announced the judgment of the Court and delivered the opinion of the Court with respect to Part III, an opinion with respect to Parts I, II, and V, in which Justice Stevens, Justice O’Connor, and Justice Sou-TER join, and an opinion with respect to Parts IV and VI, in which Justice Stevens and Justice Souter join.
These cases present First Amendment challenges to three statutory provisions that seek to regulate the broadcasting of “patently offensive” sex-related material on cable television. Cable Television Consumer Protection and Competition Act of 1992 (1992 Act or Act), 106 Stat. 1486, §§ 10(a), 10(b), and 10(c), 47 U. S. C. §§ 532(h), 532(j), and note following § 531. The provisions apply to programs broadcast over cable on what are known as “leased access channels” and “public, educational, or governmental channels.” Two of the provisions essentially permit a cable system operator to prohibit the broadcasting of “programming” that the “operator reasonably believes describes or depicts sexual or excretory activities or organs in a patently offensive manner.” 1992 *733Act, § 10(a); see § 10(c). See also In re Implementation of Section 10 of the Cable Consumer Protection and Competition Act of 1992: Indecent Programming and Other Types of Materials on Cable Access Channels, First Report and Order, 8 FCC Red 998 (1993) (First Report and Order); In re Implementation of Section 10 of the Cable Consumer Protection and Competition Act of1992, Indecent Programming and Other Types of Materials on Cable Access Channels, Second Report and Order, 8 FCC Red 2638 (1993) (Second Report and Order). The remaining provision requires cable system operators to segregate certain “patently offensive” programming, to place it on a single channel, and to block that channel from viewer access unless the viewer requests access in advance and in writing. 1992 Act, § 10(b); 47 CFR § 76.701(g) (1995).
We conclude that the first provision — which permits the operator to decide whether or not to broadcast such programs on leased access channels — is consistent with the First Amendment. The second provision, which requires leased channel operators to segregate and to block that programming, and the third provision, applicable to public, educational, and governmental channels, violate the First Amendment, for they are not appropriately tailored to achieve the basic, legitimate objective of protecting children from exposure to “patently offensive” material.
I
Cable operators typically own a physical cable network used to convey programming over several dozen cable channels into subscribers’ houses. Program sources vary from channel to channel. Most channels carry programming produced by independent firms, including “many national and regional cable programming networks that have emerged in recent years,” Turner Broadcasting System, Inc. v. FCC, 512 U. S. 622, 629 (1994), as well as some programming that the system operator itself (or an operator affili*734ate) may provide. Other channels may simply retransmit through cable the signals of over-the-air broadcast stations. Ibid. Certain special channels here at issue, called “leased channels” and “public, educational, or governmental channels,” carry programs provided by those to whom the law gives special cable system access rights.
A “leased channel” is a channel that federal law requires a cable system operator to reserve for commercial lease by unaffiliated third parties. About 10 to 15 percent of a cable system’s channels would typically fall into this category. See 47 U. S. C. § 532(b). “[P]ublic, educational, or governmental channels” (which we shall call “public access” channels) are channels that, over the years, local governments have required cable system operators to set aside for public, educational, or governmental purposes as part of the consideration an operator gives in return for permission to install cables under city streets and to use public rights-of-way. See § 531; see also H. R. Rep. No. 98-934, p. 30 (1984) (authorizing local authorities to require creation of public access channels). Between 1984 and 1992, federal law (as had much pre-1984 state law, in respect to public access channels) prohibited cable system operators from exercising any editorial control over the content of any program broadcast over either leased or public access channels. See 47 U. S. C. §§ 531(e) (public access), 532(c)(2) (leased access).
In 1992, in an effort to control sexually explicit programming conveyed over access channels, Congress enacted the three provisions before us. The first two provisions relate to leased channels. The first says:
“This subsection shall permit a cable operator to enforce prospectively a written and published policy of prohibiting programming that the cable operator reasonably believes describes or depicts sexual or excretory activities or organs in a patently offensive manner as measured by contemporary community standards.” 1992 Act, § 10(a)(2), 106 Stat. 1486.
*735The second provision, applicable only to leased channels, requires cable operators to segregate and to block similar programming if they decide to permit, rather than to prohibit, its broadcast. The provision tells the Federal Communications Commission (FCC or Commission) to promulgate regulations that will (a) require “programmers to inform cable operators if the program[ming] would be indecent as defined by Commission regulations”; (b) require “cable operators to place” such material “on a single channel”; and (e) require “cable operators to block such single channel unless the subscriber requests access to such channel in writing.” 1992 Act, § 10(b)(1). The Commission issued regulations defining the material at issue in terms virtually identical to those we have already set forth, namely, as descriptions or depictions of “sexual or excretory activities or organs in a patently offensive manner” as measured by the cable viewing community. First Report and Order ¶ ¶ 33-38, at 1003-1004. The regulations require the cable operators to place this material on a single channel and to block it (say, by scrambling). They also require the system operator to provide access to the blocked channel “within 30 days” of a subscriber’s written request for access and to reblock it within 30 days of a subscriber’s request to do so. 47 CFR § 76.701(c) (1995).
The third provision is similar to the first provision, but applies only to public access channels. The relevant statutory section instructs the FCC to promulgate regulations that will
“enable a cable operator of a cable system to prohibit the use, on such system, of any channel capacity of any public, educational, or governmental access facility for any programming which contains obscene material, sexually explicit conduct, or material soliciting or promoting unlawful conduct.” 1992 Act, § 10(c), 106 Stat. 1486.
*736The FCC, carrying out this statutory instruction, promulgated regulations defining “sexually explicit” in language almost identical to that in the statute’s leased channel provision, namely, as descriptions or depictions of “sexual or excretory activities or organs in a patently offensive manner” as measured by the cable viewing community. See 47 CFR §76.702 (1995) (incorporating definition from § 76.701(g)).
The upshot is, as we said at the beginning, that the federal law before us (the statute as implemented through regulations) now permits cable operators either to allow or to forbid the transmission of “patently offensive” sex-related materials over both leased and public access channels, and requires those operators, at a minimum, to segregate and to block transmission of that same material on leased channels.
Petitioners, claiming that the three statutory provisions, as implemented by the Commission regulations, violate the First Amendment, sought judicial review of the Commission’s First Report and Order and its Second Report and Order in the United States Court of Appeals for the District of Columbia Circuit. A panel of that Circuit agreed with petitioners that the provisions violated the First Amendment. Alliance for Community Media v. FCC, 10 F. 3d 812 (1993). The entire Court of Appeals, however, heard the case en banc and reached the opposite conclusion. It held that all three statutory provisions (as implemented) were consistent with the First Amendment. Alliance for Community Media v. FCC, 56 F. 3d 105 (1995). Four of the eleven en banc appeals court judges dissented. Two of the dissenting judges concluded that all three provisions violated the First Amendment. Two others thought that either one, or two, but not all three of the provisions, violated the First Amendment. We granted certiorari to review the en banc court’s First Amendment determinations.
*737II
We turn initially to the provision that 'permits cable system operators to prohibit “patently offensive” (or “indecent”) programming transmitted over leased access channels. 1992 Act, § 10(a). The Court of Appeals held that this provision did not violate the First Amendment because the First Amendment prohibits only “Congress” (and, through the Fourteenth Amendment, a “State”), not private individuals, from “abridging the freedom of speech.” Although the court said that it found no “state action,” 56 F. 3d, at 113, it could not have meant that phrase literally, for, of course, petitioners attack (as “abridging] . . . speech”) a congressional statute — which, by definition, is an Act of “Congress.” More likely, the court viewed this statute’s “permissive” provisions as not themselves restricting speech, but, rather, as simply reaffirming the authority to pick and choose programming that a private entity, say, a private broadcaster, would have had in the absence of intervention by any federal, or local, governmental entity.
