Concept
The Intermediate Scrutiny Test
Content-neutral regulation of protected expression
The Intermediate Scrutiny Test
Once a court classifies a speech regulation as content-neutral, it reviews the regulation under intermediate scrutiny. Turner Broadcasting System, Inc. v. FCC stated the standard for the must-carry rules that burdened cable operators: the government must show that the regulation advances important interests unrelated to the suppression of speech and does not burden substantially more speech than necessary.1 Earlier sections explained how a court decides which track a law travels; this section explains what happens on the second track. That track governs city ordinances and state rules as well as federal statutes, because the Due Process Clause of the Fourteenth Amendment has protected expression against state infringement since 1925, even though the First Amendment's text speaks only to Congress.2
The formulation makes two demands. The first concerns the government's reason: the interest must be important, and it must be unrelated to the suppression of expression, which carries forward the justification inquiry taught earlier. The second concerns fit: the regulation must not burden substantially more speech than necessary to advance that interest. Together the demands ask whether the government has a genuine reason that does not depend on the message, and whether the law costs speech no more than that reason requires.
The Court states this test in several vocabularies: one for regulations of expressive conduct, drawn from O'Brien's ban on destroying draft cards; one for time, place, and manner rules, drawn from Ward; and Turner's own wording. A student who treats each label as a separate test memorizes three lists where one analysis would do, and the next subsection shows why the labels describe a single standard. That standard is lighter than strict scrutiny because a law not aimed at ideas threatens public debate less, so the fit may be imperfect and the government need not exhaust every alternative. It is still heightened scrutiny, not minimum rationality review.3 The government needs a real interest, a real connection between that interest and the burden, and some showing that the problem exists.
The first two steps apply to every content-neutral regulation, and a neutral label guarantees nothing, as the leaflet ban and the sign ban discussed below show. The third step belongs to rules about where, when, and how expression occurs, and it protects the speaker's ability to reach an audience some other way. The fourth concerns proof, and Turner answered it across two decisions, one before and one after the government built a fuller record.
Two Formulations, One Test
Turner Broadcasting System, Inc. v. FCC offers the clearest evidence that the Court treats its two familiar formulations as one standard, because it chose the test for must-carry by citing both. The Court held that the governing standard is the intermediate scrutiny applicable to content-neutral restrictions that impose an incidental burden on speech, and it drew that standard from Ward v. Rock Against Racism, a concert-volume case from a public park, and from United States v. O'Brien, a prosecution for burning a draft card. It then quoted O'Brien: a content-neutral regulation is sustained “if it furthers an important or substantial governmental interest; if the governmental interest is unrelated to the suppression of free expression; and if the incidental restriction on alleged First Amendment freedoms is no greater than is essential to the furtherance of that interest.”4 The checklist above borrowed vocabulary from both lines without comment, and this subsection shows why that was legitimate.
O'Brien burned his draft card and was convicted under a 1965 amendment to the Selective Service Act that made knowing destruction of a card a crime.5 Chief Justice Warren accepted that the First Amendment was implicated, and he reasoned that when speech and nonspeech elements combine in one course of conduct, a sufficiently important governmental interest in regulating the nonspeech element can justify incidental limits on expressive freedom. That reasoning explains the shape of the test. The government may regulate the burning itself, so the question is whether the regulation reaches expression only incidentally. The Court announced four parts, each of which must be satisfied:
- The regulation is within the constitutional power of the government.
- The regulation furthers an important or substantial governmental interest.
- The governmental interest is unrelated to the suppression of free expression.
