¹û¶³´«Ã½app¹Ù·½

544 U.S. 581 (2005)

CLINGMAN, SECRETARY, OKLAHOMA STATE ELECTION BOARD, ET AL.
v.
BEAVER ET AL.

No. 04-37.

Supreme Court of United States.

Argued January 19, 2005.
Decided May 23, 2005.
CERTIORARI TO THE UNITED STATES COURT OF APPEALS FOR THE TENTH CIRCUIT.

JUSTICE THOMAS delivered the opinion of the Court except as to Part II-A, concluding that Oklahoma's semiclosed primary system does not violate the right to freedom of association. Any burden it imposes is minor and justified by legitimate state interests. Pp. 586-587, 591-598.

(a) The First Amendment protects citizens' right "to band together in promoting among the electorate candidates who espouse their political views." California Democratic Party v. Jones, 530 U.S. 567, 574. Regulations imposing severe burdens on associational rights must be narrowly tailored to serve a compelling state interest, but when they impose lesser burdens, "a State's important regulatory interests will usually be enough to justify reasonable, nondiscriminatory restrictions." Timmons v. Twin Cities Area New Party, 520 U. S. 351, 358. In Tashjian v. Republican Party of Conn., 479 U. S. 208, 224, n. 13, the Court *582 left open the question whether a State may prevent a political party from inviting registered voters of other parties to vote in its primary. Pp. 586-587.

(b) Oklahoma's system does not severely burden associational rights. The Court disagrees with respondents' argument that the burden Oklahoma imposes is no less severe than the burden at issue in Tashjian, and thus the Court must apply strict scrutiny as it did in Tashjian. Tashjian applied strict scrutiny without carefully examining the burden on associational rights. Not every electoral law burdening associational rights is subject to strict scrutiny, which is appropriate only if the burden is severe, e. g., Jones, supra, at 582. Requiring voters to register with a party before participating in its primary minimally burdens voters' associational rights. Moreover, Tashjian is distinguishable. Oklahoma's semiclosed primary imposes an even less substantial burden than did the Connecticut closed primary at issue in Tashjian. Unlike that law, Oklahoma's system does not require Independent voters to affiliate publicly with a party to vote in its primary, 479 U. S., at 216, n. 7. Although, like the earlier law, Oklahoma's statute does not allow parties to "broaden opportunities for joining . . . by their own act," but requires "intervening action by potential voters," ibid., this burden is not severe, since many electoral regulations require that voters take some action to participate in the primary process. Such minor barriers between voter and party do not compel strict scrutiny. See Bullock v. Carter, 405 U. S. 134, 143. To deem ordinary and widespread burdens like these severe would subject virtually every electoral regulation to strict scrutiny, hamper the ability of States to run efficient and equitable elections, and compel federal courts to rewrite state electoral codes. The Constitution does not require that result. Pp. 591-593.

(c) Oklahoma's primary advances a number of regulatory interests this Court recognizes as important: It "preserv[es] [political] parties as viable and identifiable interest groups," Nader v. Schaffer, 417 F. Supp. 837, 845 (Conn.), aff'd, 429 U. S. 989; enhances parties' electioneering and party-building efforts, 417 F. Supp., at 848; and guards against party raiding and "sore loser" candidacies by spurned primary contenders, Storer v. Brown, 415 U. S. 724, 735. Pp. 593-597.

(d) The Court declines to consider respondents' expansion of their challenge to include several of Oklahoma's ballot access and voter registration laws. Those claims were neither raised nor decided below, see, e. g., Cooper Industries, Inc. v. Aviall Services, Inc., 543 U. S. 157, 168-169, and respondents have pointed to no unusual circumstances warranting their consideration now, see Taylor v. Freeland & Kronz, 503 U. S. 638, 645-646. Pp. 597-598.

