dissenting.
Just as “[c]onfidence in the integrity of our electoral processes is essential to the functioning of our participatory democracy,” Purcell v. Gonzalez, 549 U.S. 1, 4, 127 S. Ct. 5, 166 L. Ed. 2d 1 (2006) (per curiam), so too is citizen participation in those processes, which necessarily entails political speech and association under the First Amendment. In my view, compelled disclosure
[561 U.S. 229]
of signed referendum and initiative petition1 under the Washington Public Records Act (PRA), Wash. Rev. Code § 42.56.001 et seq. (2008), severely burdens those rights and chills citizen participation in the referendum process. Given those burdens, I would hold that Washington’s decision to subject all referendum petitions to public disclosure is unconstitutional because there will always be a less restrictive means by which Washington can vindicate its stated interest in preserving the integrity of its referendum process. I respectfully dissent.
I
This case concerns the interaction of two distinct sets of Washington statutes. The first set, codified in Washington’s Election Code, regulates the referendum and initiative process. These statutes require, among other things, that referendum signers write their names and addresses on petition sheets, and mandate that this information be disclosed to Washington’s secretary of state for canvassing and verification. See, e.g., §§ 29A.72.130, 29A.72.230 (2008). Petitioners do not contend that these requirements violate their First Amendment rights; that is, they do not argue that the Constitution allows them to support a referendum measure without disclosing their names to the State.
The second set of statutes—the PRA—is not a referendum or election regulation. Rather, the PRA requires disclosure of all nonexempt “public records” upon request by any person. See §§ 42.56.010(2), 42.56.070. Washington has concluded that signed referendum petitions are “public records” subject to disclosure under the PRA, and has “routinely disclosed petitions in response to public records requests.” Brief for Respondent Reed 5-6.
[561 U.S. 230]
Petitioners do not challenge the constitutionality of the PRA generally. They contend only that Washington violates their First Amendment rights by construing the PRA to apply to signed referendum petitions. See Brief for Petitioners 35-39. As the Court notes, the parties dispute whether this challenge is best conceived as a facial challenge or an as-applied challenge. See ante, at 194, 177 L. Ed. 2d, at 501. In my view, the Court correctly concludes that petitioners must “satisfy our standards for a facial challenge” because their claim, and the relief that they seek, “reach beyond” their “particular circumstances.” Ibid.
We typically disfavor facial chal*524lenges. See Washington State Grange v. Washington State Republican Party, 552 U.S. 442, 449, 128 S. Ct. 1184, 170 L. Ed. 2d 151 (2008). They “often rest on speculation,” can lead courts unnecessarily to anticipate constitutional questions or formulate broad constitutional rules, and may prevent governmental officers from implementing laws “in a manner consistent with the Constitution.” Id., at 450-451, 128 S. Ct. 1184, 170 L. Ed. 2d 151. For those reasons, we rejected in Washington State Grange political parties’ preenforcement facial challenge to a Washington initiative that allowed candidates in a primary election to self-designate their political party preference on the primary election ballot. See id., at 458-459, 128 S. Ct. 1184, 170 L. Ed. 2d 151. Because the challenge was a preenforcement one, Washington “had no opportunity to implement” the initiative, id., at 450, 128 S. Ct. 1184, 170 L. Ed. 2d 151, so the political parties’ arguments that it violated their association rights all depended “on the possibility that voters will be confused as to the meaning of the party-preference designation,” id., at 454, 128 S. Ct. 1184, 170 L. Ed. 2d 151. Moreover, a facial challenge was inappropriate because the regulation did “not on its face impose a severe burden on political parties’ associational rights.” Id., at 444, 128 S. Ct. 1184, 170 L. Ed. 2d 151.
Those considerations point in the opposite direction here. Washington’s construction of the PRA “on its face impose [s] a severe burden,” ibid.— compelled disclosure of privacy in political association protected by the First Amendment, see
[561 U.S. 231]
infra this page and 232, 177 L. Ed. 2d, at 524, 525—on all referendum signers. And Washington has had several “opportu-nit[ies] to implement” the PRA’s disclosure requirements with respect to initiative petitions. Washington State Grange, supra, at 450, 128 S. Ct. 1184, 170 L. Ed. 2d 151. Indeed, Washington admits that “[a] 11 petitions for initiatives, referendum, recall, and candidate nomination are public records subject to disclosure.” Brief for Respondent Reed 59; see also App. 26 (listing six completed requests for disclosure of signed initiative petitions since 2006). Washington thus has eliminated any “possibility” that referendum petition signers “will be confused as to” how the State will respond to a request under the PRA to disclose their names and addresses. Washington State Grange, 552 U.S., at 454, 128 S. Ct. 1184, 170 L. Ed. 2d 151.
