commingling which gregarious man enjoys. Those with the right of free movement use it at times for mischievous purposes. But that is true of many liberties we enjoy. We nevertheless place our faith in them, and against restraint, knowing that the risk of abusing liberty so as to give rise to punishable conduct is part of the price we pay for this free society."

Aptheker, 378 U.S. at 519-20 (Douglas, J., Concurring). Plaintiffs have asked for nothing more than the "cultural, political, and social ... commingling" that free movement permits. For example, one would like to go to swimming practice, Complaint at ¶¶ 4, another to ballet performances, id. at ¶¶ 11, and another to dances and late-night movies, id. at ¶¶ 16. Viewed in isolation, these activities are of no great constitutional moment; viewed together, they constitute the rhythm of daily life for our city's youth, and the fruits of a stable pluralist society tolerant of individual liberty. Thus, even if this case raises a purely intrastate question-which is not at all clear21-precedents recognize a fundamental right to walk through public streets without thereby subjecting oneself to police custody.22

The plurality apparently fears that "lightly extend[ing]" the right to movement will require searching review of trivial or incidental impediments to movement that do not bear any relation to the "basic notions" that animate the right. Op. at 9. These concerns are misplaced. As with any right, the right to free movement is not unlimited; reasonable burdens, including those that are "incidental[ ] and remote[ ]"-are acceptable. Williams, 179 U.S. at 274; see also Shapiro, 394 U.S. at 629; Califano v. Aznavorian, 439 U.S. 170, 177 (1978); Lutz, 899 F.2d at 269. Cf. Glucksburg, 117 S. Ct. at 2282 n.8 (Souter, J., Concurring); Burdick v. Takushi, 504 U.S. 428, 434 (1992). For example, the Supreme Court has noted that the government might bar travel to certain regions in emergencies and may constrain the travel options of certain classes of citizens, such as felons. See Zemel v. Rusk, 381 U.S. 1, 15 (1965); Jones v. Helms, 452 U.S. at 420. Likewise, regulating conduct in public spaces and legitimate law enforcement objectives, see, e.g., Terry v. Ohio, 392 U.S. 1 (1968), may also justify burdens on free movement. These limits should vitiate the plurality's concern that recognizing a right to free movement would impair a state's authority to operate traffic lights. See Op. at 9. The right to free movement does not shield all conduct of which movement is a component, but simply protects an individual from police interference for mere presence, without more, on a public street.

Moreover, the plurality's preoccupation with incidental burdens is misplaced. Whatever else the curfew might be, it is not an incidental burden. The curfew does not cover a few specifically identified people, it covers a class of thousands; it does not apply to a few discrete areas, but to an entire city; it does not constrain specific types of movement, but with few exceptions bars all movement in public; it is not confined to a brief period, but extends for roughly 25% of the day. In short, the imagined consequences of recognizing the proposed right are inapposite, exaggerated, and can be addressed by settled doctrine.

Having concluded in Part II that the curfew burdens a fundamental right, I join the court in holding, as has the Fourth Circuit, see Schleifer v. City of Charlottesville, 159 F.3d 843, 847 (4th Cir. 1998), cert. denied, 119 S. Ct. 1252 (1999), that the appropriate standard of review is intermediate scrutiny. See Op. at 14; see also Hutchins, 144 F.3d at 809-10 (opinion of Rogers, J.).23

Fifth Amendment substantive due process and equal protection scrutiny is generally two-tiered: strict scrutiny applies to burdens on fundamental rights, while rational basis scrutiny applies to burdens on rights that do not qualify as fundamental. See, e.g., Glucksberg, 117 S. Ct. at 2271; Heller v. Doe, 509 U.S. 312, 320 (1993). Under either standard, courts must determine whether the state's interest in imposing a challenged burden is sufficiently weighty, and whether the state's means are sufficiently tailored to its ends. Strict scrutiny demands narrow tailoring to a compelling interest, see Reno v. Flores, 507 U.S. 292, 302 (1993), while rational basis review demands a rational relationship to a legitimate interest. See Ohio Bureau of Employment Serv. v. Hodory, 431 U.S. 471, 489 (1977). Between these poles lies intermediate scrutiny, which allows more refined analysis than usuallyfatal strict scrutiny and rarely-fatal rational basis review. To satisfy intermediate scrutiny, a burden must be substantially related to an important interest. See United States v. Virginia, 518 U.S. 515, 533 (1996). As explained in Part IV, this standard is flexible enough to respect state regulatory prerogatives while exacting enough to protect individual rights from unnecessary encroachment.