We recognize that the First Amendment, the terms of which apply to governmental action, ordinarily does not itself throw into constitutional doubt the decisions of private citizens to permit, or to restrict, speech — and this is so ordinarily even where those decisions take place within the framework of a regulatory regime such as broadcasting. Were that not so, courts might have to face the difficult, and potentially restrictive, practical task of deciding which, among any number of private parties involved in providing a program (for example, networks, station owners, program editors, and program producers), is the “speaker” whose rights may not be abridged, and who is the speech-restricting “censor.” Furthermore, as this Court has held, the editorial function itself is an aspect of “speech,” see Turner, 512 U. S., at 636, and a court’s decision that a private party, say, the station owner, is a “censor,” could itself inter*738fere with that private “censor’s” freedom to speak as an editor. Thus, not surprisingly, this Court’s First Amendment broadcasting cases have dealt with governmental efforts to restrict, not governmental efforts to provide or to maintain, a broadcaster’s freedom to pick and to choose programming. Columbia Broadcasting System, Inc. v. Democratic National Committee, 412 U. S. 94 (1973) (striking restrictions on broadcaster’s ability to refuse to carry political advertising); Red Lion Broadcasting Co. v. FCC, 395 U. S. 367 (1969) (upholding restrictions on editorial authority); FCC v. League of Women Voters of Cal., 468 U. S. 364 (1984) (striking restrictions); cf. Consolidated Edison Co. of N. Y. v. Public Serv. Comm’n of N. Y., 447 U. S. 530 (1980) (striking ban on political speech by public utility using its billing envelopes as a broadcast medium); Central Hudson Gas & Elec. Corp. v. Public Serv. Comm’n of N. Y, 447 U. S. 557 (1980) (striking restriction on public utility advertising).
Nonetheless, petitioners, while conceding that this is ordinarily so, point to circumstances that, in their view, make the analogy with private broadcasters inapposite and make these cases special ones, warranting a different constitutional result. As a practical matter, they say, cable system operators have considerably more power to “censor” program viewing than do broadcasters, for individual communities typically have only one cable system, linking broadcasters and other program providers with each community’s many subscribers. See Turner, supra, at 633 (only one cable system in most communities; nationally more than 60% of homes subscribe to cable, which then becomes the primary or sole source of video programming in the overwhelming majority of these homes). Moreover, concern about system operators’ exercise of this considerable power originally led government— local and federal — to insist that operators provide leased and public access channels free of operator editorial control. H. R. Rep. No. 98-934, at 30-31. To permit system operators to supervise programming on leased access channels will *739create the very private-censorship risk that this anticensorship effort sought to avoid. At the same time, petitioners add, cable systems have two relevant special characteristics. They are unusually involved with government, for they depend upon government permission and government facilities (streets, rights-of-way) to string the cable necessary for their services. And in respect to leased channels, their speech interests are relatively weak because they act less like editors, such as newspapers or television broadcasters, than like common carriers, such as telephone companies.
Under these circumstances, petitioners conclude, Congress’ “permissive” law, in actuality, will “abridge” their free speech. And this Court should treat that law as a con-gressionally imposed, content-based, restriction unredeemed as a properly tailored effort to serve a “compelling interest.” See Simon & Schuster, Inc. v. Members of N. Y. State Crime Victims Bd., 502 U. S. 105, 118 (1991); Sable Communications of Cal., Inc. v. FCC, 492 U. S. 115, 126 (1989). They further analogize the provisions to constitutionally forbidden content-based restrictions upon speech taking place in “public forums” such as public streets, parks, or buildings dedicated to open speech and communication. See Cornelius v. NAACP Legal Defense & Ed. Fund, Inc., 473 U. S. 788, 802 (1985); Perry Ed. Assn. v. Perry Local Educators’ Assn., 460 U. S. 37, 45 (1983); see also H. R. Rep. No. 98-934, supra, at 30 (identifying public access channels as the electronic equivalent of a “speaker’s soap box”). And, finally, petitioners say that the legal standard the law contains (the “patently offensive” standard) is unconstitutionally vague. See, e. g., Interstate Circuit, Inc. v. Dallas, 390 U. S. 676 (1968) (rejecting censorship ordinance as vague, even though it was intended to protect children).
Like petitioners, Justices Kennedy and Thomas would have us decide these cases simply by transferring and applying literally categorical standards this Court has developed in other contexts. For Justice Kennedy, leased access *740channels are like a common carrier, cablecast is a protected medium, strict scrutiny applies, § 10(a) fails this test, and, therefore, § 10(a) is invalid. Post, at 796-801,806-807. For Justice Thomas, the case is simple because the cable operator who owns the system over which access channels are broadcast, like a bookstore owner with respect to what it displays on the shelves, has a predominant First Amendment interest. Post, at 816-817, 822-824. Both categorical approaches suffer from the same flaws: They import law developed in very different contexts into a new and changing environment, and they lack the flexibility necessary to allow government to respond to very serious practical problems without sacrificing the free exchange of ideas the First Amendment is designed to protect.
The history of this Court’s First Amendment jurisprudence, however, is one of continual development, as the Constitution’s general command that “Congress shall make no law . . . abridging the freedom of speech, or of the press,” has been applied to new circumstances requiring different adaptations of prior principles and precedents. The essence of that protection is that Congress may not regulate speech except in cases of extraordinary need and with the exercise of a degree of care that we have not elsewhere required. See, e. g., Schenck v. United States, 249 U. S. 47, 51-52 (1919); Abrams v. United States, 250 U. S. 616, 627-628 (1919) (Holmes, J., dissenting); West Virginia Bd. of Ed. v. Barnette, 319 U. S. 624, 639 (1943); Texas v. Johnson, 491 U. S. 397, 418-420 (1989). At the same time, our cases have not left Congress or the States powerless to address the most serious problems. See, e. g., Chaplinsky v. New Hampshire, 315 U. S. 568 (1942); Young v. American Mini Theatres, Inc., 427 U. S. 50 (1976); FCC v. Pacifica Foundation, 438 U. S. 726 (1978).
Over the years, this Court has restated and refined these basic First Amendment principles, adopting them more particularly to the balance of competing interests and the special *741circumstances of each field of application. See, e. g., New York Times Co. v. Sullivan, 376 U. S. 254 (1964) (allowing criticism of public officials to be regulated by civil libel only if the plaintiff shows actual malice); Gertz v. Robert Welch, Inc., 418 U. S. 323 (1974) (allowing greater regulation of speech harming individuals who are not public officials, but still requiring a negligence standard); Red Lion Broadcasting Co. v. FCC, 395 U. S. 367 (1969) (employing highly flexible standard in response to the scarcity problem unique to over-the-air broadcast); Arkansas Writers’ Project, Inc. v. Ragland, 481 U. S. 221, 231-232 (1987) (requiring “compelling state interest” and a “narrowly drawn” means in context of differential taxation of media); Sable, supra, at 126, 131 (applying “compelling interest,” “least restrictive means,” and “narrowly tailored” requirements to indecent telephone communications); Turner, 512 U. S., at 641 (using “heightened scrutiny” to address content-neutral regulations of cable system broadcasts); Central Hudson Gas & Elec. Corp., 447 U. S., at 566 (restriction on commercial speech cannot be “more extensive than is necessary” to serve a “substantial” government interest).