Accordingly, I would consider petitioners’ facial challenge here. For purposes of this case, I will assume that to prevail, petitioners must satisfy our most rigorous standard, and show that there is “ ‘no set of circumstances . . . under which the’ ” PRA could be constitutionally applied to a referendum or initiative petition, “i.e., that the [PRA] is unconstitutional in all of its applications,” id., at 449, 128 S. Ct. 1184, 170 L. Ed. 2d 151 (quoting United States v. Salerno, 481 U.S. 739, 745, 107 S. Ct. 2095, 95 L. Ed. 2d 697 (1987)).
II
A
The Court correctly concludes that “an individual expresses” a “political view” by signing a referendum petition. Ante, at 194-195, 177 L. Ed. 2d, at 501-502. The Court also rightly rejects the baseless argument that such expressive activity falls “outside the scope of the First Amendment” merely be*525cause “it has legal effect in the electoral process.” Ante, at 195, 177 L. Ed. 2d, at 502. Yet, the Court does not acknowledge the full constitutional implications of these conclusions.
The expressive political activity of signing a referendum petition is a paradigmatic example of “the practice of persons sharing common views banding together to achieve a common end.” Citizens Against Rent Control I Coalition for
[561 U.S. 232]
Fair Housing v. Berkeley, 454 U.S. 290, 294, 102 S. Ct. 434, 70 L. Ed. 2d 492 (1981). A referendum supported by only one person’s signature is a nullity; it will never be placed on the ballot. The Doe petitioners recognized as much when they— and more than 120,000 other Washingtonians, see ante, at 192, 177 L. Ed. 2d, at 500—joined with petitioner Protect Marriage Washington, “a state political action committee” organized under § 42.17.040, to effect Protect Marriage Washington’s “major purpose” of collecting enough valid signatures to place Referendum 71 on the general election ballot. App. to Pet. for Cert. 29a. For these reasons, signing a referendum petition amounts to “ ‘political association’ ” protected by the First Amendment. Citizens Against Rent Control, supra, at 295, 102 S. Ct. 434, 70 L. Ed. 2d 492 (quoting Buckley v. Valeo, 424 U.S. 1, 15, 96 S. Ct. 612, 46 L. Ed. 2d 659 (1976) (per curiam)).
This Court has long recognized the “vital relationship between” political association “and privacy in one’s associations,” NAACP v. Alabama ex rel. Patterson, 357 U.S. 449, 462, 78 S. Ct. 1163, 2 L. Ed. 2d 1488 (1958), and held that “[t]he Constitution protects against the compelled disclosure of political associations and beliefs,” Brown v. Socialist Workers ’74 Campaign Comm. (Ohio), 459 U.S. 87, 91, 103 S. Ct. 416, 74 L. Ed. 2d 250 (1982). This constitutional protection “yield [s] only to a subordinating interest of the State that is compelling, and then only if there is a substantial relation between the information sought and an overriding and compelling state interest.” Id., at 91-92, 103 S. Ct. 416, 74 L. Ed. 2d 250 (internal quotation marks, citations, and brackets omitted). Thus, unlike the Court, I read our precedents to require application of strict scrutiny to laws that compel disclosure of protected First Amendment association. Buckley v. American Constitutional Law Foundation, Inc., 525 U.S. 182, 206, 212, 119 S. Ct. 636, 142 L. Ed. 2d 599 (1999) (ACLF) (Thomas, J., concurring in judgment). Under that standard, a disclosure requirement passes constitutional muster only if it is narrowly tailored—i.e., the least restrictive means—to serve a compelling state interest. See id., at 206, 119 S. Ct. 636, 142 L. Ed. 2d 599.
[561 U.S. 233]
B
Washington’s application of the PRAto a referendum petition does not survive strict scrutiny.