Nothing inherent in the definition of a fundamental right requires that "strict scrutiny" apply here. While burdens on fundamental rights trigger the most exacting review available, which as to adults is strict scrutiny, it is possible for a less stringent standard to be the most exacting available for minors. See Carey v. Population Serv. Int'l, 431 U.S. 678, 693 n.15 (1977) (plurality opinion). Even though there is a formalistic allure to treating all fundamental rights alike, and therefore applying strict scrutiny to laws regulating minors as well as adults, to do so would ignore the real, and legally accepted, differences between minors and adults. As noted in Part I, minors and adults share basic rights, but these rights have less force when used by minors as shields against regulation. Unduly intrusive judicial scrutiny of laws burdening minors would fail to respect the relative amenability of minors to regulation and would demand too much justification from government in an area in which it frequently must act. Cf. Burdick v. Takushi, 504 U.S. 428, 433-34 (1992). Given that the force of the right to movement varies with the status of the people asserting it, the standard of review must be sensitive to the context in which it is applied. As Justice Frankfurter cautioned, "[l]egal theories and their phrasing in other cases readily lead to fallacious reasoning if uncritically transferred to determination of a State's duty towards children," May v. Anderson, 345 U.S. 528, 536 (1953) (Frankfurter, J., Concurring).

When a minor's fundamental right to movement is at issue, intermediate rather than strict scrutiny is most appropriate.24 The essence of intermediate scrutiny, as distinct from rational basis review, is that the government must tailor its burden to relatively specific and important ends and justify incidents of the law that exceed or depart from those ends. Tailoring is particularly important when the rights of minors are at stake, inasmuch as substantial discrepancies between the treatment of adults and minors have often turned on unsubstantiated assumptions rather than persuasive evidence. The Supreme Court's opinion in In re Gault, 387 U.S. 1 (1967), which invalidated procedures in juvenile courts that vastly differed from procedures in adult courts, is instructive. In Gault, the Court recognized the state's special interest in providing informal Justice for juveniles, but was concerned by the magnitude of the 'reforms' that states adopted in pursuit of this interest, stating that "[s]o wide a gulf between the State's treatment of the adult and of the child requires a bridge sturdier than mere verbiage, and reasons more persuasive than cliche can provide." Id. at 29-30; see also id. at 21-22. The Gault holding reflects judicial concern for ensuring a reasonable "fit" between legitimate state ends and the means adopted to advance them in cases predicated on distinctions between juveniles and adults. Such scrutiny ensures that regulations that disproportionately burden juveniles are wellconsidered and not merely well-intentioned.

Some juvenile curfews may survive intermediate scrutiny, but the present curfew does not. The curfew has legitimate ends, but the D.C. Council inadequately tailored its means to these ends in light of the severe burdens that the curfew imposes on minors' fundamental rights.

To survive intermediate scrutiny, statutory burdens must be substantially related to an important government interest. See, e.g., Clark v. Jeter, 486 U.S. 456, 461 (1988); Hogan, 458 U.S. at 724. Review under this standard is far from "toothless," Mathews v. Lucas, 427 U.S. 495, 510 (1976), and this court has given it meaningful bite. See Lamprecht v. Federal Communications Comm'n, 958 F.2d 382, 391-98 (D.C. Cir. 1992) (per Thomas, Circuit Justice). The standard places duties on both legislatures and courts: legislative analysis must be "reasoned," and judicial analysis must be "searching." Hogan, 458 U.S. at 726, 728. Only burdens that demonstrate a reasonable fit-or "congruen[ce]"-with their benefits may withstand scrutiny. See, e.g., Turner Broad. Sys. v. Federal Communications Comm'n, 520 U.S. 180, 215 (1997); Board of Trustees v. Fox, 492 U.S. 469, 480 (1989). A legislature seeking to protect minors need not produce "scientifically certain criteria of legislation," Ginsberg, 390 U.S. at 643 (citation omitted), but neither can it rest on unsubstantiated speculation. See, e.g., Gault, 387 U.S. at 29-30. Or as this circuit has put it, "[a]ny 'predictive judgments' concerning group behavior and the differences in behavior among different groups must at the very least be sustained by meaningful evidence." Lamprecht, 958 F.2d at 393.

The curfew clearly satisfies the "important interest" requirement of intermediate scrutiny. The curfew seeks to reduce crime by and against minors, and to assist parents and guardians "in carrying out their responsibility to exercise reasonable supervision of minors." D.C. Code § 6-2181(e)(1)(3). Each is a laudable goal. See, e.g., Hodgson, 497 U.S. at 444; Schall, 467 U.S. at 264; Bellotti, 443 U.S. at 637 (plurality opinion). As the court notes, the D.C. Council was presented with a wealth of evidence of the seriousness of the juvenile crime problem in the District of Columbia. See Op. at 15. The difficulty, however, lies in the D.C. Council's Conclusion that these ends warrant the particular burdens that the curfew imposes on minors. There are many ways to reduce juvenile crime and victimization and to strengthen family units, some of which are more extreme than others. The question here is whether the curfew is too extreme given the evidence considered by the D.C. Council before adopting it. Cf. Plyler v. Doe, 457 U.S. 202, 229 n.25 (1982).