- The incidental restriction on First Amendment freedoms is no greater than is essential to the furtherance of that interest.6
Applying those parts, the Court found the 1965 amendment designed to serve the legitimate and substantial interest in the maintenance and integrity of the selective service system, and it sustained the conviction.7 The statute's breadth matters to that result. It barred every knowing destruction of a card, not merely contemptuous destruction or destruction meant to hinder the war effort, and a leaf-burning ordinance reasons the same way. A general ban on leaf burning clears the first, second, and fourth parts easily, and a court reviewing a conviction under it need not decide whether a particular burning was communicative. An ordinance reaching only burning by socialists, or only burning done contemptuously or to protest the city's policies, fails the third part, because the city's interest is then tied to suppression.8
The public-forum cases ask the same questions in different words. A time, place, and manner regulation must be justified without reference to the content of the speech, must be narrowly tailored to advance a significant governmental interest, and must leave open ample alternative channels for communicating the information.9 The vocabularies line up closely. The content-neutral justification matches O'Brien's requirement of an interest unrelated to suppression, the significant interest matches its important-or-substantial-interest part, and narrow tailoring matches its requirement that the incidental restriction be no greater than essential. Only the alternative-channels clause lacks an obvious twin, which is why it can look like an additional requirement.
Clark v. Community for Creative Non-Violence closed that apparent gap. The Court upheld a ban on camping in the parks of the capital as applied to protesters demonstrating for the homeless, and it analyzed the ban under both formulations, noting that the two tests run in parallel.10 Both framings produced one answer. The ban was content-neutral and was not applied out of disagreement with the protesters' message. It left the demonstration intact, with its symbolic city, its signs, and its day-and-night vigil, and it focused narrowly on the government's substantial interest in keeping the parks attractive and intact.11 One set of facts thus answered every question in both vocabularies. The alternative-channels clause did work inside that analysis too, because nothing barred the protesters from delivering their message to the media or the public by other means.
The merger disposes of a common error, the belief that O'Brien is a special, weaker standard for symbolic speech. It is the time, place, and manner test worded slightly differently, so both apply mid-level scrutiny when the regulation is not directed at the content of the speech, and neither applies when it is.12 O'Brien could be punished because the punishment rested on how he conveyed his message. A protester who spray-painted “Stop the draft” on a courthouse wall would have had no better claim against a property-damage charge, since the government's focus would again be the manner of expression. A government that punished either protester out of opposition to the message could not defend the punishment by pointing to the symbolic form.
The third part decides whether the test is available at all. A regulation unrelated to the message receives the intermediate standard; one that targets a harm associated with the message is content-based, the test does not apply, and strict scrutiny governs.13 An obviously content-based inquiry does not escape strict scrutiny because some event, such as an election, is involved, and a government gains nothing by describing its law as a regulation of conduct.14
Armbands and draft cards show the difference. In Tinker v. Des Moines Independent Community School District, public school students punished for wearing black armbands to protest the Vietnam War won on the strict-scrutiny track, while O'Brien, who burned a draft card to express the same disapproval of the same policies, lost under the lower standard.15 The difference lay in the government's reason. The armband suspension rested on the message, and the draft-card law rested on the administration of the draft. In Holder v. Humanitarian Law Project, the Court likewise refused to apply O'Brien to the federal material-support statute as applied to plaintiffs who wished to speak to designated terrorist groups, because whether they could speak depended on what they said. It reasoned from Cohen v. California, where a generally applicable breach-of-the-peace law was applied to a jacket bearing an epithet because of what the jacket communicated.16 The flag-burning case discussed in the next subsection ends outside the test for the same reason.
The formulation sounds more demanding than it has proved. On its face the O'Brien test appears highly protective of speech, but at the Supreme Court level it has not protected expressive conduct, because the Court gave its parts a highly deferential reading.17 O'Brien accepted the selective-service interest on the allegation that draft-card burning would interfere with administering the system, even though federal law already required possession of a draft card.18 The Court applied the same inquiry in a public-forum setting in Members of City Council of Los Angeles v. Taxpayers for Vincent, sustaining a ban on posting political posters on public property.19 The fourth part, in particular, does not demand the least speech-restrictive means, because narrow tailoring is satisfied when the means chosen do not burden substantially more speech than necessary.20
A problem on this material therefore begins with classification. If the law is content-neutral, one intermediate standard applies, and you can state it in either vocabulary. If the government's interest is related to the suppression of expression, the test never comes into play.