*583 JUSTICE THOMAS, joined by THE CHIEF JUSTICE, JUSTICE SCALIA, and JUSTICE KENNEDY, concluded in Part II-A that a voter unwilling to disaffiliate from another party in order to vote in the LPO's primary forms little "association" with the LPO — nor the LPO with him. See Tashjian, supra, at 235. But even if Oklahoma's system burdens an associational right, the burden is less severe than others this Court has upheld as constitutional. The reasons underpinning Timmons, supra, show that Oklahoma's system burdens the LPO only minimally. As in Timmons, Oklahoma's law does not regulate the LPO's internal processes, its authority to exclude unwanted members, or its capacity to communicate with the public. And just as in Timmons, in which a Minnesota law conditioned a party's ability to nominate the candidate of its choice on the candidate's willingness to disaffiliate from another party, Oklahoma conditions a party's ability to welcome a voter into its primary on the voter's willingness to dissociate from his current party of choice. If a party may be prevented from associating with its desired standard bearer because he refuses to disaffiliate from another party, it may also be prevented from associating with a voter who refuses to do the same. Oklahoma's system imposes an even slighter burden on voters than on the LPO. Disaffiliation is not difficult: Other parties' registered members who wish to vote in the LPO primary simply need to file a form changing their registration. Voters are not "locked in" to an unwanted party affiliation, see Kusper v. Pontikes, 414 U. S. 51, 60-61, because with only nominal effort they are free to vote in the LPO primary. Pp. 587-591.

JUSTICE O'CONNOR, joined by JUSTICE BREYER except as to Part III, agreed with most of the Court's reasoning, but wrote separately to emphasize two points. First, the Libertarian Party of Oklahoma (LPO) and voters registered with another party have constitutionally cognizable interests in associating with one another through the LPO's primary, and these interests should not be minimized to dispose of this case. Second, while the Court is correct that only Oklahoma's semiclosed primary law is properly under review, that standing alone it imposes only a modest, nondiscriminatory burden on respondents' associational rights, and that this burden is justified by the State's legitimate regulatory interests, there are some grounds for concern that other Oklahoma laws governing party recognition and changes in party affiliation may unreasonably restrict voters' ability to participate in the LPO's primary. A realistic assessment of regulatory burdens on associational rights would, in an appropriate case, require examination of the cumulative effects of the State's overall primary scheme; and any finding of a more severe burden would trigger more probing review of the State's justifications. Pp. 598-608.

*584 THOMAS, J., delivered an opinion, which was for the Court except as to Part II-A. REHNQUIST, C. J., and SCALIA and KENNEDY, JJ., joined that opinion in full, and O'CONNOR and BREYER, JJ., joined except as to Part II-A. O'CONNOR, J., filed an opinion concurring in part and concurring in the judgment, in which BREYER, J., joined except as to Part III, post, p. 598. STEVENS, J., filed a dissenting opinion, in which GINSBURG, J., joined, and in which SOUTER, J., joined as to Parts I, II, and III, post, p. 608.

Wellon B. Poe, Jr., Assistant Attorney General of Oklahoma, argued the cause for petitioners. With him on the briefs was W. A. Drew Edmondson, Attorney General.

James C. Linger argued the cause and filed a brief for respondents.^[*]

JUSTICE THOMAS delivered the opinion of the Court, except as to Part II-A.

Oklahoma has a semiclosed primary system, in which a political party may invite only its own party members and voters registered as Independents to vote in the party's primary. The Court of Appeals held that this system violates the right to freedom of association of the Libertarian Party of Oklahoma (LPO) and several Oklahomans who are registered members of the Republican and Democratic Parties. We hold that it does not.

I

Oklahoma's election laws provide that only registered members of a political party may vote in the party's primary, *585 see Okla. Stat. Ann., Tit. 26, § 1-104(A) (West 1997), unless the party opens its primary to registered Independents as well, see § 1-104(B)(1). In May 2000, the LPO notified the secretary of the Oklahoma State Election Board that it wanted to open its upcoming primary to all registered Oklahoma voters, without regard to their party affiliation. See § 1-104(B)(4) (requiring notice when a party opens its primary to Independents). Pursuant to § 1-104, the secretary agreed as to Independent voters, but not as to voters registered with other political parties. The LPO and several Republican and Democratic voters then sued for declaratory and injunctive relief in the United States District Court for the Western District of Oklahoma, alleging that Oklahoma's semiclosed primary law unconstitutionally burdens their First Amendment right to freedom of political association. App. 20.