The curfew has three essential elements: it operates on a defined class in defined places at defined times. The District government defends each definition with statistical evidence cataloging a severe epidemic of juvenile crime and victimization. While juveniles are the source of and victims of an intolerably large volume of crime in the District, examination of the record reveals that the evidence does not fit the definitions that the D.C. Council crafted. See Craig v. Boren, 429 U.S. 190, 200 (1976).

First, the evidence upon which the D.C. Council relied is too broad because it documents a problem that the curfew does not address. The curfew applies only to persons under 17, but the statistics include crimes by youths as old as 17 and victimization of youths as old as 19.25 See 942 F. Supp. at 675; Annie E. Casey Foundation, Kids Count Data Book: State Profiles of Child Well-Being at 49 (1995). This statistical anomaly is more than technical because approximately 42% of all juvenile referrals in the District of Columbia courts from 1990-1994 involved youths over age 16.26 Relying on data that includes youths aged 17 therefore significantly overstates the problem that a curfew limited to those under 17 can solve. The District government is of course free to limit a curfew to whatever ages it deems appropriate, but it may justify the curfew only with data that is relevant to the targeted ages. Here the District has not explained why the curfew targets substantially less crime and victimization than outlined in the data offered to support it, and the court accordingly has no basis for deferring to the legislature's decision to impose a curfew that excludes minors seventeen and older while burdening minors under seventeen.

Second, the evidence on which the D.C. Council relied is also too narrow because it does not indicate when juvenile crime and victimization occur.27 Such information is critical to assessing a curfew, which does not directly affect crime outside of curfew hours.28 Again, this evidentiary defect is more than merely technical because uncontested evidence indicates that, nationwide, juvenile victimization is most prevalent during after-school hours at around 3-4 p.m.,29 and FBI statistics show that violent juvenile crime peaks in the mid- to late-afternoon.30 The D.C. Council has discretion to address only part of a larger problem, and therefore may enact a curfew even if it will not solve all juvenile crime. Cf. New Orleans v. Dukes, 427 U.S. 297, 303 (1976). However, before burdening a fundamental right, the legislature must have a clear picture of the problem it is addressing. See Craig, 429 U.S. at 200-04.31 Intermediate scrutiny, by contrast with rational basis review, requires that a legislature pay more attention to detail than the record indicates was expended in the instant case; otherwise, a court cannot determine if an ordinance is appropriately tailored to the details it addresses. See Phillips v. Borough of Keyport, 107 F.3d 164, 174 (3d Cir. 1997) (in banc). Here, the D.C. Council had ample evidence of a general juvenile crime problem, but far too little evidence describing the specific problem that it chose to address in an extraordinarily burdensome way.

The weakness of the evidence that the D.C. Council did consider is particularly troubling in light of evidence it did not consider. As the district court noted, the D.C. Council ignored evidence showing that more than 90% of all juveniles do not commit any crimes, at night or otherwise. See 942 F. Supp. at 676. The curfew thus burdens a far larger class of minors than are responsible for crime or at risk because of it. If the D.C. Council had decided that the benefits of the curfew for a subset of the affected class (or the public in general) were worth the costs to the entire class, the court might properly defer to legislative discretion. But because there is virtually no record to indicate that the D.C. Council assessed the extent to which the affected class was responsible for or at risk from the targeted activities, and whether the targeted ages and hours were a significant component of the perceived problem, the foundation for deference evaporates. This view is consistent with the purpose of intermediate scrutiny, which does not require the least restrictive means necessary to satisfy important governmental interests, but does result in judicial invalidation of laws that burden "substantially" more rights than necessary. Ward v. Rock Against Racism, 491 U.S. 781, 799 (1989); cf. Pickett v. Brown, 462 U.S. 1, 17-18 (1983); Plyler, 457 U.S. at 228-29.