This tradition teaches that the First Amendment embodies an overarching commitment to protect speech from government regulation through close judicial scrutiny, thereby enforcing the Constitution’s constraints, but without imposing judicial formulas so rigid that they become a strait jacket that disables government from responding to serious problems. This Court, in different contexts, has consistently held that government may directly regulate speech to address extraordinary problems, where its regulations are appropriately tailored to resolve those problems without imposing an unnecessarily great restriction on speech. Justices Kennedy and Thomas would have us further declare which, among the many applications of the general approach that this Court has developed over the years, we are applying here. But no definitive choice among competing *742analogies (broadcast, common carrier, bookstore) allows us to declare a rigid single standard, good for now and for all future media and purposes. That is not to say that we reject all the more specific formulations of the standard — they appropriately cover the vast majority of cases involving government regulation of speech. Rather, aware as we are of the changes taking place in the law, the technology, and the industrial structure related to telecommunications, see, e. g., Telecommunications Act of 1996, 110 Stat. 56; S. Rep. No. 104-23 (1995); H. R. Rep. No. 104-204 (1995), we believe it unwise and unnecessary definitively to pick one analogy or one specific set of words now. See Columbia Broadcasting, 412 U. S., at 102 (“The problems of regulation are rendered more difficult because the broadcast industry is dynamic in terms of technological change; solutions adequate a decade ago are not necessarily so now, and those acceptable today may well be outmoded 10 years hence”); Pacifica, supra, at 748 (“We have long recognized that each medium of expression presents special First Amendment problems”). We therefore think it premature to answer the broad questions that Justices Kennedy and Thomas raise in their efforts to find a definitive analogy, deciding, for example, the extent to which private property can be designated a public forum, compare post, at 791-793, 794 (Kennedy, J., concurring in part, concurring in judgment in part, and dissenting in part), with post, at 826-829 (Thomas, J., concurring in judgment in part and dissenting in part); whether public access channels are a public forum, post, at 791-792 (opinion of Kennedy, J.); whether the Government’s viewpoint neutral decision to limit a public forum is subject to the same scrutiny as a selective exclusion from a pre-existing public forum, post, at 799-803 (opinion of Kennedy, J.); whether exclusion from common carriage must for all purposes be treated like exclusion from a public forum, post, at 797-798 (opinion of Kennedy, J.); and whether the interests of the owners of communica*743tions media always subordinate the interests of all other users of a medium, post, at 816-817 (opinion of Thomas, J.).
Rather than decide these issues, we can decide these cases more narrowly, by closely scrutinizing § 10(a) to assure that it properly addresses an extremely important problem, without imposing, in light of the relevant interests, an unnecessarily great restriction on speech. The importance of the interest at stake here — protecting children from exposure to patently offensive depictions of sex; the accommodation of the interests of programmers in maintaining access channels and of cable operators in editing the contents of their channels; the similarity of the problem and its solution to those at issue in Pacifica; and the flexibility inherent in an approach that permits private cable operators to make editorial decisions, lead us to conclude that § 10(a) is a sufficiently tailored response to an extraordinarily important problem.
First, the provision before us comes accompanied with an extremely important justification, one that this Court has often found compelling — the need to protect children from exposure to patently offensive sex-related material. Sable Communications, 492 U. S., at 126; Ginsberg v. New York, 390 U. S. 629, 639-640 (1968); New York v. Ferber, 458 U. S. 747, 756-757 (1982).
Second, the provision arises in a very particular context— congressional permission for cable operators to regulate programming that, but for a previous Act of Congress, would have had no path of access to cable channels free of an operator’s control. The First Amendment interests involved are therefore complex, and require a balance between those interests served by the access requirements themselves (increasing the availability of avenues of expression to programmers who otherwise would not have them), H. R. Rep. No. 98-934, at 31-36, and the disadvantage to the First Amendment interests of cable operators and other programmers (those to whom the cable operator would have assigned *744the channels devoted to access). See Turner, 512 U. S., at 635-637.
Third, the problem Congress addressed here is remarkably similar to the problem addressed by the FCC in Pacifica, and the balance Congress struck is commensurate with the balance we approved there. In Pacifica this Court considered a governmental ban of a radio broadcast of “indecent” materials, defined in part, like the provisions before us, to include
“ ‘language that describes, in terms patently offensive as measured by contemporary community standards for the broadcast medium, sexual or excretory activities and organs, at times of the day when there is a reasonable risk that children may be in the audience.’ ” 438 U. S., at 732 (quoting 56 F. C. C. 2d 94, 98 (1975)).
The Court found this ban constitutionally permissible primarily because “broadcasting is uniquely accessible to children” and children were likely listeners to the program there at issue — an afternoon radio broadcast. 438 U. S., at 749-750. In addition, the Court wrote, “the broadcast media have established a uniquely pervasive presence in the lives of all Americans,” id., at 748, “[p]atently offensive, indecent material... confronts the citizen, not only in public, but also in the privacy of the home,” generally without sufficient prior warning to allow the recipient to avert his or her eyes or ears, ibid.; and “[a]dults who feel the need may purchase tapes and records or go to theaters and nightclubs” to hear similar performances, id., at 750, n. 28.
All these factors are present here. Cable television broadcasting, including access channel broadcasting, is as “accessible to children” as over-the-air broadcasting, if not more so. See Heeter, Greenberg, Baldwin, Paugh, Srig-ley, & Atkin, Parental Influences on Viewing Style, in Cable-viewing 140 (C. Heeter & B. Greenberg eds. 1988) (children spend more time watching television and view more channels *745than do their parents, whether their household subscribes to cable or receives television over the air). Cable television systems, including access channels, “have established a uniquely pervasive presence in the lives of all Americans.” Pacifica, supra, at 748. See Jost, The Future of Television, 4 The CQ Researcher 1131, 1146 (Dec. 23, 1994) (63% of American homes subscribe to cable); Greenberg, Heeter, D’Alessio, & Sipes, Cable and Noncable Viewing Style Comparisons, in Cableviewing, supra, at 207 (cable households spend more of their day, on average, watching television, and will watch more channels, than households without cable service). “Patently offensive” material from these stations can “confron[t] the citizen” in the “privacy of the home,” Pa-cifica, supra, at 748, with little or no prior warning. Cable-viewing, supra, at 217-218 (while cable subscribers tend to use guides more than do broadcast viewers, there was no difference among these groups in the amount of viewing that was planned, and, in fact, cable subscribers tended to sample more channels before settling on a program, thereby making them more, not less, susceptible to random exposure to unwanted materials). There is nothing to stop “adults who feel the need” from finding similar programming elsewhere, say, on tape or in theaters. In fact, the power of cable systems to control home program viewing is not absolute. Over-the-air broadcasting and direct broadcast satellites already provide alternative ways for programmers to reach the home and are likely to do so to a greater extent in the near future. See generally Telecommunications Act of 1996, §201, 110 Stat. 107 (advanced television services), §205 (direct broadcast satellite), § 302 (video programming by telephone companies), and §304 (availability of navigation devices to enhance multichannel programming); L. Johnson, Toward Competition in Cable Television (1994).