After a hearing, the District Court declined to enjoin Oklahoma's semiclosed primary law for the 2000 primaries. After a 2-day bench trial following the primary election, the District Court found that Oklahoma's semiclosed primary system did not severely burden respondents' associational rights. Further, it found that any burden imposed by the system was justified by Oklahoma's asserted interest in "preserving the political parties as viable and identifiable interest groups, [and] insuring that the results of a primary election . . . accurately reflect the voting of the party members." Memorandum Opinion, Case No. CIV-00-1071-F (WD Okla., Jan. 24, 2003), App. to Pet. for Cert. 55-56 (hereinafter Memorandum Opinion) (internal quotation marks omitted). The District Court therefore upheld the semiclosed primary statute as constitutional. Id., at 72-73.

On appeal, the Court of Appeals for the Tenth Circuit reversed the judgment of the District Court. The Court of Appeals concluded that the State's semiclosed primary statute imposed a severe burden on respondents' associational rights, and thus was constitutional only if the statute was *586 narrowly tailored to serve a compelling state interest. 363 F. 3d 1048, 1057-1058 (2004). Finding none of Oklahoma's interests compelling, the Court of Appeals enjoined Oklahoma from using its semiclosed primary law. Id., at 1060-1061. Because the Court of Appeals' decision not only prohibits Oklahoma from using its primary system but also casts doubt on the semiclosed primary laws of 23 other States,^[1] we granted certiorari. 542 U. S. 965 (2004).

II

The Constitution grants States "broad power to prescribe the `Times, Places and Manner of holding Elections for Senators and Representatives,' Art. I, §4, cl. 1, which power is matched by state control over the election process for state offices." Tashjian v. Republican Party of Conn., 479 U. S. 208, 217 (1986); Timmons v. Twin Cities Area New Party, 520 U. S. 351, 358 (1997) (quoting Tashjian). We have held that the First Amendment, among other things, protects the right of citizens "to band together in promoting among the electorate candidates who espouse their political views." California Democratic Party v. Jones, 530 U. S. 567, 574 (2000). Regulations that impose severe burdens on associational rights must be narrowly tailored to serve a compelling state interest. Timmons, 520 U. S., at 358. However, *587 when regulations impose lesser burdens, "a State's important regulatory interests will usually be enough to justify reasonable, nondiscriminatory restrictions." Ibid. (internal quotation marks omitted).

In Tashjian, this Court struck down, as inconsistent with the First Amendment, a closed primary system that prevented a political party from inviting Independent voters to vote in the party's primary. 479 U. S., at 225. This case presents a question that Tashjian left open: whether a State may prevent a political party from inviting registered voters of other parties to vote in its primary. Id., at 224, n. 13. As Tashjian acknowledged, opening a party's primary "to all voters, including members of other parties, . . . raise[s] a different combination of considerations." Ibid. We are persuaded that any burden Oklahoma's semiclosed primary imposes is minor and justified by legitimate state interests.

A

At the outset, we note that Oklahoma's semiclosed primary system is unlike other laws this Court has held to infringe associational rights. Oklahoma has not sought through its electoral system to discover the names of the LPO's members, see NAACP v. Alabama ex rel. Patterson, 357 U. S. 449, 451 (1958); to interfere with the LPO by restricting activities central to its purpose, see NAACP v. Claiborne Hardware Co., 458 U. S. 886, 895 (1982); NAACP v. Button, 371 U. S. 415, 423-426 (1963); to disqualify the LPO from public benefits or privileges, see Keyishian v. Board of Regents of Univ. of State of N. Y., 385 U. S. 589, 595-596 (1967); or to compel the LPO's association with unwanted members or voters, see Jones, supra, at 577. The LPO is free to canvass the electorate, enroll or exclude potential members, nominate the candidate of its choice, and engage in the same electoral activities as every other political party in Oklahoma. Oklahoma merely prohibits the LPO from leaving the selection of its candidates to people who are members of *588 another political party. Nothing in § 1-104 prevents members of other parties from switching their registration to the LPO or to Independent status.^[2] The question is whether the Constitution requires that voters who are registered in other parties be allowed to vote in the LPO's primary.