If the curfew did not burden fundamental rights, these evidentiary defects would not warrant judicial intervention under rational basis scrutiny. See Exxon Corp. v. Governor of Maryland, 437 U.S. 117, 124 (1978) (citations omitted). Unlike an administrative agency, which generally must explain the basis for the rules it promulgates, see 5 U.S.C. § 553(c); Securities & Exch. Comm'n v. Chenery Corp., 318 U.S. 80, 88 (1943), legislatures need offer express rationales for statutes, and courts rarely scrutinize the legislative process to determine if adequate evidence justifies its work product. Cf. Turner Broad. Sys., 520 U.S. at 195-96. But when legislation substantially burdens a fundamental right or relies on a disfavored class distinction, judicial scrutiny intensifies to examine the need for and scope of challenged statutes. See, e.g., Mills v. Habluetzel, 456 U.S. 91, 101 n.9 (1982); Trimble v. Gordon, 430 U.S. 762, 771-72 (1977); Lamprecht, 958 F.2d at 391-92 (per Thomas, Circuit Justice). In such cases, the state cannot rely on its lawyers to sift through the record and cobble evidentiary shards into a posthoc rationalization. See Craig, 429 U.S. at 200 n.7; cf. Maine v. Taylor, 477 U.S. 131, 149 (1986); Hughes v. Oklahoma, 441 U.S. 322, 338 n.20 (1979). Rather, for a legislative judgment to warrant judicial deference, there must be a contemporaneous factual foundation from which the court can conclude that there is a close nexus between the burden on fundamental rights and the important state interest. See, e.g., Turner Broad. Sys. v. Federal Communications Comm'n, 512 U.S. 622, 666 (1994) (plurality opinion). The importance of the District of Columbia's interest is evident, but the congruence between the particular curfew it enacted and that interest is only minimally developed.

The Supreme Court has repeatedly demonstrated that, under intermediate scrutiny, it will not tolerate a severe burden on a fundamental right simply because a legislature has concluded that the law is necessary. Rather, the Court has independently examined the evidence before the legislature to determine whether an adequate foundation justified the challenged burdens. For example, in Craig, the Supreme Court held that the Oklahoma legislature lacked an adequate basis for permitting women to consume low-alcohol beer at a younger age then men.32 See 429 U.S. at 204. The Court recognized that the state had legitimate interests in traffic safety and public health, see id. at 199-200, but found that the gender distinction did not "closely serve[ ]" these interests. Id. at 200. Although the state proffered statistics showing that young men were more likely than women to be arrested or injured in alcohol-related traffic incidents, the Court questioned the accuracy of these statistics and closely scrutinized the numerical correlations on which the state relied, concluding that the data provided an "unduly tenuous 'fit.' " Id. at 201. Likewise, the Court noted that the data did not address the "salient" characteristics of the challenged burden because it did not expressly relate sex, age, and consumption of the specific type of alcohol at issue. See id. at 202-03. The looseness of these statistics is disturbingly parallel to the evidentiary shortcomings in the instant case because the present record lacks evidence of a connection between the salient characteristics of age, time, and violence.

As in Craig, a plurality of the Supreme Court in Turner Broadcasting refused to accept that interests which in the "abstract" were important could "in fact" justify a particular burden. 512 U.S. at 664. In Turner, where the Supreme Court was asked to affirm a decision by Congress to require cable operators to carry local broadcast signals, the Court recognized that Congress was entitled to "substantial deference," but refused to uphold the statute because the record provided insufficient evidence of a "genuine" problem creating a "need" for the particular burdens that Congress imposed. Id. at 665. Rather than rely on legislative "findings," the Court remanded for further development of facts sufficient to permit the judiciary to fulfill its "obligation to exercise independent judgment" and test Congress's inferences against the record. Id. at 666. The Court also rejected statistics proffered by the government because they were either too general or failed to address the salient features of the regulations. For example, statistics showing that the programming rules would prevent broadcasters from being dropped from cable systems were unhelpful because they did not explain what the consequence of such action would be, and whether there was a "serious risk of financial difficulty" for broadcasters absent the regulation. Id. at 667. Likewise, the Court faulted the "paucity" of evidence describing the precise burdens that the statute imposed on cable operators because the absence of such evidence precluded the court from determining whether the burdens were substantially broader than necessary to achieve Congress's goals. See id. at 667-68. This evidentiary failure is similar to the problem in the instant case: this court lacks sufficient evidence to determine whether the curfew restrains too many minors in too severe a manner in light of the volume of crime for which minors of the targeted ages are responsible during the targeted hours.

This court has been similarly vigilant when applying intermediate scrutiny. In Lamprecht, the court, writing through Circuit Justice Thomas, reviewed gender preferences within the FCC's scheme for licencing radio stations. Recognizing that it must defer to the policy judgments of Congress and the FCC, the court nevertheless demanded "meaningful evidence" of a link between the rule and an important purpose. 958 F.2d at 393. It then went on to dissect the statistics supporting the gender distinction, concluding that awarding women licences solely on the basis of gender did not advance the goal of programming diversity because, among other reasons, stations owned primarily by women were only 1.25 times more likely to broadcast "women's programming" than stations owned by men. See id. at 397. The court concluded that this correlation, and similar evidence, was an insufficient predicate to survive intermediate scrutiny. See id. at 398.