A Substantial Interest Unrelated to Suppression
The first component asks what the government is trying to accomplish, and Turner supplies the model answer. Congress declared that the must-carry provisions serve three interrelated interests: preserving the benefits of free, over-the-air local broadcast television, promoting the widespread dissemination of information from a multiplicity of sources, and promoting fair competition in the market for television programming.22 The Court held that none of these interests is related to the suppression of free expression or to the content of any speaker's message, and that each, viewed in the abstract, is an important governmental interest.23
The component has two parts, and each guards against a different danger. The demand for an important or substantial interest keeps the government from burdening speech for trivial reasons, because even a neutral rule reduces the expression in circulation and a real burden calls for a real justification. The demand that the interest be unrelated to suppression decides which track the law travels. An interest that arises only from what speech says belongs under strict scrutiny, while an interest that exists whatever is said poses less risk of masking an effort to drive ideas out of public debate.24 A court tells the two apart by asking what causal connection the government asserts between its interest and the speech, because the ultimate interest the government names, such as safety or order, will almost always sound unrelated to expression.
Each of Turner's interests rests on a fact about the television market. The competition interest is always substantial, even when the regulated entities are engaged in expressive activity, so an interest in how a market is structured can support regulating speakers without becoming an interest in what they say.25 Justice Stevens found the access interest self-evident: protecting access to television for the millions of homes without cable, and assuring a multiplicity of sources, were “unquestionably substantial.”26 The First Amendment does not of its own force require access to the electronic media, but a statute or an FCC regulation may mandate some public access, and broadcasting is subject to greater regulation in the public interest than print is.27 The harm Congress feared, households losing free broadcast service, does not depend on what any station broadcasts, and that is what makes the interest unrelated to suppression.
Justice O'Connor's partial dissent shows that the second part is a real battleground. Joined by Justices Scalia and Ginsburg and in part by Justice Thomas, she read Congress's references to the diversity and local-affairs focus of broadcast programming as showing that its preference for broadcasters was justified with reference to content.28 An interest stated in general terms can therefore still turn on content when the government's own findings tie it to what the favored speech says, so a court reads the findings and not just the label.
Public-forum cases supply a more familiar list. Content-neutral regulations are typically defended by the orderly movement of crowds, peace and quiet, or the aesthetic attractiveness of public spaces.29 All the Justices in Metromedia, Inc. v. City of San Diego agreed that aesthetics is a legitimate and substantial interest in judging restraints on speech, and the plurality treated traffic safety as substantial too.30 These interests share one feature: they are easy to find. Turner had no difficulty concluding that each of its three interests was important,31 and in Watchtower Bible and Tract Society of New York, Inc. v. Village of Stratton the Court struck down a permit requirement for door-to-door canvassing because posted “No Solicitation” signs protected residents' privacy with far less burden on speech.32 A government can nearly always name an interest of this kind, so the contest on most facts moves to whether the interest is unrelated to the message, and then to proof and tailoring.
Taxpayers for Vincent shows an interest that clears both parts. The Court held that the city could ban political posters from public property to further its interest in traffic safety and aesthetics, so long as the ordinance did not discriminate between types of speech.33 The interest is unrelated to suppression for the reason the diagnostic gives: a pile of signs on a pole is clutter whatever the signs say.