In other words, the Republican and Democratic voters who have brought this action do not want to associate with the LPO, at least not in any formal sense. They wish to remain registered with the Republican, Democratic, or Reform parties, and yet to assist in selecting the Libertarian Party's candidates for the general election. Their interest is in casting a vote for a Libertarian candidate in a particular primary election,^[3] rather than in banding together with fellow citizens committed to the LPO's political goals and ideas. See Jones, supra, at 573-574, n. 5 ("As for the associational `interest' in selecting the candidate of a group to which one does not belong, that falls far short of a constitutional right, if indeed it *589 can even fairly be characterized as an interest"). And the LPO is happy to have their votes, if not their membership on the party rolls.

However, a voter who is unwilling to disaffiliate from another party to vote in the LPO's primary forms little "association" with the LPO — nor the LPO with him. See Tashjian, supra, at 235 (SCALIA, J., dissenting). That same voter might wish to participate in numerous party primaries, or cast ballots for several candidates, in any given race. The issue is not "dual associations," post, at 601 (O'CONNOR, J., concurring in part and concurring in judgment), but seemingly boundless ones. "If the concept of freedom of association is extended" to a voter's every desire at the ballot box, "it ceases to be of any analytic use." Tashjian, supra, at 235 (SCALIA, J., dissenting); cf. Democratic Party of United States v. Wisconsin ex rel. La Follette, 450 U. S. 107, 130 (1981) (Powell, J., dissenting) ("[Not] every conflict between state law and party rules concerning participation in the nomination process creates a burden on associational rights").

But even if Oklahoma's semiclosed primary system burdens an associational right, the burden is less severe than others this Court has upheld as constitutional. For instance, in Timmons, we considered a Minnesota election law prohibiting multiparty, or "fusion," candidacies in which a candidate appears on the ballot as the nominee of more than one party. 520 U. S., at 353-354. Minnesota's law prevented the New Party, a minor party under state law, from putting forward the same candidate as a major party. The New Party challenged the law as unconstitutionally burdening its associational rights. Id., at 354-355. This Court concluded that the burdens imposed by Minnesota's law — "though not trivial — [were] not severe." Id., at 363.

The burdens were not severe because the New Party and its members remained free to govern themselves internally and to communicate with the public as they wished. Ibid. *590 Minnesota had neither regulated the New Party's internal decisionmaking process, nor compelled it to associate with voters of any political persuasion, see Jones, 530 U. S., at 577. The New Party and its members simply could not nominate as their candidate any of "those few individuals who both have already agreed to be another party's candidate and also, if forced to choose, themselves prefer that other party." Timmons, supra, at 363.

The same reasons underpinning our decision in Timmons show that Oklahoma's semiclosed primary system burdens the LPO only minimally. As in Timmons, Oklahoma's law does not regulate the LPO's internal processes, its authority to exclude unwanted members, or its capacity to communicate with the public. And just as in Timmons, in which Minnesota conditioned the party's ability to nominate the candidate of its choice on the candidate's willingness to disaffiliate from another political party, Oklahoma conditions the party's ability to welcome a voter into its primary on the voter's willingness to dissociate from his current party of choice. If anything, it is "[t]he moment of choosing the party's nominee" that matters far more, Jones, 530 U. S., at 575, for that is "`the crucial juncture at which the appeal to common principles may be translated into concerted action, and hence to political power in the community,'" ibid. (quoting Tashjian, 479 U. S., at 216). If a party may be prevented from associating with the candidate of its choice — its desired "`standard bearer,'" Timmons, supra, at 359; Jones, supra, at 575 — because that candidate refuses to disaffiliate from another political party, a party may also be prevented from associating with a voter who refuses to do the same.

Oklahoma's semiclosed primary system imposes an even slighter burden on voters than on the LPO. Disaffiliation is not difficult: In general, "anyone can `join' a political party merely by asking for the appropriate ballot at the appropriate time or (at most) by registering within a state-defined reasonable period of time before an election." Jones, supra, *591 at 596 (STEVENS, J., dissenting). In Oklahoma, registered members of the Republican, Democratic, and Reform Parties who wish to vote in the LPO primary simply need to file a form with the county election board secretary to change their registration. See Okla. Stat. Ann., Tit. 26, § 4-119 (West Supp. 2005). Voters are not "locked in" to an unwanted party affiliation, see Kusper v. Pontikes, 414 U. S. 51, 60-61 (1973), because with only nominal effort they are free to vote in the LPO primary. For this reason, too, the registration requirement does not unduly hinder the LPO from associating with members of other parties. To attract members of other parties, the LPO need only persuade voters to make the minimal effort necessary to switch parties.