Fourth, the permissive nature of § 10(a) means that it likely restricts speech less than, not more than, the ban at issue in Pacifica. The provision removes a restriction as to *746some speakers — namely, cable operators. See supra, at 743. Moreover, although the provision does create a risk that a program will not appear, that risk is not the same as the certainty that accompanies a governmental ban. In fact, a glance at the programming that cable operators allow on their own (nonaccess) channels suggests that this distinction is not theoretical, but real. See App. 393 (regular channel broadcast of Playboy and “Real Sex” programming). Finally, the provision’s permissive nature brings with it a flexibility that allows cable operators, for example, not to ban broadcasts, but, say, to rearrange broadcast times, better to fit the desires of adult audiences while lessening the risks of harm to children. See First Report and Order ¶ 31, at 1003 (interpreting the Act’s provisions to allow cable operators broad discretion over what to do with offensive materials). In all these respects, the permissive nature of the approach taken by Congress renders this measure appropriate as a means of achieving the underlying purpose of protecting children.
Of course, cable system operators may not always rearrange or reschedule patently offensive programming. Sometimes, as petitioners fear, they may ban the programming instead. But the same may be said of Pacifica’s ban. In practice, the FCC’s daytime broadcast ban could have become a total ban, depending upon how private operators (programmers, station owners, networks) responded to it. They would have had to decide whether to reschedule the daytime show for nighttime broadcast in light of comparative audience demand and a host of other practical factors that similarly would determine the practical outcomes of the provisions before us. The upshot, in both cases, must be uncertainty as to practical consequences — of the governmental ban in the one case and of the permission in the other. That common uncertainty makes it difficult to say the provision here is, in any respect, more restrictive than the order in *747Pacifica. At the same time, in the respects we discussed, the provision is significantly less restrictive.
The existence of this complex balance of interests persuades us that the permissive nature of the provision, coupled with its viewpoint-neutral application, is a constitutionally permissible way to protect children from the type of sexual material that concerned Congress, while accommodating both the First Amendment interests served by the access requirements and those served in restoring to cable operators a degree of the editorial control that Congress removed in 1984.
Our basic disagreement with Justice Kennedy is narrow. Like him, we believe that we must scrutinize § 10(a) with the greatest care. Like Justices Kennedy and Thomas, we believe that the interest of protecting children that § 10(a) purports to serve is compelling. But we part company with Justice Kennedy on two issues. First, Justice Kennedy’s focus on categorical analysis forces him to disregard the cable system operators’ interests. Post, at 805-806. We, on the other hand, recognize that in the context of cable broadcast that involves an access requirement (here, its partial removal), and unlike in most cases where we have explicitly required “narrow tailoring,” the expressive interests of cable operators do play a legitimate role. Cf. Turner, 512 U. S., at 636-637. While we cannot agree with Justice Thomas that everything turns on the rights of the cable owner, see post, at 823-824, we also cannot agree with Justice Kennedy that we must ignore the expressive interests of cable operators altogether. Second, Justice Kennedy’s application of a very strict “narrow tailoring” test depends upon an analogy with a category (“the public forum cases”), which has been distilled over time from the similarities of many cases. Rather than seeking an analogy to a category of cases, however, we have looked to the cases themselves. And, as we have said, we found that Pacifica provides the *748closest analogy and lends considerable support to our conclusion.
Petitioners and Justice Kennedy, see post, at 797-798, 803-804, argue that the opposite result is required by two other cases: Sable Communications of Cal., Inc. v. FCC, 492 U. S. 115 (1989), a case in which this Court found unconstitutional a statute that banned “indecent” telephone messages, and Turner, in which this Court stated that cable broadcast receives full First Amendment protection. See 512 U. S., at 637-641. The ban at issue in Sable, however, was not only a total governmentally imposed ban on a category of communications, but also involved a communications medium, telephone service, that was significantly less likely to expose children to the banned material, was less intrusive, and allowed for significantly more control over what comes into the home than either broadcasting or the cable transmission system before us. See 492 U. S., at 128. The Court’s distinction in Turner, furthermore, between cable and broadcast television, relied on the inapplicability of the spectrum scarcity problem to cable. See 512 U. S., at 637-641. While that distinction was relevant in Turner to the justification for structural regulations at issue there (the “must carry” rules), it has little to do with a case that involves the effects of television viewing on children. Those effects are the result of how parents and children view television programming, and how pervasive and intrusive that programming is. In that respect, cable and broadcast television differ little, if at all. See supra, at 744-745. Justice Kennedy would have us decide that all common carriage exclusions are subject to the highest scrutiny, see post, at 796-799, and then decide these cases on the basis of categories that provide imprecise analogies rather than on the basis of a more contextual assessment, consistent with our First Amendment tradition, of assessing whether Congress carefully and appropriately addressed a serious problem.
*749Petitioners also rely on this Court’s “public forum” cases. They point to Perry Ed. Assn. v. Perry Local Educators’ Assn., 460 U. S., at 45, a case in which this Court said that “public forums” are “places” that the government “has opened for use by the public as a place for expressive activity,” or which “by long tradition . . . have been devoted to assembly and debate.” Ibid. See also Cornelius v. NAACP Legal Defense & Ed. Fund, Inc., 473 U. S., at 801 (assuming public forums may include “private property dedicated to public use”). They add that the Government cannot “enforce a content-based exclusion” from a public forum unless “necessary to serve a compelling state interest” and “narrowly drawn.” Perry, supra, at 45. They further argue that the statute’s permissive provisions unjustifiably exclude material, on the basis of content, from the “public forum” that the Government has created in the form of access channels. Justice Kennedy adds by analogy that the decision to exclude certain content from common carriage is similarly subject to strict scrutiny, and here does not satisfy that standard of review. See post, at 796-799, 805-807.
For three reasons, however, it is unnecessary, indeed, unwise, for us definitively to decide whether or how to apply the public forum doctrine to leased access channels. First, while it may be that content-based exclusions from the right to use common carriers could violate the First Amendment, see post, at 796-800 (opinion of Kennedy, J.), it is not at all clear that the public forum doctrine should be imported wholesale into the area of common carriage regulation. As discussed above, we are wary of the notion that a partial analogy in one context, for which we have developed doctrines, can compel a full range of decisions in such a new and changing area. See supra, at 739-743. Second, it is plain from this Court’s cases that a public forum “may be created for a limited purpose.” Perry, supra, at 46, n. 7; see also Cornelius, supra, at 802 (“[T]he government ‘is not required to indefinitely retain the open character of the facility’”)
*750(quoting Perry, supra, at 46). Our cases have not yet determined, however, that government’s decision to dedicate a public forum to one type of content or another is necessarily subject, to the highest level of scrutiny. Must a local government, for example, show a compelling state interest if it builds a band shell in the park and dedicates it solely to classical music (but not to jazz)? The answer is not obvious. Cf. Perry, supra, at 46, n. 7. But, at a minimum, these cases do not require us to answer it. Finally, and most important, the effects of Congress’ decision on the interests of programmers, viewers, cable operators, and children are the same, whether we characterize Congress’ decision as one that limits access to a public forum, discriminates in common carriage, or constrains speech because of its content. If we consider this particular limitation of indecent television programming acceptable as a constraint on speech, we must no less accept the limitation it places on access to the claimed public forum or on use of a common carrier.