Justice Brennan's dissent in Vincent explains why the label alone cannot settle the second part. He warned that an aesthetic objective can be a facade for suppressing the messages signs typically carry, and he would have accepted the objective as substantial and unrelated to suppression only if the government showed that it pursues the objective seriously and comprehensively and in ways unrelated to the restriction of speech.34 The majority upheld the ban, but the concern survives in sign regulation generally, because a municipality risks undermining its claim to an aesthetic goal to the extent it grants exemptions.35
The second town has lost the evenhandedness on which the Vincent majority relied, and it has produced the facade the dissent described.36
Texas v. Johnson shows an interest that fails the second part. Texas defended a flag-burning conviction with two interests, preventing breaches of the peace and preserving the flag as a symbol of nationhood and national unity.37 The Court held the first interest not implicated on the record and the second related to the suppression of expression.38 The breach-of-the-peace theory amounted to a claim that an audience that takes serious offense at expression is necessarily likely to disturb the peace, a presumption the Court's precedents do not countenance. The flag interest fails because concern for the flag's symbolism arises only when a person's treatment of the flag communicates some message. An interest tied to the message sends the law to strict scrutiny, which presumes the regulation invalid and demands narrowly tailored or necessary means to achieve a compelling interest, so the intermediate test never comes into play.39
Set beside Vincent, the flag case isolates the question the first component asks. A campaign sign and a burning flag are both communicative acts, and each government offered an interest that sounds respectable. The sign ban answers a harm, accumulated clutter, that would exist whatever the signs said; the flag statute answers a harm that exists only for observers who grasp what the burning expresses. The label the government attaches tells a court little, and the causal path from the speech to the harm tells it nearly everything.
A student applying the first component therefore runs two checks in order. The interest must be one a court will treat as important or substantial, which is usually the easy check, and the path from speech to harm must then run through something other than the message. An interest that passes both earns the government only the chance to be tested on proof and fit.
Narrow Tailoring Without Least Restriction
Turner states the rule for the second component, the fit between a content-neutral regulation and its interest. A regulation need not be the least speech-restrictive means of advancing the government's interests. It is narrowly tailored “so long as the . . . regulation promotes a substantial government interest that would be achieved less effectively absent the regulation,” and it must not “burden substantially more speech than is necessary to further the government's legitimate interests.”41 The formulation comes from Ward and contains two limits that pull in different directions. The first relaxes the demand: a court may not strike a rule down merely by imagining a gentler alternative. The second keeps the demand alive: the rule still fails if it sweeps in substantially more speech than the interest calls for.
The least-restrictive-means requirement belongs to the other track. Strict scrutiny presumes the regulation invalid and asks whether it is a necessary means to a compelling interest, so a case turns on whether any less restrictive alternative would have worked.42 Ward refused to carry that habit into the content-neutral setting. Time, place, or manner regulations must be narrowly tailored, but they have never received strict scrutiny, so the same degree of tailoring is not required of them, and “least-restrictive-alternative analysis is wholly out of place.”43 The tracks differ because they guard against different dangers. A law aimed at a message may disguise an effort to suppress ideas, so the government must prove that nothing gentler would serve. A neutral rule poses less of that danger, and Turner accordingly forbids only a rule that burdens substantially more speech than needed.44
Ward illustrates the relaxed demand. New York City required performers at the Central Park bandshell to use a city sound technician, and the Court sustained the rule. A regulation is not invalid simply because a court concludes that some less speech-restrictive alternative could adequately serve the government's interest, so long as the means chosen are not substantially broader than necessary.45
The second half of the formulation forbids a regulation under which “a substantial portion of the burden on speech does not serve to advance its goals.”46 Schneider v. State supplies the classic example. A city banned all leaflet distribution on its streets to keep them clean. The interest in preventing litter was legitimate, but litter comes from those who drop paper, and a law punishing the litterer would advance the interest effectively while burdening substantially less speech; the ban fell on leafleting that never littered.47 Justice O'Connor's summary in Turner is the shortest statement of the rule: a government that wants to avoid littering may ban littering, but it may not ban all leafleting.48
A sign ban defended by a similar-sounding concern came out the other way, and the contrast shows what the substantial-portion test measures. Leafleting is direct communication that lasts only while the speaker is on the scene, so the litter the city feared is a by-product that a law against littering can reach. Posted signs stay unattended until someone removes them, so the visual blight comes from the medium of expression itself, and a ban on posting signs on public property fits that harm.49 The question is where the harm comes from. If the speech activity itself creates it, banning the activity fits the problem. If it comes from conduct the government can reach directly, a ban on the speech sweeps in activity that never contributes to the harm, and the fit fails.