B

Respondents argue that this case is no different from Tashjian. According to respondents, the burden imposed by Oklahoma's semiclosed primary system is no less severe than the burden at issue in Tashjian, and hence we must apply strict scrutiny as we did in Tashjian. We disagree. At issue in Tashjian was a Connecticut election statute that required voters to register with a political party before participating in its primary. 479 U.S., at 210-211. The State's Republican Party, having adopted a rule that allowed Independent voters to participate in its primary, contended that Connecticut's closed primary infringed its right to associate with Independent voters. Ibid. Applying strict scrutiny, this Court found that the interests Connecticut advanced to justify its ban were not compelling, and thus that the State could not constitutionally prevent the Republican Party from inviting into its primary willing Independent voters. Id., at 217-225.

Respondents' reliance on Tashjian is unavailing. As an initial matter, Tashjian applied strict scrutiny with little discussion of the magnitude of the burdens imposed by Connecticut's closed primary on parties' and voters' associational *592 rights. Post, at 605 (O'CONNOR, J., concurring in part and concurring in judgment). But not every electoral law that burdens associational rights is subject to strict scrutiny. See, e. g., Nader v. Schaffer, 417 F. Supp. 837, 849 (Conn.) ("There must be more than a minimal infringement on the rights to vote and of association . . . before strict judicial review is warranted"), aff'd, 429 U. S. 989 (1976). Instead, as our cases since Tashjian have clarified, strict scrutiny is appropriate only if the burden is severe. Jones, supra, at 582; Timmons, 520 U. S., at 358. In Tashjian itself, Independent voters could join the Connecticut Republican Party as late as the day before the primary. 479 U. S., at 219. As explained above, supra, at 590-591, requiring voters to register with a party prior to participating in the party's primary minimally burdens voters' associational rights.

Nevertheless, Tashjian is distinguishable. Oklahoma's semiclosed primary imposes an even less substantial burden than did the Connecticut closed primary at issue in Tashjian. In Tashjian, this Court identified two ways in which Connecticut's closed primary limited citizens' freedom of political association. The first and most important was that it required Independent voters to affiliate publicly with a party to vote in its primary. 479 U. S., at 216, n. 7. That is not true in this case. At issue here are voters who have already affiliated publicly with one of Oklahoma's political parties. These voters need not register as Libertarians to vote in the LPO's primary; they need only declare themselves Independents, which would leave them free to participate in any party primary that is open to registered Independents. See Okla. Stat. Ann., Tit. 26, § 1-104(B)(1) (West 1997).

The second and less important burden imposed by Connecticut's closed primary system was that political parties could not "broaden opportunities for joining . . . by their own act, without any intervening action by potential voters." Tashjian, 479 U. S., at 216, n. 7. Voters also had to act by registering themselves in a particular party. Ibid. That is *593 equally true of Oklahoma's semiclosed primary system: Voters must register as Libertarians or Independents to participate in the LPO's primary. However, Tashjian did not characterize this burden alone as severe, and with good reason. Many electoral regulations, including voter registration generally, require that voters take some action to participate in the primary process. See, e. g., Rosario v. Rockefeller, 410 U. S. 752, 760-762 (1973) (upholding requirement that voters change party registration 11 months in advance of the primary election). Election laws invariably "affec[t] — at least to some degree — the individual's right to vote and his right to associate with others for political ends." Anderson v. Celebrezze, 460 U. S. 780, 788 (1983).