Consequently, if one wishes to view the permissive provisions before us through a “public forum” lens, one should view those provisions as limiting the otherwise totally open nature of the forum that leased access channels provide for communication of other than patently offensive sexual material — taking account of the fact that the limitation was imposed in light of experience gained from maintaining a totally open “forum.” One must still ask whether the First Amendment forbids the limitation. But unless a label alone were to make a critical First Amendment difference (and we think here it does not), the features of these cases that we have already discussed — the Government’s interest in protecting children, the “permissive” aspect of the statute, and the nature of the medium — sufficiently justify the “limitation” on the availability of this forum.
Finally, petitioners argue that the definition of the materials subject to the challenged provisions is too vague, thereby granting cable system operators too broad a program-*751screening authority. Cf. Hoffman Estates v. Flipside, Hoffman Estates, Inc., 455 U. S. 489, 498 (1982) (citing Grayned v. City of Rockford, 408 U. S. 104, 108-109 (1972)) (vague laws may lead to arbitrary enforcement); Dombrowski v. Pfister, 380 U. S. 479, 486-487 (1965) (uncertainty may perniciously chill speech). That definition, however, uses language similar to language previously used by this Court for roughly similar purposes.
The provisions, as augmented by FCC regulations, permit cable system operators to prohibit
“programming that the cable operator reasonably believes describes or depicts sexual or excretory activities or organs in a patently offensive manner as measured by contemporary community standards.” 1992 Act, § 10(a), 106 Stat. 1486.
See also 47 CFR § 76.702 (1995) (reading approximately the same definition into § 10(c)). This language is similar to language adopted by this Court in Miller v. California, 413 U. S. 15, 24 (1973), as a “guideline]” for identifying materials that States may constitutionally regulate as obscene. In Miller, the Court defined obscene sexual material (material that lacks First Amendment protection) in terms of
“(a) whether the average person, applying contemporary community standards would find that the work, taken as a whole, appeals to the prurient interest . . . ; (b) whether the work depicts or describes, in a patently offensive way, sexual conduct specifically defined by the applicable state law; and (c) whether the work, taken as a whole, lacks serious literary, artistic, political, or scientific value.” Ibid, (emphasis added; internal quotation marks omitted).
The language, while vague, attempts to identify the category of materials that Justice Stewart thought could be described only in terms of “I know it when I see it.” Jacobellis v. Ohio, 378 U. S. 184, 197 (1964) (concurring opinion). In *752§ 10(a) and the FCC regulations, without Miller’s qualifiers, the language would seem to refer to material that would be offensive enough to fall within that category but for the fact that the material also has “serious literary, artistic, political or scientific value” or nonprurient purposes.
This history suggests that the statute’s language aims at the kind of programming to which its sponsors referred— pictures of oral sex, bestiality, and rape, see 138 Cong. Rec. 981, 985 (1992) (statement of Sen. Helms) — and not at scientific or educational programs (at least unless done with a highly unusual lack of concern for viewer reaction). Moreover, as this Court pointed out in Pacifica, what is “patently offensive” depends on context (the kind of program on which it appears), degree (not “an occasional expletive”), and time of broadcast (a “pig” is offensive in “the parlor” but not the “barnyard”). 438 U. S., at 748, 750. Programming at 2 o’clock in the morning is seen by a basically adult audience and the “patently offensive” must be defined with that fact in mind.
Further, the statute protects against overly broad application of its standards insofar as it permits cable system operators to screen programs only pursuant to a “written and published policy.” 1992 Act, § 10(a), 106 Stat. 1486. A cable system operator would find it difficult to show that a leased access program prohibition reflects a rational “policy” if the operator permits similarly “offensive” programming to run elsewhere on its system at comparable times or in comparable ways. We concede that the statute’s protection against overly broad application is somewhat diminished by the fact that it permits a cable operator to ban programming that the operator “reasonably believes” is patently offensive. Ibid. (emphasis added). But the “reasonable] belie[f]” qualifier here, as elsewhere in the law, seems designed not to expand the category at which the law aims, but, rather, to provide a legal excuse, for (at least) one honest mistake, from liability that might otherwise attach. Cf. Waters v. Churchill, 511 *753U. S. 661, 682 (1994) (Souter, J., concurring) (public employer’s reasonable belief that employee engaged in unprotected speech excuses liability); United States v. United States Gypsum Co., 438 U. S. 422, 453-455, and n. 29 (1978) (“‘meeting competition’ ” defense in antitrust based on reasonable belief in the necessity to meet competition); Pierson v. Ray, 386 U. S. 547, 555-557 (1967) (police officer has defense to constitutional claim, as did officers of the peace at common law in actions for false arrest, when the officer reasonably believed the statute whose violation precipitated the arrest was valid). And the contours of the shield — reasonableness— constrain the discretion of the cable operator as much as they protect it. If, for example, a court had already found substantially similar programming to be beyond the pale of “patently offensive” material, or if a local authority overseeing the local public, governmental, or educational channels had indicated that materials of the type that the cable operator decides to ban were not “patently offensive” in that community, then the cable operator would be hard pressed to claim that the exclusion of the material was “reasonable.” We conclude that the statute is not impermissibly vague.
For the reasons discussed, we conclude that § 10(a) is consistent with the First Amendment.
III
The statute s second provision significantly differs from the first, for it does not simply permit, but rather requires, cable system operators to restrict speech — by segregating and blocking “patently offensive” sex-related material appearing on leased channels (but not on other channels). 1992 Act, § 10(b). In particular, as previously mentioned, see supra, at 735, this provision and its implementing regulations require cable system operators to place “patently offensive” leased channel programming on a separate channel; to block that channel; to unblock the channel within 30 days of a subscriber’s written request for access; and to *754reblock the channel within 30 days of a subscriber’s request for reblocking. 1992 Act, § 10(b); 47 CFR §§ 76.701(b), (c), (g) (1995). Also, leased channel programmers must notify cable operators of an intended “patently offensive” broadcast up to 30 days before its scheduled broadcast date. §§ 76.701(d), (g).
These requirements have obvious restrictive effects. The several up-to-30-day delays, along with single channel segregation, mean that a subscriber cannot decide to watch a single program without considerable advance planning and without letting the “patently offensive” channel in its entirety invade his household for days, perhaps weeks, at a time. These restrictions will prevent programmers from broadcasting to viewers who select programs day by day (or, through “surfing,” minute by minute); to viewers who would like occasionally to watch a few, but not many, of the programs on the “patently offensive” channel; and to viewers who simply tend to judge a program’s value through channel reputation, i. e., by the company it keeps. Moreover, the “written notice” requirement will further restrict viewing by subscribers who fear for their reputations should the operator, advertently or inadvertently, disclose the list of those who wish to watch the “patently offensive” channel. Cf. Lamont v. Postmaster General, 381 U. S. 301, 307 (1965) (finding unconstitutional a requirement that recipients of Communist literature notify the Post Office that they wish to receive it). Further, the added costs and burdens that these requirements impose upon a cable system operator may encourage that operator to ban programming that the operator would otherwise permit to run, even if only late at night.
The Government argues that, despite these adverse consequences, the “segregate and block” requirements are lawful because they are “the least restrictive means of realizing” a “‘compelling interest,’” namely, ‘“protecting the physical and psychological well-being of minors.’” See Brief for Federal Respondents 11 (quoting Sable, 492 U. S., at 126). *755It adds that, in any event, the First Amendment, as applied in Pacifica, “does not require that regulations of indecency on television be subject to the strictest” First Amendment “standard of review.” Brief for Federal Respondents 11.