What separates an irrelevant alternative from a relevant one is the word substantially. Watchtower treats the availability of effective but less restrictive alternatives as one of several fact-driven variables that determine the fit, and it found a permit requirement for door-to-door canvassing insufficiently tailored in part because enforceable “No Solicitation” signs protected residents' privacy with far less burden on speech.50 The Court made the same point in striking down Ladue's sign ban: more temperate measures could in large part satisfy the city's regulatory needs without harm to its citizens' First Amendment rights.51 An obvious alternative that works as well and burdens substantially less speech is evidence that most of the rule's burden does not serve its purpose. It operates as proof of overbreadth, not as a separate requirement, so an exam stem that stresses a gentler method should not be answered with the strict-scrutiny rule that the government must use the least restrictive means.
Turner itself contains the best illustration of the overbreadth argument. Justice O'Connor concluded that the must-carry rules were content-based, and she added that even if she were wrong they failed content-neutral scrutiny, because they restrict too much speech that does not implicate the interests in fair competition and the survival of free television.52 A cable operator's choice of a programmer over a broadcaster may sometimes reflect anticompetitive motives or endanger a broadcaster's survival, but the rules disadvantage cable programmers even where neither is true. That some speech within a broad category causes harm does not justify restricting the whole category, and Congress could protect broadcasters in danger of failing, or bar operators from preferring programmers they own, without restricting programmers in markets where free broadcasting would thrive anyway.53 That is the substantial-portion test applied to a statute, and the Court's response shows how contested the application can be. It placed the burden on the government to show that the remedy does not burden substantially more speech than necessary, found the record insufficient, and remanded; after the remand it sustained the rules five to four.54
Ample Alternative Channels
Turner's two-demand statement of the standard does not speak directly to whether a speaker can still reach an audience some other way. That question belongs to the time, place, and manner formulation, under which a regulation must leave open “ample alternative channels for communication of the information.”56 The first subsection showed that the two formulations describe one inquiry; this clause is the piece with no obvious twin in O'Brien's wording, which is why it can look optional. It is a working component with its own job.
The element asks what remains to the speaker after the regulation takes effect. If a law leaves no time, no place, and no manner in which the information can be communicated, the element fails; beyond that floor there is no set formula, and the general availability of traditional means of communication ordinarily suffices.57 The clause exists because the other elements look only at the government's side of the ledger. A rule can be neutral, serve a real interest, and fit that interest closely, yet leave a speaker no practical way to reach listeners, as when it closes the one medium a speaker of modest means can afford. The clause does not, however, guarantee the speaker's favorite channel. Every rule about where or how people speak takes away some preferred method, and if preference controlled, no such rule could stand.
Heffron v. International Society for Krishna Consciousness shows the element satisfied. A state fair's fixed-location rule limited where literature could be sold or distributed, and the Court upheld it as a reasonable time, place, and manner regulation. A state fair on public property is a limited public forum designed to present the products and views of many exhibitors to a large number of people in a limited area, and the rule applied even-handedly to all distributors, so it was content-neutral.58 The Krishnas lost the method they preferred, and the arrangement may not have been ideal from their perspective, but absent a showing that the alternative channels are illusory or demonstrably inadequate, the factor should be easily satisfied.59
Clark disposed of the element even more quickly. The camping ban closed one way of making the protesters' point, but the regulation left the demonstration intact, and no one suggested any barrier to delivering the message about the plight of the homeless to the media or the public by other means.60 Together the two decisions mark the ordinary case. The government takes away one location or method, speakers can still talk, write, and reach the press, and the element is met without close inquiry into whether the leftover channels equal the lost one.