These minor barriers between voter and party do not compel strict scrutiny. See Bullock v. Carter, 405 U. S. 134, 143 (1972). To deem ordinary and widespread burdens like these severe would subject virtually every electoral regulation to strict scrutiny, hamper the ability of States to run efficient and equitable elections, and compel federal courts to rewrite state electoral codes. The Constitution does not require that result, for it is beyond question "that States may, and inevitably must, enact reasonable regulations of parties, elections, and ballots to reduce election- and campaign-related disorder." Timmons, supra, at 358; Storer v. Brown, 415 U. S. 724, 730 (1974). Oklahoma's semiclosed primary system does not severely burden the associational rights of the State's citizenry.

C

When a state electoral provision places no heavy burden on associational rights, "a State's important regulatory interests will usually be enough to justify reasonable, nondiscriminatory restrictions." Timmons, supra, at 358 (internal quotation marks omitted); Anderson, supra, at 788. Here, Oklahoma's semiclosed primary advances a number of regulatory interests that this Court recognizes as important: It *594 "preserv[es] [political] parties as viable and identifiable interest groups," Nader, 417 F. Supp., at 845; enhances parties' electioneering and party-building efforts, id., at 848; and guards against party raiding and "sore loser" candidacies by spurned primary contenders, Storer, supra, at 735.

First, as Oklahoma asserts, its semiclosed primary "preserv[es] the political parties as viable and identifiable interest groups, insuring that the results of a primary election, in a broad sense, accurately reflec[t] the voting of the party members." Amended and Supplemental Trial Brief of Defendants 10, Record Doc. 63 (quoting without attribution Nader, supra, at 845). The LPO wishes to open its primary to registered Republicans and Democrats, who may well vote in numbers that dwarf the roughly 300 registered LPO voters in Oklahoma. See Memorandum Opinion 31-32 (at least 95% of voters in LPO's 1996 primary were independents, not Libertarians). If the LPO is permitted to open its primary to all registered voters regardless of party affiliation, the candidate who emerges from the LPO primary may be "unconcerned with, if not . . . hostile to," the political preferences of the majority of the LPO's members. Nader, supra, at 846. It does not matter that the LPO is willing to risk the surrender of its identity in exchange for electoral success. Oklahoma's interest is independent and concerns the integrity of its primary system. The State wants to "avoid primary election outcomes which would tend to confuse or mislead the general voting population to the extent [it] relies on party labels as representative of certain ideologies." Brief for Petitioners 12 (quoting without attribution Nader, supra, at 845); Eu v. San Francisco County Democratic Central Comm., 489 U. S. 214, 228 (1989).

Moreover, this Court has found that "`[i]n facilitating the effective operation of [a] democratic government, a state might reasonably classify voters or candidates according to political affiliations.'" Nader, supra, at 845-846 (quoting Ray v. Blair, 343 U. S. 214, 226, n. 14 (1952)). But for that *595 classification to mean much, Oklahoma must be allowed to limit voters' ability to roam among parties' primaries. The purpose of party registration is to provide "a minimal demonstration by the voter that he has some `commitment' to the party in whose primary he wishes to participate." Nader, supra, at 847. That commitment is lessened if party members may retain their registration in one party while voting in another party's primary. Opening the LPO's primary to all voters not only would render the LPO's imprimatur an unreliable index of its candidate's actual political philosophy, but it also "would make registered party affiliations significantly less meaningful in the Oklahoma primary election system." Memorandum Opinion 59. Oklahoma reasonably has concluded that opening the LPO's primary to all voters regardless of party affiliation would undermine the crucial role of political parties in the primary process. Cf. Jones, 530 U. S., at 574.

Second, Oklahoma's semiclosed primary system, by retaining the importance of party affiliation, aids in parties' electioneering and party-building efforts. "It is common experience that direct solicitation of party members — by mail, telephone, or face-to-face contact, and by the candidates themselves or by their active supporters — is part of any primary election campaign." Nader, supra, at 848. Yet parties' voter turnout efforts depend in large part on accurate voter registration rolls. See, e. g., Council of Alternative Political Parties v. State Div. of Elections, 344 N. J. Super. 225, 231-232, 781 A. 2d 1041, 1045 (2001) ("It is undisputed that the voter registration lists, with voter affiliation information, . . . provide essential information to the [party state committees] for other campaign and party-building activities, including canvassing and fundraising").