1
Washington first contends that it has a compelling interest in “transparency and accountability,” which it claims encompasses several subordinate interests: preserving the integrity of its election process, preventing corruption, deterring fraud, and correcting mistakes by the secretary of state or by petition signers. See Brief for Respondent Reed 40-42; 57-59.
It is true that a State has a substantial interest in regulating its referendum and initiative processes “to protect the[ir] integrity and reliability.” ACLF, 525 U.S., at 191, 119 S. Ct. *526636, 142 L. Ed. 2d 599. But Washington points to no precedent from this Court recognizing “correcting errors” as a distinct compelling interest that could support disclosure regulations. And our cases strongly suggest that preventing corruption and deterring fraud bear less weight in this particular electoral context: the signature-gathering stage of a referendum or initiative drive. The Court has twice observed that “ ‘the risk of fraud or corruption, or the appearance thereof, is more remote at the petition stage of an initiative than at the time of balloting.’ ” Id., at 203, 119 S. Ct. 636, 142 L. Ed. 2d 599 (quoting Meyer v. Grant, 486 U.S. 414, 427, 108 S. Ct. 1886, 100 L. Ed. 2d 425 (1988)). Similarly, because “ [referenda are held on issues, not candidates for public office,” the “risk of corruption perceived in cases involving candidate elections simply is not present in a popular vote on a public issue.” First Nat. Bank of Boston v. Bellotti, 435 U.S. 765, 790, 98 S. Ct. 1407, 55 L. Ed. 2d 707 (1978) (citations omitted).
We should not abandon those principles merely because Washington and its amici can point to a mere eight instances of initiative-related fraud, see Brief for Respondent Reed 42; Brief for State of Ohio et al. as Amici Curiae 22-24, among the 809 initiative measures placed on state ballots in this
[561 U.S. 234]
country between 1988 and 2008, see Initiative and Referendum Institute, Initiative Use 2 (Feb. 2009), online at http ://www. iandrinstitute. org/IRI % 20 Initiative % 20 Use % 20 (1904-2008).pdf (as visited June 21, 2010, and available in Clerk of Court’s case file). If anything, these meager figures reinforce the conclusion that the risks of fraud or corruption in the initiative and referendum process are remote and thereby undermine Washington’s claim that those two interests should be considered compelling for purposes of strict scrutiny.
Thus, I am not persuaded that Washington’s interest in protecting the integrity and reliability of its referendum process, as the State has defined that interest, is compelling. But I need not answer that question here. Even assuming the interest is compelling, on-demand disclosure of a referendum petition to any person under the PRA is “a blunderbuss approach” to furthering that interest, Colorado Republican Federal Campaign Comm. v. Federal Election Comm’n, 518 U.S. 604, 642, 116 S. Ct. 2309, 135 L. Ed. 2d 795 (1996) (Thomas, J., concurring in judgment and dissenting in part) (internal quotation marks omitted), not the least restrictive means of doing so. The events that prompted petitioners’ complaint in this case demonstrate as much.
As Washington explained during oral argument, after the secretary of state receives signed referendum petitions, his “first step ... is to take them to his archiving section and to have them digitized. As soon as they’re digitized, they’re available on disks for anyone who requests them” under the PRA. Tr. of Oral Arg. 30. In this case, two organizations announced their intention to obtain the digitized names and addresses of referendum signers and post them “online, in a searchable format.” Ante, at 193, 177 L. Ed. 2d, at 500.