Decisions of other circuits affirming curfews do not suggest a contrary methodology, as the curfews under review were founded upon sturdier evidence. In Schleifer, the Fourth Circuit reviewed a curfew enacted by Charlottesville based on specific data documenting a crime problem in that city with reference to the age of offenders, see 159 F.3d at 850, the time of occurrence, see id., and the place of occurrence, see id. at 851. Moreover, the city supplemented evidence of the effects of curfews in other cities with specific analysis relating these studies to local circumstances. See id. at 850. This greater effort at tailoring established the requisite congruence and thus led the Fourth Circuit to conclude that the curfew is "a meaningful step towards solving a real, not fanciful problem." Id. at 849.33 By contrast, there is little basis in the present record on which the court may rely to make the same statement about the D.C. curfew, or to conclude that the curfew is not substantially over-restrictive.

Given the inadequacy of the District's statistics, all that remains to justify the curfew are bare assumptions about the demographics of crime and conventional political wisdom. Neither is sufficient to justify a sweeping restriction of minors' fundamental right to movement. See Turner Broad. Sys., 512 U.S. at 664 (plurality opinion); Weinberger v. Wiesenfield, 420 U.S. 636, 643 (1975); Gault, 387 U.S. at 29-30. Cf. Cleburne Living Ctr., 473 U.S. at 448-49. If the legislature wants to solve pressing problems by carving exceptions to fundamental rights, intermediate scrutiny requires that it use a restrained and delicate blade; here, the D.C. Council sliced broadly with too little regard for available evidence.

Nor can the evidentiary deficiencies be overcome by looking to the experiences of other cities, as the court and the District of Columbia urge. The experience of other cities with law enforcement tools may be relevant and may provide useful information to inform the D.C. Council's decisions. But this is not the same as saying that the tools used by other cities can be imported without consideration of the characteristics of the two communities. In concluding that the D.C. Council could properly rely on the experiences of New Orleans, San Antonio, and Dallas with juvenile curfews, the court relies on Renton v. Playtime Theaters, Inc., 475 U.S. 41, 52-53 (1986), in which the Supreme Court acknowledged that intermediate scrutiny permits one jurisdiction to rely on evidence accumulated by another addressing a similar problem. Compare City of Richmond v. J.A. Croson Co., 488 U.S. 469, 505 (1989). Yet under Renton, a city may rely on data collected in another city only "so long as whatever evidence the city relies upon is reasonably believed to be relevant to the problem that the city addresses." Renton, 475 U.S. at 52-53. Indeed, Renton and Seattle, the city that had gathered the data on which Renton relied, chose different remedies to their common problem. See id. at 52. By contrast, the D.C. Council appears to have adopted the Dallas ordinance "wholesale" without attempting to tailor it, save in a few very inconsequential ways, to the District's circumstances. 942 F. Supp. at 678. The need for substantial tailoring precludes off-the-rack solutions on the scale present here. See Renton, 475 U.S. at 52-53. Thus, while Renton's reasoning may be applicable, the D.C. Council failed to establish a fit between local circumstances and the borrowed ordinance and data.34

Finally, efficacy can be no substitute for constitutional scrutiny. See Op. at 20. Assuming that the decline in arrests of juveniles during curfew hours demonstrates the curfew's effectiveness during its brief three-month period of operation, the efficacy of the curfew cannot alone save it from constitutional infirmity.35 The fact that well-enforced nocturnal juvenile curfews reduce crime is hardly surprising; minors cannot readily injure the public when not permitted to mingle with it. But it is equally clear that a nocturnal adult curfew would also reduce crime, as would extending the present juvenile curfew to cover the entire day.36 Yet both options would be extreme, and raise the same question as the instant case: whether the severity of the District of Columbia's remedy is warranted by a substantial relation to an important interest. A court reviewing an adult curfew could not substitute effectiveness as a proxy for constitutional propriety, and this court likewise must look beyond any apparent attractiveness of the curfew to determine if it is a constitutionally acceptable exercise of legislative authority.

In a time too-often punctuated by reports of senseless youth violence and untimely death, and of promising lives lost to the sadly familiar vices of the streets, minors are easy targets of ambitious law enforcement measures, as well as well-intentioned government paternalism, and cannot readily defend their rights in political fora. When challenges to legislative reforms are presented, it falls to the courts to ensure that the political branches respect minors' rights even as they exercise their considerable discretion to assess and promote minors' best interests in the face of pervasive threats. See, e.g., Gault, 387 U.S. at 21-22. The court appropriately concludes that intermediate scrutiny best serves this important but limited judicial role of protecting fundamental rights while deferring to delicate legislative judgments. Applying such scrutiny to the record at hand, the court falters, however, attempting to finesse the congruence required by intermediate scrutiny. Accordingly, I respectfully Dissent, concluding that in the absence of a record warranting deference the curfew does not survive the heightened scrutiny that accompanies the burdens it places on minors' right to free movement.