When voters are no longer required to disaffiliate before participating in other parties' primaries, voter registration rolls cease to be an accurate reflection of voters' political preferences. And without registration rolls that accurately *596 reflect likely or potential primary voters, parties risk expending precious resources to turn out party members who may have decided to cast their votes elsewhere. See Brief for State of South Dakota et al. as Amici Curiae 20-21. If encouraging citizens to vote is an important state interest, see Jones, supra, at 587 (KENNEDY, J., concurring), then Oklahoma is entitled to protect parties' ability to plan their primaries for a stable group of voters. Tr. of Oral Arg. 26.

Third, Oklahoma has an interest in preventing party raiding, or "the organized switching of blocs of voters from one party to another in order to manipulate the outcome of the other party's primary election." Anderson, 460 U. S., at 788-789, n. 9; Jones, supra, at 572. For example, if the outcome of the Democratic Party primary were not in doubt, Democrats might vote in the LPO primary for the candidate most likely to siphon off votes from the Republican candidate in the general election. Or a Democratic primary contender who senses defeat might launch a "sore loser" candidacy by defecting to the LPO primary, taking with him loyal Democratic voters, and thus undermining the Democratic Party in the general election.^[4]Storer, 415 U. S., at 735. Oklahoma has an interest in "temper[ing] the destabilizing effects" of precisely this sort of "party splintering and excessive factionalism." *597 Timmons, 520 U. S., at 367; cf. Davis v. Bandemer, 478 U. S. 109, 144-145 (1986) (O'CONNOR, J., concurring in judgment). Oklahoma's semiclosed primary system serves that interest by discouraging voters from temporarily defecting from another party to vote in the LPO primary. While the State's interest will not justify "unreasonably exclusionary restrictions," Timmons, 520 U. S., at 367, we have "repeatedly upheld reasonable, politically neutral regulations" like Oklahoma's semiclosed primary law, id., at 369 (internal quotation marks omitted).

III

Beyond their challenge to Oklahoma's semiclosed primary law, § 1-104, respondents have expanded their challenge before this Court to include other Oklahoma election laws. Respondents contend that several of the State's ballot access and voter registration laws, taken together, severely burden their associational rights by effectively preventing them from changing their party affiliations in advance of a primary election. Brief for Respondents 15-18 (discussing the joint operation of Okla. Stat. Ann., Tit. 26, §§ 1-108, 1-109, 1-110, 4-112, and 4-119 (West Supp. 2005)).

Though the LPO has unsuccessfully challenged one of these provisions before, see Rainbow Coalition of Okla. v. Oklahoma State Election Bd., 844 F. 2d 740 (CA10 1988) (rejecting First Amendment challenge by LPO and other political parties to Oklahoma's ballot access provision, § 1-108 (West 1981 and Supp. 1987)), respondents raise this argument for the first time in their brief on the merits to this Court. Before the District Court and the Court of Appeals, the only associational burden of which respondents complained was that imposed by § 1-104 (West 1997), i. e., the need to disaffiliate from one party in order to vote in another party's primary. See, e. g., Appellants' Opening Brief in No. 03-6058 (CA10), pp. 5, 8-10, 30 (challenging only § 1-104 as applied to respondents); Plaintiffs' Amended Trial Brief *598 9-25, Record Doc. 65 (same); Amended Complaint 6-9, id., Doc. 23 (same). As a result, there is virtually no evidence in the record on how other electoral regulations operate in tandem with § 1-104, whether these other laws actually burden respondents' associational rights, and whether these laws advance important or even compelling state interests. We ordinarily do not consider claims neither raised nor decided below, Cooper Industries, Inc. v. Aviall Services, Inc., 543 U. S. 157, 168-169 (2004) (citing Adarand Constructors, Inc. v. Mineta, 534 U. S. 103, 109 (2001) (per curiam)), and respondents have pointed to no unusual circumstances that would warrant considering other portions of Oklahoma's electoral code this late in the day, see Taylor v. Freeland & Kronz, 503 U. S. 638, 645-646 (1992). We therefore decline to consider this aspect of their challenge.

* * *

Oklahoma remains free to allow the LPO to invite registered voters of other parties to vote in its primary. But the Constitution leaves that choice to the democratic process, not to the courts. The judgment of the Court of Appeals is reversed, and the case is remanded for further proceedings.