There is no apparent reason why Washington must broadly disclose referendum signers’ names and addresses in this manner to vindicate the interest that it invokes here. Washington—which is in possession of that information because
[561 U.S. 235]
of refer*527endum regulations that petitioners do not challenge, see supra, at 229, 177 L. Ed. 2d, at 523—could put the names and addresses of referendum signers into a similar electronic database that state employees could search without subjecting the name and address of each signer to wholesale public disclosure. The secretary could electronically cross-reference the referendum database against the “statewide voter registration list” contained in Washington’s “statewide voter registration database,” § 29A.08.651(1),2 to ensure that each referendum signer meets Washington’s residency and voter registration requirements, see § 29A.72.130. Doing so presumably would drastically reduce or eliminate possible errors or mistakes that Washington argues the secretary might make, see Brief for Respondent Reed 42, since it would allow the secretary to verify virtually all of the signatures instead of the mere “3 to 5%” he “ordinarily checks,” ante, at 198, 177 L. Ed. 2d, at 504 (internal quotation marks omitted).3
An electronic referendum database would also enable the secretary to determine whether multiple entries correspond to a single registered voter, thereby detecting whether a voter had signed the petition more than once. In addition, the database would protect victims of “forgery” or “ ‘bait and switch’ fraud.” Ibid. In Washington, “a unique identifier is assigned to each legally registered voter in the state.” § 29A.08.651(4). Washington could create a Web site, linked to the electronic referendum database, where a voter concerned that his name had been fraudulently signed could conduct a search using his unique identifier to ensure that his name was absent from the database—without requiring disclosure
[561 U.S. 236]
of the names and addresses of all the voluntary, legitimate signers.
Washington admits that creating this sort of electronic referendum database “could be done.” Tr. of Oral Arg. 51. Implementing such a system would not place a heavy burden on Washington; “the Secretary of State’s staff’ already uses an “electronic voter registration database” in its “verification process.” Id., at 50.
Washington nevertheless contends that its citizens must “have access to public records ... to independently evaluate whether the Secretary properly determined to certify or not to certify a referendum to the ballot.” Brief for Respondent Reed 41. “[W]ithout the access to signed petitions that the PRA provides,” Washington argues, its “citizens could not fulfill their role as the final judge of public business.” Ibid, (internal quotation marks omitted).
But Washington’s Election Code already gives Washington voters access to referendum petition data. Under § 29A.72.230, “[t]he verification and canvass of signatures on the [referendum] petition may be observed by persons representing the advocates and opponents of the proposed measure so long as they make no record of the names, addresses, or other information on the petitions or related *528records during the verification process except upon” court order. Each side is entitled to at least two such observers, although the secretary may increase that number if, in his opinion, doing so would not “cause undue delay or disruption of the verification process.” Ibid.
Washington does not explain why this existing access, which petitioners do not challenge here, is insufficient to permit its citizens to oversee the verification process under § 29A.72.230, or to decide intelligently whether to pursue a court challenge under § 29A.72.240. Moreover, if Washington had implemented the more narrowly tailored electronic referendum database discussed above, observers could see the secretary of state’s employees examine the data using
[561 U.S. 237]
exactly the same techniques they would use if the data were released to them under the PRA. Obtaining a digitized list to navigate on their own computer would not allow an observer to learn any additional information.
Washington law also contains several other measures that preserve the integrity of the referendum process. First, it is a crime in Washington to forge a signature on a referendum petition, or to knowingly sign one more than once. See § 29A.84.230. Second, referendum supporters must gather a large number of valid signatures—four percent of the votes cast for Governor in the immediately preceding gubernatorial election—to place a referendum petition on the ballot. § 29A.72.150. Third, Washington’s required referendum petition form limits each petition to a single subject. See § 29A.72.130. Fourth, a large, plain-English warning must appear at the top of the referendum petition, alerting signers to the law’s requirements. See §29A.72.140. Fifth, Washington prescribes the text of the declaration that a circulator must submit along with the signed petition sheets. See § 29A.72.130. Sixth, Washington prescribes verification and canvassing methods. See § 29A.72.230.
The Court’s dismissive treatment of those provisions, see ante, at 198, 177 L. Ed. 2d, at 503-504, is perplexing, given the analysis that the Court endorsed in ACLF. There, the Court held that two disclosure requirements governing Colorado’s initiative process were unconstitutional, see 525 U.S., at 186-187, 119 S. Ct. 636, 142 L. Ed. 2d 599, specifically finding that they were “not warranted by the state interests (administrative efficiency, fraud detection, informing voters) alleged to justify” them, and emphasizing that its “judgment [wa]s informed by other means Colorado employs to accomplish its regulatory purposes,” id., at 192, 119 S. Ct. 636, 142 L. Ed. 2d 599. The entire last section of the Court’s opinion detailed those “less problematic measures” by which Colorado “can and. d[id] meet” its “substantial interests in regulating the ballot-initiative process.” Id., at 204, 119 S. Ct. 636, 142 L. Ed. 2d 599 (emphasis added). With one exception—a law deeming an initiative void if the circulator violated any
[561 U.S. 238]
law applicable to the circulation process— those Colorado laws correspond exactly to the Washington regulatory requirements listed above. See id., at 205, 119 S. Ct. 636, 142 L. Ed. 2d 599. Including the observer provision, § 29A.72.230, and the provision permitting court review of the secretary’s decision to certify (or not to certify) a referendum petition, § 29A.72.240, Washington thus appears to provide even more of the “less problematic measures” than Colorado did to “pro*529tect the integrity of the initiative process,” id., at 204, 119 S. Ct. 636, 142 L. Ed. 2d 599, and I see no reason why Washington’s identical provisions should not “inform” the analysis here.