The instinct that any list of remaining options satisfies the element is the one City of Ladue v. Gilleo corrects. Ladue banned almost all signs on residential property, and a resident who displayed a sign in her window challenged the ordinance. Assuming the ordinance was content-neutral, the Court held that it closed off an important medium of speech and that no adequate substitutes existed; Justice O'Connor, concurring, would have decided first whether the ordinance was content-based.61 The city answered that residents could still use hand-held signs, letters, handbills, flyers, telephone calls, newspaper advertisements, bumper stickers, speeches, and neighborhood or community meetings, and the Court was “not persuaded that adequate substitutes exist for the important medium of speech that Ladue has closed off.”62
The Court's reasons show what the element measures. Displaying a sign from one's own residence often carries a message quite distinct from placing the same sign someplace else or conveying the same text or picture by other means, because the location tells viewers who the speaker is, and residential signs are an unusually cheap and convenient form of communication.63 A list of substitutes therefore counts only to the extent that each can carry the same message to the same audience at a cost the speaker can bear. The Court added that its decision leaves the city free to address the ills of residential signs by more temperate measures.64
Two sign ordinances aimed at visual clutter thus came out opposite ways, and the reason lies in what each one closed. Vincent upheld a ban on posting signs on public property,65 which reached only public property and so left signs on private property available. Ladue's ordinance reached the one place where a sign speaks as the resident's own word. The answer turned on whether the closed channel was unique and whether its effect could be duplicated elsewhere, and not on the number of listed substitutes.
The element is applied with far less force to adult theaters. In City of Renton v. Playtime Theatres, Inc., the theaters proved that the five percent of the city where they could legally locate was not commercially viable, yet the Court found the alternative-channels part satisfied; in the context of this adult expression, the alternative channels did not need to be ample or even available.66 The Court asked only whether the city had effectively denied the owners a reasonable opportunity to open and operate a theater, and owners who must fend for themselves in the real estate market on an equal footing with other buyers have no First Amendment claim.67 The dilution tracks the lesser magnitude the Court assigns to erotic expression compared with political debate, so an answer about ordinary political speech should apply Ladue's search for real substitutes and not Renton's search for a reasonable opportunity.68
Step four does the work in the hard cases. An exam stem that recites a long list of remaining options invites a count, but the count settles nothing until each option is measured against the channel the government closed. A rule that removes a convenient method and leaves the audience reachable, as in the fair and park cases, passes. A rule that removes a medium whose location, form, or price is part of its value, as in the residential-sign case, does not.
Proof and Deference
The evidentiary half of intermediate scrutiny comes from two Turner decisions: Turner I, 512 U.S. 622 (1994), and Turner II, 520 U.S. 180 (1997). Turner I accepted that the interests in preserving local broadcasting and promoting a multiplicity of information sources were important in the abstract, and it held that this did not show that must-carry would in fact advance them.70 A government that defends a speech regulation as a remedy for past harms or a safeguard against anticipated ones must do more than assert that a problem exists. It must demonstrate that the recited harms are real, not merely conjectural, and that the regulation will in fact alleviate them in a direct and material way.71 A regulation perfectly reasonable in the face of a given problem may be highly capricious if the problem does not exist, so an interest's importance gives a court no basis to uphold a rule unless the facts show that the rule is needed.
Turner I turned that demand into two questions. The first asks whether the government has adequately shown that the harm is real, which in Turner meant whether the economic health of local broadcasting was in genuine jeopardy and in need of the protection must-carry afforded. The second arises only if the answer is yes, and it asks whether the remedy does not “burden substantially more speech than is necessary to further the government's legitimate interests.”72 The government bears the burden on each question, so a court does not begin by presuming that a neutral regulation is justified and asking the challenger to explain why it is not.