We agree with the Government that protection of children is a “compelling interest.” See supra, at 743. But we do not agree that the “segregate and block” requirements properly accommodate the speech restrictions they impose and the legitimate objective they seek to attain. Nor need we here determine whether, or the extent to which, Pacifica does, or does not, impose some lesser standard of review where indecent speech is at issue, compare 438 U. S., at 745-748 (opinion of Stevens, J.) (indecent materials enjoy lesser First Amendment protection), with id., at 761-762 (Powell, J., concurring in part and concurring in judgment) (refusing to accept a lesser standard for nonobscene, indecent material). That is because once one examines this governmental restriction, it becomes apparent that, not only is it not a “least restrictive alternative” and is not “narrowly tailored” to meet its legitimate objective, it also seems considerably “more extensive than necessary.” That is to say, it fails to satisfy this Court’s formulations of the First Amendment’s “strictest,” as well as its somewhat less “strict,” requirements. See, e. g., Sable, 492 U. S., at 126 (“compelling interest” and “least restrictive means” requirements applied to indecent telephone communications); id., at 131 (requiring “narrowly tailored” law); Turner, 512 U. S., at 641 (using “heightened scrutiny” to address content-neutral structural regulations of cable systems); id., at 662 (quoting “‘no greater than . . . essential’ ” language from United States v. O’Brien, 391 U. S. 367, 377 (1968), as an example of “heightened,” less-than-strictest, First Amendment scrutiny); Central Hudson, 447 U. S., at 566 (restriction on commercial speech cannot be “more extensive than is necessary”); Florida Bar v. Went For It, Inc., 515 U. S. 618, 624 (1995) (restriction must be “narrowly drawn”); id., at 632 (there must be a *756“reasonable” “fit” with the objective that legitimates speech restriction). The provision before us does not reveal the caution and care that the standards underlying these various verbal formulas impose upon laws that seek to reconcile the critically important interest in protecting free speech with very important, or even compelling, interests that sometimes warrant restrictions.
Several circumstances lead us to this conclusion. For one thing, the law, as recently amended, uses other means to protect children from similar “patently offensive” material broadcast on mleased cable channels, i. e., broadcast over any of a system’s numerous ordinary, or public access, channels. The law, as recently amended, requires cable operators to “scramble or . . . block” such programming on any (unleased) channel “primarily dedicated to sexually-oriented programming.” Telecommunications Act of 1996, §505, 110 Stat. 136 (emphasis added). In addition, cable operators must honor a subscriber’s request to block any, or all, programs on any channel to which he or she does not wish to subscribe. §504, ibid. And manufacturers, in the future, will have to make television sets with a so-called “V-chip”— a device that will be able automatically to identify and block sexually explicit or violent programs. § 551, id., at 139-142.
Although we cannot, and do not, decide whether the new provisions are themselves lawful (a matter not before us), we note that they are significantly less restrictive than the provision here at issue. They do not force the viewer to receive (for days or weeks at a time) all “patently offensive” programming or none; they will not lead the viewer automatically to judge the few by the reputation of the many; and they will not automatically place the occasional viewer’s name on a special list. They therefore inevitably lead us to ask why, if they adequately protect children from “patently offensive” material broadcast on ordinary channels, they would not offer adequate protection from similar leased channel broadcasts as well? Alternatively, if these provi*757sions do not adequately protect children from “patently offensive” material broadcast on ordinary channels, how could one justify more severe leased channel restrictions when (given ordinary channel programming) they would yield so little additional protection for children?
The record does not answer these questions. It does not explain why, under the new Act, blocking alone — without written access requests — adequately protects children from exposure to regular sex-dedicated channels, but cannot adequately protect those children from programming on similarly sex-dedicated channels that are leased. It does not explain why a simple subscriber blocking request system, perhaps a phone-call-based system, would adequately protect children from “patently offensive” material broadcast on ordinary non-sex-dedicated channels (i. e., almost all channels) but a far more restrictive segregate/block/written-access system is needed to protect children from similar broadcasts on what (in the absence of the segregation requirement) would be non-sex-dedicated channels that are leased. Nor is there any indication Congress thought the new ordinary channel protections less than adequate.
The answers to the questions are not obvious. We have no empirical reason to believe, for example, that sex-dedicated channels are all (or mostly) leased channels, or that “patently offensive” programming on non-sex-dedicated channels is found only (or mostly) on leased channels. To the contrary, the parties’ briefs (and major city television guides) provide examples of what seems likely to be such programming broadcast over both kinds of channels.
We recognize, as the Government properly points out, that Congress need not deal with every problem at once. Cf. Semler v. Oregon Bd. of Dental Examiners, 294 U. S. 608, 610 (1935) (the legislature need not “strike at all evils at the same time”); and Congress also must have a degree of leeway in tailoring means to ends. Columbia Broadcasting, 412 U. S., at 102-103. But in light of the 1996 statute, it seems *758fair to say that Congress now has tried to deal with most of the problem. At this point, we can take Congress’ different, and significantly less restrictive, treatment of a highly similar problem at least as some indication that more restrictive means are not “essential” (or will not prove very helpful). Cf. Boos v. Barry, 485 U. S. 312, 329 (1988) (existence of a less restrictive statute suggested that a challenged ordinance, aimed at the same problem, was overly restrictive).
The record’s description and discussion of a different alternative — the “lockbox” — leads, through a different route, to a similar conclusion. The Cable Communications Policy Act of 1984 required cable operators to provide
“upon the request of a subscriber, a device by which the subscriber can prohibit viewing of a particular cable service during periods selected by the subscriber.” 47 U.S. C. § 544(d)(2).
This device — the “lockbox” — would help protect children by permitting their parents to “lock out” those programs or channels that they did not want their children to see. See FCC 85-179, ¶ 132, 50 Fed. Reg. 18637, 18655 (1985) (“[T]he provision for lockboxes largely disposes of issues involving the Commission’s standard for indecency”). The FCC, in upholding the “segregate and block” provisions, said that lockboxes protected children (including, say, children with inattentive parents) less effectively than those provisions. See First Report and Order ¶ ¶ 14-15, 8 FCC Red, at 1000. But it is important to understand why that is so.
The Government sets forth the reasons as follows:
“In the case of lockboxes, parents would have to discover that such devices exist; find out that their cable operators offer them for sale; spend the time and money to buy one; learn how to program the lockbox to block undesired programs; and, finally, exercise sufficient vigilance to ensure that they have, indeed, locked out *759whatever indecent programming they do not wish their children to view.” Brief for Federal Respondents 37.
We assume the accuracy of this statement. But the reasons do not show need for a provision as restrictive as the one before us. Rather, they suggest a set of provisions very much like those that Congress placed in the 1996 Act.