It is readily apparent that Washington can vindicate its stated interest in “transparency and accountability” through a number of more narrowly tailored means than wholesale public disclosure. Accordingly, this interest cannot justify applying the PRA to a referendum petition.
2
Washington also contends that it has a compelling interest in “providing relevant information to Washington voters,” and that on-demand disclosure to the public is a narrowly tailored means of furthering that interest. Brief for Respondent Reed 44. This argument is easily dispatched, since this Court has already rejected it in a similar context.
In McIntyre v. Ohio Elections Comm’n, 514 U.S. 334, 115 S. Ct. 1511, 131 L. Ed. 2d 426 (1995), the Court held that an Ohio law prohibiting anonymous political pamphleting violated the First Amendment. One of the interests Ohio had invoked to justify that law was identical to Washington’s here: the “interest in providing the electorate with relevant information.” Id., at 348, 115 S. Ct. 1511, 131 L. Ed. 2d 426. The Court called that interest “plainly insufficient to support the constitutionality of [Ohio’s] disclosure requirement.” Id., at 349, 115 S. Ct. 1511, 131 L. Ed. 2d 426. “The simple interest in providing voters with additional relevant information does not justify a state requirement that a writer make statements or disclosures she would otherwise omit .’’ Id., at 348, 115 S. Ct. 1511, 131 L. Ed. 2d 426. “Don’t underestimate the
[561 U.S. 239]
common man,” we advised. Id., at 348, n. 11, 115 S. Ct. 1511, 131 L. Ed. 2d 426 (internal quotation marks omitted).
“People are intelligent enough to evaluate the source of an anonymous writing. They can see it is anonymous. They know it is anonymous. They can evaluate its anonymity along with its message. . . . And then, once they have done so, it is for them to decide what is ‘responsible,’ what is valuable, and what is truth.” Ibid, (internal quotation marks omitted).
See also Bellotti, 435 U.S., at 777, 98 S. Ct. 1407, 55 L. Ed. 2d 707 (“The inherent worth of the speech in terms of its capacity for informing the public does not depend upon the identity of its source”).
This observation applies equally to referendum measures. People are intelligent enough to evaluate the merits of a referendum without knowing who supported it. Thus, just as this informational interest did not justify the Ohio law in McIntyre, it does not justify applying the PRA to referendum petitions.
C
The foregoing analysis applies in every case involving disclosure of a referendum measure’s supporters, as it must for petitioners’ facial challenge to succeed. See Washington State Grange, 552 U.S., at 449, 128 S. Ct. 1184, 170 L. Ed. 2d 151 (quoting Salerno, 481 U.S., at 745, 107 S. Ct. 2095, 95 L. Ed. 2d 697). Washington does not argue that the strength of its transparency and accountability interest rises or falls based on the topic of a referendum. Nor would such an *530argument be convincing. We have no basis to assume that Washington’s interest in maintaining the integrity of its referendum process is high for a charter-school referendum but low for an unemployment insurance referendum, or that a library or land-use referendum is more likely to be a target of fraud or corruption than a referendum on insurance coverage and benefits. See ante, at 200-201, 177 L. Ed. 2d, at 505-506. The strength of Washington’s interest remains constant across all types of referendum measures.