The Court agreed that courts must accord substantial deference to the predictive judgments of Congress. Sound policymaking often requires legislators to forecast future events and to anticipate their likely impact through deductions and inferences for which complete empirical support may be unavailable, and Congress is not obliged to make a record of the kind an agency or a court makes to accommodate judicial review.73 A question about how an industry will develop with or without a statute calls for judgment about the future, and deference seems especially appealing to the Court where new technologies make the effects hard to foresee.74 Evidence can come from many places: when Renton zoned adult theaters, it relied heavily on the experience and studies of Seattle.75 The Court has been hesitant to make firm rules at the intersection of the First Amendment and cable television, and Turner belongs to a complex series of decisions in which the degree of deference is not easily predictable.76
Deference to a prediction leaves the court's review in place. The legislature supplies the forecast, and the court still asks whether evidence supports it. That division explains Turner I, which reached the Court on summary judgment without findings of fact from the district court. The Court was unable to conclude that the government had shown either that local broadcasting was in genuine jeopardy or that must-carry avoided burdening substantially more speech than necessary.77 A general finding that an industry is endangered told the Court what Congress believed, and it did not show that the belief was well founded or that the remedy fit it. Justice Kennedy, joined on this point only by Chief Justice Rehnquist and Justices Blackmun and Souter, found the record insufficient to show that must-carry was narrowly tailored and concluded that the case should be remanded.78
The remand was one vote short, and the fifth vote came from a Justice who thought no remand was necessary. Justice Stevens would have affirmed the summary judgment for the government, because to him the importance of the interests was apparent: the public interests in protecting access to television for the millions of homes without cable and in assuring a multiplicity of sources were “unquestionably substantial.”79 He added that an industry need not be in its death throes before Congress may act to protect it from economic harm threatened by a monopoly.80 To reach a disposition that commanded the support of a majority of the Court, he nevertheless joined Justice Kennedy's opinion, and his vote supplied the fifth for the remand.81 His difference with the four concerned whether the existing record already proved enough, and it left the standard itself in place. The Justices who dissented in part, convinced that the rules were content-based and called for strict scrutiny, stood on a different premise altogether.82
After further factfinding on remand, the lower court found that substantial evidence supported Congress's prediction that without must-carry a sizeable number of over-the-air broadcast signals would be denied cable carriage. Applying the intermediate standard enunciated in O'Brien, the Supreme Court affirmed, five to four.83 Justice Kennedy held that a court “cannot displace Congress' judgment respecting content-neutral regulations with our own, so long as its policy is grounded on reasonable factual findings supported by evidence that is substantial for a legislative determination.”84 The evidence answered both of Turner I's questions. It showed that the future of free local broadcasting would be in economic jeopardy without must-carry, and it showed that the majority of cable systems had not canceled any programming to satisfy must-carry, so that the burden was congruent to the benefit.85
The two records explain the two outcomes. In Turner I the government held a general congressional finding of economic danger and little to show that local broadcasters were in genuine jeopardy or that the remedy burdened no more speech than needed, and summary judgment for the government could not stand on that record. In Turner II the same statute came back under the same standard, and the record now showed that many broadcasters would otherwise lose carriage and that most cable systems had dropped nothing to comply. The standard stayed fixed between the two decisions; the evidence is what changed.
Accepting the standard does not end disagreement about how to apply it. Justice O'Connor wrote an extended dissent in Turner II, joined by all the other dissenters from Turner I. Justice Breyer, who had succeeded Justice Blackmun, joined the majority but concurred separately to note important First Amendment interests on both sides of the equation, and he found the diversity rationale persuasive but not the competition rationale.86 Judges who accept the same two questions can therefore disagree about whether a given body of evidence is substantial.
Two limits keep this deference from swallowing the test. The first is classification. Deference applies to content-neutral rules supported by substantial evidence, and a content-based law faces strict scrutiny, which presumes the regulation invalid and demands narrowly tailored or necessary means to achieve a compelling interest.87 The second is that intermediate scrutiny still demands proof, because the Court subjects most content-neutral regulations to heightened rather than minimum rationality review.88