I agree with Judge Rogers that the District of Columbia juvenile curfew implicates a fundamental right to free movement and that the right should be defined without regard to the age of the right-holder. See Rogers Op. at 4-12. Although I still believe that the curfew should be subject to strict scrutiny and that the compelling interest prong of the analysis can adequately account for "the government's legitimate need to regulate minors," Hutchins v. District of Columbia, 144 F.3d 798, 826 (D.C. Cir. 1998) (Tatel, J., Concurring in the judgment), I join Judge Rogers's Conclusion that this curfew fails to survive even intermediate scrutiny. Modeled nearly verbatim on a Dallas juvenile curfew "without apparent determination that circumstances here warranted exactly the same solution," Rogers Op. at 2, and made permanent by the D.C. Council without any assessment of its effectiveness simply to avoid mooting this litigation when the initial temporary measure expired, see Hutchins, 144 F.3d at 827 (Tatel, J., Concurring in the judgment), the D.C. curfew applies at specific times to juveniles of specific ages despite virtually no record evidence that the particular restrictions will deter crime by and against the city's youth. See Rogers Op. at 22-25. Indeed, to conclude on this record that the juvenile curfew survives intermediate scrutiny, as this court now does, strips an already elastic standard of any semblance of heightened review, with grave consequences for other rights protected by intermediate scrutiny. See Mississippi Univ. for Women v. Hogan, 458 U.S. 718, 724 (1982); Craig v. Boren, 429 U.S. 190, 197 (1976). I write separately to express my view that quite apart from the question of its constitutionality with respect to the rights of minors, the D.C. curfew fails to survive the strict scrutiny triggered by the restriction it imposes on parents' fundamental right to control the upbringing of their children.

I know that many parents believe that the curfew reinforces their efforts to ensure their children's safety and proper upbringing. Indeed, one of the curfew's stated purposes is to "[a]id[ ] parents or guardians in carrying out their responsibility to exercise reasonable supervision of minors entrusted to their care." D.C. Code Ann. § 6-2181(e)(3) (Supp. 1998). As Chief Judge Edwards and Judge Silberman observe, moreover, the law contains several "defenses" that to some extent preserve parents' control over their children's activities. See Silberman Op. at 22 (citing D.C. Code § 6-2183(b)(1)(A), (B), (D), (E), (G)); Edwards Op. at 6 (same).

Restating the legislative judgment that "the curfew facilitates rather than usurps parental authority," Silberman Op. at 22, however, does not answer plaintiffs' assertion of parental rights. Whatever views the Judges of this court, members of the D.C. Council, or even the majority of D.C. parents may have regarding the range of discretion needed for proper parenting, the relevant fact is that plaintiffs in this case disagree. In their complaint, see Complaint ¶¶ ¶¶ 5, 8, 10, 16, 33, 43, and uncontroverted affidavits, they claim that the curfew interferes with their ability to raise their children as they see fit. For example, Kimberly Denise Dean, a plaintiff who lives in Northeast Washington, said this:

"I am the mother of Natiya Daniel Tapper, who is 14 years old, and subject to the District of Columbia's new curfew law. I have one other child, Qiana Shontay Dean, who is 17 years old."

"I have taken great care to raise my daughters and hope they will grow up to be responsible adults. Naturally, this includes setting limits on their activities, such as hours by which they should be in at night. However, the curfew law, [sic] takes away my parental discretion to set those limits. As a responsible parent, I do not often allow my fourteen-year-old child, Natiya, to go out after 11:00 p.m. However, there are times when I decide after careful consideration that she should be allowed to participate in activities that require her to be out after 11:00 p.m."

"For instance, last May I allowed Natiya to help Qiana celebrate her seventeenth birthday. Qiana, Natiya and a couple of Qiana's girlfriends ate dinner at a local restaurant, saw a late-movie, and then completed the celebration with an early breakfast. My daughters did not arrive home until after 2 a.m."

"...."

"Soon Natiya will be in high school, and I expect that, like her sister, she will become more involved in social activities that will keep her out late at night. I will try to make wise decisions about whether to allow Natiya to engage in these activities when the time comes. The curfew law, if allowed to stand, will unfairly restrict Natiya's legitimate social activities and interests, as well as my ability to raise Natiya in the way that I see fit."