[561 U.S. 240]
So too does the strength of a signer’s First Amendment interest. The First Amendment rights at issue here are associational rights, and a long, unbroken line of this Court’s precedents holds that privacy of association is protected under the First Amendment. See supra, at 231-232, 177 L. Ed. 2d, at 524-525. The loss of associational privacy that comes with disclosing referendum petitions to the general public under the PRA constitutes the same harm as to each signer of each referendum, regardless of the topic. To be sure, a referendum signer may be more willing to disclose to the general public his political association with persons signing certain referendum measures than his association with others. But that choice belongs to the voter; the State may not make it for him by ascribing a lower level of First Amendment protection to an associational interest that some think a voter may be (or should be) more willing to disclose. Cf. Rosenberger v. Rector and Visitors of Univ. of Va., 515 U.S. 819, 828, 115 S. Ct. 2510, 132 L. Ed. 2d 700 (1995) (“In the realm of private speech or expression, government regulation may not favor one speaker over another”).
Finally, the less restrictive means available to vindicate Washington’s transparency and accountability interest can be employed for all referendum measures, regardless of topic. There is nothing measure-specific about an electronic database or additional observers. And the forgery prohibition and other existing requirements in Washington law that help “protect the integrity of the initiative process,” ACLF, 525 U.S., at 204, 119 S. Ct. 636, 142 L. Ed. 2d 599, apply equally to all referendum measures.
Because the strength of Washington’s interest in transparency and a signer’s individual First Amendment interest in privacy of political association remain constant across all referendum topics, and because less restrictive means to protect the integrity of the referendum process are not topic specific, I would hold that on-demand public disclosure of referendum petitions under the PRA is not narrowly tailored for any referendum.
[561 U.S. 241]
III
Significant practical problems will result from requiring as-applied challenges to protect referendum signers’ constitutional rights.
A
The Court’s approach will “require substantial litigation over an extended time” before a potential signer of any referendum will learn whether, if he signs a referendum, his associational privacy right will remain intact. Citizens United v. Federal Election Comm’n, 558 U.S. 310, 326, 130 S. Ct. 876, 175 L. Ed. 2d 753 (2010). And the tenacious litigant’s reward for trying to protect his First Amendment rights? An “interpretive process [that] itself would create an inevitable, pervasive, and serious risk of *531chilling protected speech pending the drawing of fine distinctions that, in the end, would themselves be questionable.” Id., at 327, 130 S. Ct. 876, 175 L. Ed. 2d 753. The large number of such fine and questionable distinctions in these types of cases reinforces my view that as-applied challenges provide no more than “a hollow assurance” that referendum signers’ First Amendment rights will be protected. Id., at 484, 130 S. Ct. 876, 175 L. Ed. 2d 753 (Thomas, J., concurring in part and dissenting in part). Consider just a few examples.
In Washington, a referendum sponsor must file the proposed referendum with the secretary of state before collecting signatures. See § 29A.72.010. May the sponsor seek an injunction against disclosure through an as-applied challenge before filing the proposed measure, or simultaneously with its filing? Because signature gathering will not have started, the sponsor will not be able to present any evidence specific to signers or potential signers of that particular referendum, showing “a reasonable probability that the compelled disclosure [of personal information] will subject them to threats, harassment, or reprisals from either Government officials or private parties.” Ante, at 200, 177 L. Ed. 2d, at 505 (internal quotation marks omitted). Thus, to succeed at that stage of litigation,
[561 U.S. 242]
plaintiffs must point to (at least) one other instance of harassment arising from a similar referendum. The Court has never held that such evidence would be acceptable; but if it is, that necessarily means that some signers, at some point, will have suffered actual “threats, harassment, and reprisals” for engaging in protected First Amendment activity.
If the sponsor must wait at least until signature gathering has started on his referendum to file an as-applied challenge, it is still unclear what sort of evidence of “threats, harassment, or reprisals” directed toward his supporters would satisfy the Court’s standard. How many instances of “threats, harassment, or reprisals” must a signer endure before a court may grant relief on an as-applied challenge? And how dispersed throughout the group of the necessary 120,000 signers, see ante, at 192, 177 L. Ed. 2d, at 500, must these threats be?