No provision, we concede, short of an absolute ban, can offer certain protection against assault by a determined child. We have not, however, generally allowed this fact alone to justify “ ‘ “reducing] the adult population . . . to .. . only what is fit for children.”’” Sable, 492 U. S., at 128 (quoting Bolger v. Youngs Drug Products Corp., 463 U. S. 60, 73 (1983), in turn quoting Butler v. Michigan, 352 U. S. 380, 383 (1957)); see Sable, supra, at 130, and n. 10. But, leaving that problem aside, the Government’s list of practical difficulties would seem to call, not for “segregate and block” requirements, but, rather, for informational requirements, for a simple coding system, for readily available blocking equipment (perhaps accessible by telephone), for imposing cost burdens upon system operators (who may spread them through subscription fees); or perhaps even for a system that requires lockbox defaults to be set to block certain channels (say, sex-dedicated channels). These kinds of requirements resemble those that Congress has recently imposed upon all but leased channels. For that reason, the “lockbox” description and the discussion of its frailties reinforces our conclusion that the leased channel provision is overly restrictive when measured against the benefits it is likely to achieve. (We add that the record’s discussion of the “lockbox” does not explain why the law now treats leased channels more restrictively than ordinary channels.)
There may, of course, be other explanations. Congress may simply not have bothered to change the leased channel provisions when it introduced a new system for other channels. But responses of this sort, like guesses about the comparative seriousness of the problem, are hot legally adequate. *760In other cases, where,.as here, the record before Congress or before an agency provides no convincing explanation, this Court has not been willing to stretch the limits of the plausible, to create hypothetical nonobvious explanations in order to justify laws that impose significant restrictions upon speech. See, e. g., Sable, supra, at 130 (“[T]he congressional record presented to us contains no evidence as to how effective or ineffective the FCC’s most recent regulations were or might prove to be”); Simon & Schuster, 502 U. S., at 120; Minneapolis Star & Tribune Co. v. Minnesota Comm’r of Revenue, 460 U. S. 575, 585-586 (1983); Arkansas Writers’ Project, 481 U. S., at 231-232.
Consequently, we cannot find that the “segregate and block” restrictions on speech are a narrowly, or reasonably, tailored effort to protect children. Rather, they are overly restrictive, “sacrificing]” important First Amendment interests for too “speculative a gain.” Columbia Broadcasting, 412 U. S., at 127; see League of Women Voters, 468 U. S., at 397. For that reason they are not consistent with the First Amendment.
IV
The statute’s third provision, as implemented by FCC regulation, is similar to its first provision, in that it too permits a cable operator to prevent transmission of “patently offensive” programming, in this case on public access channels. 1992 Act, § 10(c); 47 CFR § 76.702 (1995). But there are four important differences.
The first is the historical background. As Justice Kennedy points out, see post, at 788-790, cable operators have traditionally agreed to reserve channel capacity for public, governmental, and educational channels as part of the consideration they give municipalities that award them cable franchises. See H. R; Rep. No. 98-934, at 30. In the terms preferred by Justice Thomas, see post, at 827-828, the requirement to reserve capacity for public access channels is similar to the reservation of a public easement, or a dedica*761tion of land for streets and parks, as part of a municipality’s approval of a subdivision of land. Cf. post, at 793-794 (opinion of Kennedy, J.). Significantly, these are channels over which cable operators have not historically exercised editorial control. H. R. Rep. No. 98-934, supra, at 30. Unlike § 10(a) therefore, § 10(c) does not restore to cable operators editorial rights that they once had, and the countervailing First Amendment interest is nonexistent, or at least much diminished. See also post, at 792-793 (opinion of Kennedy, J.).
The second difference is the institutional background that has developed as a result of the historical difference. When a “leased channel” is made available by the operator to a private lessee, the lessee has total control of programming during the leased time slot. See 47 U. S. C. § 532(c)(2). Public access channels, on the other hand, are normally subject to complex supervisory systems of various sorts, often with both public and private elements. See § 531(b) (franchising authorities “may require rules and procedures for the use of the [public access] channel capacity”). Municipalities generally provide in their cable franchising agreements for an access channel manager, who is most commonly a nonprofit organization, but may also be the municipality, or, in some instances, the cable system owner. See D. Brenner, M. Price, & M. Myerson, Cable Television and Other Non-broadcast Video ¶ 6.04[7] (1993); P. Aufderheide, Public Access Cable Programming, Controversial Speech, and Free Expression (1992) (hereinafter Aufderheide), reprinted in App. 61, 63 (surveying 61 communities; the access manager was: a nonprofit organization in 41, a local government official in 12, the cable operator in 5, and an unidentified entity in 3); D. Agosta, C. Rogoff, & A. Norman, The Participate Report: A Case Study of Public Access Cable Television in New York State 28 (1990) (hereinafter Agosta), attached as Exh. K to Joint Comments for the Alliance for Community Media et al., filed with the FCC under MM Docket No. 92-*762258 (materials so filed hereinafter FCC Record) (“In 88% [of New York public access systems] access channels were programmed jointly between the cable operator and another institution such as a university, library, or non-profit access organization”); id., at 28-32, FCC Record; Comments of National Cable Television Association Inc., at 14, FCC Record (“Operators often have no involvement in PEG channels that are run by local access organizations”). Access channel activity and management are partly financed with public funds — through franchise fees or other payments pursuant to the franchise agreement, or from general municipal funds, see Brenner, Price, & Myerson, supra, ¶ 6.04[3][c]; Aufder-heide, App. 59-60 — and are commonly subject to supervision by a local supervisory board. See, e. g., D. C. Code Ann. § 43-1829 (1990 and Supp. 1996); Lynchburg City Code § 12.1-44(d)(2) (1988).
This system of public, private, and mixed nonprofit elements, through its supervising boards and nonprofit or governmental access managers, can set programming policy and approve or disapprove particular programming services. And this system can police that policy by, for example,, requiring indemnification by programmers, certification of compliance with local standards, time segregation, adult content advisories, or even by prescreening individual programs. See Second Report and Order ¶ 26, 8 FCC Red, at 2642 (“[F]rom the comments received, it appears that a number of access organizations already have in place procedures that require certification statements [of compliance with local standards], or their equivalent, from access programmers”); Comments of Boston Community Access and Programming Foundation, App. 163-164; Aufderheide, id., at 69-71; Comments of Metropolitan Area Communications Commission 2, FCC Record; Reply Comments of Waycross Community Television 4-6, FCC Record; Reply Comments of Columbus Community Cable Access, Inc., App. 329; Reply Comments of City of St. Paul, id., at 318, 325; Reply Comments of Erik *763Mollberg, Public Access Coordinator, Ft. Wayne, Ind., 3, FCC Record; Comments of Defiance Community Television 3, FCC Record; Comments of Nutmeg Public Access Television, Inc., 3-4, FCC Record. Whether these locally accountable bodies prescreen programming, promulgate rules for the use of public access channels, or are merely available to respond when problems arise, the upshot is the same: There is a locally accountable body capable of addressing the problem, should it arise, of patently offensive programming broadcast to children, making it unlikely that many children will in fact be exposed to programming considered patently offensive in that community. See 56 F. 3d, at 127-128; Second Report and Order ¶ 26, 8 FCC Red 2642.
Third, the existence of a system aimed at encouraging and securing programming that the community considers valuable strongly suggests that a “cable operator’s veto” is less likely necessary to achieve the statute’s basic objective, protecting children, than a similar veto in the context of leased channels. Of course, the system of access managers and supervising boards can make mistakes, which the operator might in some cases correct with its veto power. Balanced against this potential benefit, however, is the risk that the veto itself may be mistaken; and its use, or threatened use, could prevent the presentation of programming, that, though borderline, is not “patently offensive” to its targeted audience. See Aufderheide, App. 64-66 (describing the programs that were considered borderline by access managers, including sex education, health education, broadcasts of politically marginal groups, and various artistic experiments). And this latter threat must bulk large within a system that already has publicly accountable systems for maintaining responsible programs.