Dean Decl. ¶¶ ¶¶ 2-4, 6 [JA 402-03]. Another plaintiff, Robert Jablon of Northwest Washington, said:

"My wife and I have taken great care to try to raise our children so that they will-we hope-grow into responsible adults.... [J]ust as part of teaching children about responsible behavior involves setting limits, part of that teaching also involves showing them that rules are not rigid, and that reasonable exceptions should be made when there is good justification. Accordingly, my wife and I allow [our eleven-year-old son] Joel to stay out late from time to time, or to go out early in the morning, when in our view there is an appropriate reason. For example, we regularly allow our son Joel to walk our family dog, Calle, around the block before going to bed at night, which could be after midnight during the summer or before 6:00 a.m. We have also allowed Joel to ride his bike to and from a neighborhood friend's house four or five blocks away when Joel is invited to attend a movie and to return home after midnight on a weekend night or during the week in summer.... It usurps our role as parents for the government to step in and tell us and our children that we cannot make those decisions for ourselves, and it threatens to make us, as well as our children, criminals if we exercise parental discretion in customary, reasonable ways."

Jablon Decl. ¶¶ 3 [JA 423-24].

Even if walking the family dog could be classified as an "errand" under the curfew's defenses, see Edwards Op. at 6, no fair reading of the law would allow parents to permit their children during curfew hours to participate in a birthday celebration or ride a bike to a friend's house. The curfew likewise eliminates parents' discretion to allow their children to take an early-morning jog through the neighborhood, go to a restaurant with friends after a school dance, or-as the District conceded at oral argument-"go out to a friend's house to do math homework at night" unaccompanied by an adult. Oral Arg. Tr. at 17. The D.C. law makes criminals of parents who consent to their children's participation during curfew hours in a wide range of social, educational, and recreational activities-non-criminal activities that some parents (however few or many) consider fundamental to their children's growth and well-being. See D.C. Code § 6-2183(a)(2), (d) (providing for enforcement and criminal penalties).

Thus, not only do I disagree that "[t]he curfew's defenses allow the parents almost total discretion over their children's activities during curfew hours," Silberman Op. at 22, I think the curfew squarely implicates the well-established "liberty of parents and guardians to direct the upbringing and education of children under their control." Pierce v. Society of Sisters, 268 U.S. 510, 534-35 (1925). As the Supreme Court stated in Wisconsin v. Yoder, "The history and culture of Western civilization reflect a strong tradition of parental concern for the nurture and upbringing of their children. This primary role of the parents in the upbringing of their children is now established beyond debate as an enduring American tradition." 406 U.S. 205, 232 (1972). See Meyer v. Nebraska, 262 U.S. 390, 399 (1923); see also Bellotti v. Baird, 443 U.S. 622, 639 n.18 (1979) (opinion of Powell, J.); Stanley v. Illinois, 405 U.S. 645, 651 (1972).

State interference with this long-recognized parental right to raise children demands strict judicial scrutiny. It is in the context of family, in addition to school and other societal institutions, that children of this diverse and democratic nation begin to develop habits of responsibility necessary for self-governance and to observe not only the formal rules established by government but also the informal rules and understandings that undergird civil society. Through parents, children first learn to relate conduct to consequences, to exercise freedom with responsibility, and to respect the views of others. Ms. Dean's and Mr. Jablon's affidavits describe precisely that process: They are attempting to teach their children in the way they think best, granting them more freedom when they demonstrate responsibility. As Justice Powell said, "[t]his affirmative process of teaching, guiding, and inspiring by precept and example is essential to the growth of young people into mature, socially responsible citizens." Bellotti, 443 U.S. at 638 (opinion of Powell, J.). The parenting process described by Justice Powell-the very process that the curfew curtails for these plaintiffs-is likewise essential, in my view, to equipping young people with the confidence they need to resist the many destructive influences of society. Schools and other governmental institutions, to be sure, are indispensable to this learning process. Parents, however, retain a critical role because "[w]e have believed in this country that this process, in large part, is beyond the competence of impersonal political institutions." Id.

Heightened constitutional protection for parental autonomy is required for another reason. In Yoder, the Supreme Court's unqualified characterization of parents' "primary role" in child-rearing as "an enduring American tradition" reflected its recognition that " '[t]he fundamental theory of liberty upon which all governments in this Union repose excludes any general power of the State to standardize its children....' " 406 U.S. at 232-33 (quoting Pierce, 268 U.S. at 535). Indeed, we refuse to regard "[t]he child [as] the mere creature of the state," Pierce, 268 U.S. at 535, because insistence on a particular theory of parenting, like "affirmative sponsorship of particular ethical, religious, or political beliefs[,] is something we expect the State not to attempt in a society constitutionally committed to the ideal of individual liberty and freedom of choice," Bellotti, 443 U.S. at 638 (opinion of Powell, J.). Of course, this does not mean that Ms. Dean's and Mr. Jablon's authority to raise their children is impervious to state regulation. It does mean that to be valid, limitations on parental rights not only must seek to achieve compelling objectives (which the D.C. juvenile curfew does), but also must demonstrate a close fit-substantiated by record evidence-between means and ends (which the curfew does not).