More importantly, the Court’s standard does not appear to require actual “threats, harassment, or reprisals,” but merely a “ ‘reasonable probability’ ” that disclosure of the signers’ names and addresses will lead to such activity. Ante, at 200, 177 L. Ed. 2d, at 505 (emphasis added). What sort of evidence suffices to satisfy this apparently more relaxed, though perhaps more elusive, standard? Does one instance of actual harassment directed toward one signer mean that the “reasonable probability” requirement is met? And again, how widespread must this “reasonable probability” be? The Court does not answer any of these questions, leaving a vacuum to be filled on a case-by-case basis. This will, no doubt, result in the “drawing of’ arbitrary and “questionable” “fine distinctions” by even the most well-intentioned district or circuit judge. Citizens United, 558 U.S., at 327, 130 S. Ct. 876, 175 L. Ed. 2d 753.
B
In addition, as I have previously explained, the state of technology today creates at least some probability that signers of every referendum will be subjected to threats, harassment, *532or reprisals if their personal information is disclosed.
[561 U.S. 243]
“ ‘[T]he advent of the Internet’ enables” rapid dissemination of “ ‘the information needed’ to” threaten or harass every referendum signer. Id., at 484, 130 S. Ct. 876, 175 L. Ed. 2d 753 (opinion of Thomas, J.). “Thus, ‘disclosure permits citizens . . . to react to the speech of [their political opponents] in a proper’—or undeniably improper—‘way’ long before a plaintiff could prevail on an as-applied challenge.” Ibid.
The Court apparently disagrees, asserting that “there is no reason to assume that any burdens imposed by disclosure of typical referendum petitions would be remotely like the burdens plaintiffs fear in this case.” Ante, at 201, 177 L. Ed. 2d, at 505. That conclusion rests on the premise that some referendum measures are so benign that the fact of public disclosure will not chill protected First Amendment activity. I am not convinced that this premise is correct.
The historical evidence shows that the referendum and initiative process first gained popularity as a means of “provid [ing] an occasional safety valve for interests that failed to get a fair hearing in the legislatures.” T. Cronin, Direct Democracy: The Politics of Initiative, Referendum, and Recall 59 (1989). Unsurprisingly, such interests tended to be controversial by nature. Early examples include “the single tax, prohibition, women’s suffrage, prolabor legislation, and the graduated income tax.”Id., at 58. And proponents of initiative measures tended to include politically marginalized groups such as the “Farmer’s Alliance” in rural States; “[t]housands of labor federations, notably the miners”; and “the Women’s Suffrage Association,” which “saw the initiative and referendum as a possible new means to overcome” repeated failed attempts in state legislatures to secure for women the right to vote. Id., at 50-51.
These characteristics of initiative and referendum drives persist today. Consider, for example, the goal of increasing ethics in government—a seemingly laudable and unobjectionable goal. So thought some citizens of Utah, who, frustrated with the state legislature’s failure to pass ethics laws
[561 U.S. 244]
commensurate with their preferences, filed a “21-page initiative tar-getting] legislative conduct with a broad array of reforms that would significantly change how business gets done on Utah’s Capitol Hill.” McKitrick, Suit Demands Secrecy for Ethics Petition Signers, Salt Lake Tribune, Apr. 15, 2010, p. A4 (hereinafter Salt Lake Tribune). But Utah law provides that “[initiative packets,” which contain the names and addresses (and, in some cases, birth-dates) of petition signers, “are public once they are delivered to the county clerks” for verification and canvassing. Utah Code Ann. § 20A-7-206(7) (2009 Lexis Supp. Pamphlet).
The attorneys sponsoring that initiative moved for an injunction to prevent disclosure of the initiative packets under § 20A-7-206(7) because, they claimed, ‘“[t]he [state] Republican Party has said it will target our folks.’ ” Salt Lake Tribune A4. According to these attorneys, a facially benign initiative may well result in political retribution and retaliation in a State where Republicans currently hold the offices of Governor, Lieutenant Governor, attorney general, state treasurer, state auditor, and a super-majority in both the Utah House of Representatives (71%) and the Utah Senate (72%), see State Yellow Book: Who’s Who in the Executive and Legislative Branches of the 50 State Gov*533ernments 650-651,1292-1294 (Spring 2010), as well as four of the five seats in the State’s delegation to the United States Congress, see GPO, 2009-2010 Official Congressional Directory, 111th Cong., pp. 299, 307 (2009).