Finally, our examination of the legislative history and the record before us is consistent with what common sense suggests, namely, that the public/nonprofit programming control systems now in place would normally avoid, minimize, or *764eliminate any child-related problems concerning “patently offensive” programming. We have found anecdotal references to what seem isolated instances of potentially indecent programming, some of which may well have occurred on leased, not public access, channels. See 138 Cong. Rec. 984, 990 (1992) (statement of Sen. Wirth) (mentioning “abuses” on Time Warner’s New York City channel); but see Comments of Manhattan Neighborhood Network, App. 235, 238 (New York access manager noting that leased, not public access, channels regularly carry sexually explicit programming in New York, and that no commercial programs or advertising are allowed on public access channels); Brief for Time Warner Cable as Amicus Curiae 2-3 (indicating that relevant “abuses” likely occurred on leased channels). See also 138 Cong. Rec., at 989 (statement of Sen. Fowler) (describing solicitation of prostitution); id., at 985 (statement of Sen. Helms) (identifying newspaper headline referring to mayor’s protest of a “strip act”); 56 F. 3d, at 117-118 (recounting comments submitted to the FCC describing three complaints of offensive programming); Letter from Mayor of Rancho Palos Verdes, FCC Record; Resolution of San Antonio City Council, No. 92-49-40, FCC Record.
But these few examples do not necessarily indicate a significant nationwide pattern. See 56 F. 3d, at 127-128 (public access channels “did not pose dangers on the order of magnitude of those identified on leased access channels,” and “local franchising authorities could respond” to such problems “by issuing Tules and procedures’ or other ‘requirements’”). The Commission itself did not report any examples of “indecent” programs on public access channels. See Second Report and Order, 8 FCC Red, at 2638; see also Comments of Boston Community Access and Programming Foundation, App. 162-163 (noting that the FCC’s Notice of Proposed Rulemaking, 7 FCC Red 7709 (1992), did not identify any “inappropriate” programming that actually exists on public *765access channels). Moreover, comments submitted to the FCC undermine any suggestion that prior to 1992 there were significant problems of indecent programming on public access channels. See Agosta 10,28, FCC Record (surveying 76 public access systems in New York over two years, and finding “only two examples of controversial programming, and both had been settled by the producers and the access channel”); Reply Comments of Staten Island Community Television 2, FCC Record (“Our access channels have been on the air since 1986 without a single incident which would be covered by Section 10 of the new law”); Reply Comments of Waycross Community Television, at 2, FCC Record (“[I]n-decent and obscene programs ... [have] never been cablecast through Waycross Community Television during our entire ten year programming history”); Reply Comments of Cambridge Community Television, App. 314 (“In Cambridge less than one hour out of 15,000 hours of programming CCTV has run in the past five year[s] may have been affected by the Act”); ibid. (“CCTV feels that there simply is not a problem which needs to be fixed”); Reply Comments of Columbus Community Cable Access, Inc., id., at 329 (“ACTV is unaware of any actions taken by the cable operators under [a local law authorizing them to prohibit “legally obscene matter”] within the last 10 years”); Reply Comments of Cincinnati Community Video, Inc., id., at 316 (“[I]n 10 years of access operations with over 30,000 access programs cablecast not a single obscenity violation has ever occurred”); Comments of Defiance Community Television, at 2-3, FCC Record (in eight years of operation, “there has never been a serious problem with the content of programming on the channel”).
At most, we have found borderline examples as to which people’s judgment may differ, perhaps acceptable in some communities but not others, of the type that petitioners fear the law might prohibit. See, e. g., Aufderheide, App. 64-66; Brief for Petitioners in No. 95-124, p. 7 (describing depiction *766of a self-help gynecological examination); Comments of Time Warner Entertainment Co., App. 252 (describing an Austin, Tex., program that included “nude scenes from a movie,” and an Indianapolis, Ind., “ ‘safe sex’ ” program). It is difficult to see how such borderline examples could show a compelling need, nationally, to protect children from significantly harmful materials. Compare 138 Cong. Rec., at 985 (statement of Sen. Helms) (justifying regulation of leased access channels in terms of programming that depicts “bestiality” and “rape”). In the absence of a factual basis substantiating the harm and the efficacy of its proposed cure, we cannot assume that the harm exists or that the regulation redresses it. See Turner, 512 U. S., at 664-665.
The upshot, in respect to the public access channels, is a law that could radically change present programming-related relationships among local community and nonprofit supervising boards and access managers, which relationships are established through municipal law, regulation, and contract. In doing so, it would not significantly restore editorial rights of cable operators, but would greatly increase the risk that certain categories of programming (say, borderline offensive programs) will not appear. At the same time, given present supervisory mechanisms, the need for this particular provision, aimed directly at public access channels, is not obvious. Having carefully reviewed the legislative history of the Act, the proceedings before the FCC, the record below, and the submissions of the parties and amici here, we conclude that the Government cannot sustain its burden of showing that § 10(c) is necessary to protect children or that it is appropriately tailored to secure that end. See, e.g., Columbia Broadcasting, 412 U. S., at 127; League of Women Voters, 468 U. S., at 398-399; Sable, 492 U. S., at 126. Consequently, we find that this third provision violates the First Amendment.
*767V
Finally, we must ask whether § 10(a) is severable from the two other provisions. The question is one of legislative intent: Would Congress still “have passed” § 10(a) “had it known” that the remaining “provision^ were] invalid”? Brockett v. Spokane Arcades, Inc., 472 U. S. 491, 506 (1985). If so, we need not invalidate all three provisions' New York v. Ferber, 458 U. S., at 769, n. 24 (citing United States v. Thirty-seven Photographs, 402 U. S. 363 (1971)).
Although the 1992 Act contains no express “severability clause,” we can find the Act’s “severability” intention in its structure and purpose. It seems fairly obvious Congress would have intended its permissive “leased access” channels provision, § 10(a), to stand irrespective of § 10(c)’s legal fate. That is because the latter provision concerns only public, educational, and governmental channels. Its presence had little, if any, effect upon “leased access” channels; hence its absence in respect to those channels could not make a significant difference.
The “segregate and block” requirement’s invalidity does make a difference, however, to the effectiveness of the permissive “leased access” provision, § 10(a). Together they told the cable system operator: “Either ban a ‘patently offensive’ program or ‘segregate and block’ it.” Without the “segregate and block” provision, cable operators are afforded broad discretion over what to do with a patently offensive program, and because they will no longer bear the costs of segregation and blocking if they refuse to ban such programs, cable operators may choose to ban fewer programs.
Nonetheless, this difference does not make the two provisions unseverable. Without the “segregate and block” provision, the law simply treats leased channels (in respect to patently offensive programming) just as it treats all other channels. And judging by the absence of similar segregate and block provisions in the context of these other channels, Congress would probably have thought that § 10(a), standing *768alone, was an effective (though, perhaps, not the most effective) means of pursuing its objective. Moreover, we can find no reason why, in light of Congress’ basic objective (the protection of children), Congress would have preferred no provisions at all to the permissive provision standing by itself. That provision, capable of functioning on its own, still helps to achieve that basic objective. Consequently, we believe the valid provision is severable from the others.
VI
For these reasons, the judgment of the Court of Appeals is affirmed insofar as it upheld § 10(a); the judgment of the Court of Appeals is reversed insofar as it upheld § 10(b) and § 10(c).
It is so ordered.