Relying on Prince v. Massachusetts, 321 U.S. 158 (1944), the District argues that parental authority in child-rearing implicates no fundamental right and is subject to reasonable regulation. To be sure, Prince, which was decided before the modern three-tier framework for reviewing equal protection and fundamental rights claims, said that "the state has a wide range of power for limiting parental freedom and authority in things affecting the child's welfare." Id. at 167. But as I read Prince, it stands not for the broad proposition that reasonable state regulations may override parental judgments on matters of child welfare, but for the now-settled principle that religious practices may be circumscribed by reasonable, neutral laws of general applicability.

Prince sustained the conviction of a Jehovah's Witness under a Massachusetts child labor law for allowing her nineyear-old niece to distribute religious magazines on the street. Characterizing the issue before the Court, Prince's opening paragraph states: "The case brings for review another episode in the conflict between Jehovah's Witnesses and state authority. This time Sarah Prince appeals from convictions for violating Massachusetts' child labor laws, by acts said to be a rightful exercise of her religious convictions." Id. at 159. In other words, Prince claimed a right to allow her niece, not to ply the trade of a newsgirl or magazine seller, but to proselytize, to engage in "the public proclaiming of religion." Id. at 170; see id. at 164 (stating that Prince claimed "the parent's [liberty] to bring up the child in the way he should go, which for appellant means to teach him the tenets and the practices of their faith") (emphasis added). Although Prince's niece offered her magazines for "5per copy," thus technically bringing her conduct under the child labor law, the Court observed that she "received no money" on the evening the offenses occurred, id. at 162, and that while "specified small sums are generally asked and received[,] ... the publications may be had without the payment if so desired," id. at 161 n.4. As suggested by the decision's analogy between child labor and compulsory vaccination laws, see id. at 166, Prince is thus neither a case about "child labor" nor a vindication of state power to trump parental authority, but a case limiting free exercise of religion in the face of otherwise valid state regulation.

Confirming this view, the Supreme Court recently situated Prince in the line of cases establishing that "the right of free exercise does not relieve an individual of the obligation to comply with a 'valid and neutral law of general applicability on the ground that the law proscribes (or prescribes) conduct that his religion prescribes (or proscribes).' " Employment Div., Dep't of Human Resources v. Smith, 494 U.S. 872, 879 (1990) (quoting United States v. Lee, 455 U.S. 252, 263 n.3 (1982) (Stevens, J., Concurring in the judgment)); see id. at 880 (characterizing Prince as finding "no constitutional infirmity in 'excluding [these children] from doing there what no other children may do' ") (quoting Prince, 321 U.S. at 171). Explaining the different approach taken in Wisconsin v. Yoder, where the Court demanded "more than merely a 'reasonable relation to some purpose within the competency of the State' " in holding compulsory school attendance laws inapplicable to Amish parents who refused to send their children to school, 406 U.S. at 233, quoted in Smith, 494 U.S. at 881 n.1, Smith said that Yoder implicated not only free exercise but also "the right of parents, acknowledged in Pierce ..., to direct the education of their children," id. at 881. Smith thus makes clear that a square assertion of parental rights elevates the standard of review applicable to a free exercise claim otherwise subject to rational basis scrutiny. See id. ("The only decisions in which we have held that the First Amendment bars application of a neutral, generally applicable law to religiously motivated action have involved not the Free Exercise Clause alone, but the Free Exercise Clause in conjunction with other constitutional protections, such as ... [parental rights]...."). In light of Smith, I am unconvinced by the District's reliance on Prince for the proposition that rational basis review applies to parental rights claims. Smith leaves no doubt that if the child labor law in Prince, like the compulsory school attendance law in Yoder, had genuinely implicated a parental right distinct from the right of free exercise, then some form of heightened scrutiny should have applied. See Smith, 494 U.S. at 881; accord City of Boerne v. Flores, 117 S. Ct. 2157, 2161 (1997).

In sum, the inquiry triggered by plaintiffs' claim of a fundamental right is not whether the curfew on the whole helps or hinders parental control-that is a policy question for D.C. lawmakers, not federal Judges-but rather whether the District has provided sufficient justification for imposing the particular restrictions on parental control to which these plaintiffs object. On this question, I stand by my view that although the District's goal of reducing crime by and against juveniles is important enough to justify restrictions on parental liberty under either strict or intermediate scrutiny, the method it chose so plainly lacks an evidentiary link to the stated goal that it fails the tailoring prong of both strict and intermediate scrutiny. See Hutchins, 144 F.3d at 826-27 (Tatel, J., Concurring in the judgment); Rogers Op. at 22-31. I respectfully Dissent