The difficulty in predicting which referendum measures will prove controversial—combined with Washington’s default position that signed referendum petitions will be disclosed on demand, thereby allowing anyone to place this information on the Internet for broad dissemination—raises the significant probability that today’s decision will “inhibit the exercise of legitimate First Amendment activity” with respect to referendum and initiative petitions. Colorado Republican, 518 U.S., at 634, 116 S. Ct. 2309, 135 L. Ed. 2d 795 (Thomas, J., concurring in judgment
[561 U.S. 245]
and dissenting in part). “[DJisclosure requirements enable private citizens and elected officials to implement political strategies specifically calculated, to curtail campaign-related activity and prevent the lawful, peaceful exercise of First Amendment rights.” Citizens United, 558 U.S., at 483, 130 S. Ct. 876, 175 L. Ed. 2d 753 (Thomas, J., concurring in part and dissenting in part). Our cases have long recognized this reality;4 as the Court recently reiterated, the First Amendment does not require “case-by-case determinations” if “archetypical” First Amendment rights “would be chilled in the meantime.” Id., at 329, 130 S. Ct. 876, 175 L. Ed. 2d 753.
This chill in protected First Amendment activity harms others besides the dissuaded signer. We have already expressed deep skepticism about restrictions that “mak[ej it less likely that” a referendum “will garner the number of signatures necessary to place the matter on the ballot, thus limiting [the] ability to make the matter the focus of statewide discussion.” Meyer, 486 U.S., at 423, 108 S. Ct. 1886, 100 L. Ed. 2d 425. Such restrictions “inevitabl[y] . . . reduc[e] the total quantum of speech on a public issue.” Ibid, The very public that the PRA is supposed to serve is thus harmed by the way Washington implements that statute here.
Petitioners do not argue that the Constitution gives supporters of referendum petitions a right to act without anyone
[561 U.S. 246]
knowing their identities. Thus, Washington’s requirements that referendum supporters sign their names and addresses to a referendum petition, and that this information be disclosed to the State for canvassing and verification, see Wash. Rev. Code § 29A.72.230, are not at issue. And, petitioners do not contend that Washington’s citizens may never obtain access to referendum data. Thus, Washington’s rules allowing access to at least two representative observers from each side, see ibid,, and autho*534rizing courts to review the secretary of state’s verification and canvassing decision if those observers are dissatisfied with the secretary’s decision, see § 29A.72.240, are also not in question.
The Court is asked to assess the constitutionality of the PRA only with regard to referendum petitions. The question before us is whether all signers of all referendum petitions must resort to “substantial litigation over an extended time,” Citizens United, supra, at 326, 130 S. Ct. 876, 175 L. Ed. 2d 753, to prevent Washington from trenching on their protected First Amendment rights by subjecting their referendum petition signatures to on-demand public disclosure. In my view, they need not.
. Generally speaking, in a referendum, voters approve or reject an Act already passed by the legislature. In an initiative, voters adopt or reject an entirely new law, either a statute or a constitutional amendment. See T. Cronin, Direct Democracy: The Politics of Initiative, Referendum, and Recall 2 (1989).
. Under Washington law, this “computerized list must serve as the single system for storing and maintaining the official list of registered voters throughout the state’’ and “must contain the name and registration information of every legally registered voter in the state.’’ Wash. Rev. Code §§ 29A.08.651(2)-(3) (2008).
. See § 29A.72.230 (permitting the secretary of state to verify and canvass referendum petitions using approved statistical sampling methods).
. See, e.g, NAACP v. Alabama ex rel. Patterson, 357 U.S. 449, 462, 78 S. Ct. 1163, 2 L. Ed. 2d 1488 (1958) (noting the “hardly . . . novel perception that compelled disclosure of affiliation with groups engaged in advocacy may constitute’’ an “effective . . . restraint on freedom of association’’); Bates v. Little Rock, 361 U.S. 516, 523, 80 S. Ct. 412, 4 L. Ed. 2d 480 (1960) (“Freedoms such as’’ the “freedom of association for the purpose of advancing ideas and airing grievances’’ are “protected not only against heavy-handed frontal attack, but also from being stifled by more subtle government interference’’); see also id., at 528, 80 S. Ct. 412, 4 L. Ed. 2d 480 (Black and Douglas, JJ., concurring) (“First Amendment rights are beyond abridgment either by legislation that directly restrains their exercise or by suppression or impairment through harassment, humiliation, or exposure by government" (emphasis added)).