It is so ordered.

JUSTICE O'CONNOR, with whom JUSTICE BREYER joins except as to Part III, concurring in part and concurring in the judgment.

I join the Court's opinion except for Part II-A. Although I agree with most of the Court's reasoning, I write separately to emphasize two points. First, I think respondents' claim implicates important associational interests, and I see no reason to minimize those interests to dispose of this case. Second, I agree with the Court that only Oklahoma's semiclosed primary law is properly before us, that standing alone it imposes only a modest, nondiscriminatory burden on respondents' associational rights, and that this burden is justified *599 by the State's legitimate regulatory interests. I note, however, that there are some grounds for concern that other state laws may unreasonably restrict voters' ability to change party registration so as to participate in the Libertarian Party of Oklahoma's (LPO) primary. A realistic assessment of regulatory burdens on associational rights would, in an appropriate case, require examination of the cumulative effects of the State's overall scheme governing primary elections; and any finding of a more severe burden would trigger more probing review of the justifications offered by the State.

I

Nearly every State in the Nation now mandates that political parties select their candidates for national or statewide office by means of primary elections. See Galderisi & Ezra, Congressional Primaries in Historical and Theoretical Context, in Congressional Primaries and the Politics of Representation 11, 17, and n. 34 (P. Galderisi, M. Ezra, & M. Lyons eds. 2001). Primaries constitute both a "`crucial juncture'" in the electoral process, California Democratic Party v. Jones, 530 U. S. 567, 575 (2000) (quoting Tashjian v. Republican Party of Conn., 479 U. S. 208, 216 (1986)), and a vital forum for expressive association among voters and political parties, see Kusper v. Pontikes, 414 U. S. 51, 58 (1973) ("[A] basic function of a political party is to select the candidates for public office to be offered to the voters at general elections[, and a] prime objective of most voters in associating themselves with a particular party must surely be to gain a voice in that selection process"). It is here that the parties invite voters to join in selecting their standard bearers. The outcome is pivotal, of course, for it dictates the range of choices available at — and often the presumptive winner of — the general election.

"No right is more precious in a free country than that of having a voice in the election of those who make the laws under which, as good citizens, we must live," Wesberry v. *600 Sanders, 376 U. S. 1, 17 (1964), and "[t]he right to associate with the political party of one's choice is an integral part of this basic constitutional freedom," Kusper, supra, at 57. The Court has repeatedly reaffirmed that the First and Fourteenth Amendments protect the rights of voters and parties to associate through primary elections. See, e. g., California Democratic Party, supra, at 574-575; Tashjian, supra, at 214; Kusper, supra, at 56-57. Indeed, constitutional protection of associational rights is especially important in this context because the aggregation of votes is, in some sense, the essence of the electoral process. To have a meaningful voice in this process, the individual voter must join together with like-minded others at the polls. And the choice of who will participate in selecting a party's candidate obviously plays a critical role in determining both the party's message and its prospects of success in the electoral contest. See California Democratic Party, supra, at 575; see also Democratic Party of United States v. Wisconsin ex rel. La Follette, 450 U. S. 107, 122 (1981) ("[T]he freedom to associate for the `common advancement of political beliefs' necessarily presupposes the freedom to identify the people who constitute the association" (quoting Kusper, supra, at 56)).

The plurality questions whether the LPO and voters registered with another party have any constitutionally cognizable interest in associating with one another through the LPO's primary. See ante, at 588-589. Its doubts on this point appear to stem from two implicit premises: first, that a voter forms a cognizable association with a political party only by registering with that party; and second, that a voter can only form a cognizable association with one party at a time. Neither of these premises is sound, in my view. As to the first, registration with a political party surely may signify an important personal commitment, which may be accompanied by faithful voting and even activism beyond the polls. But for many voters, registration serves principally as a mandatory (and perhaps even ministerial) prerequisite *601 to participation in the party's primaries. The act of casting a ballot in a given primary may, for both the voter and the party, constitute a form of association that is at least as important as the act of registering. See La Follette, supra, at 130, n. 2 (Powell, J., dissenting) ("[T]he act of voting in the Democratic primary fairly can be described as an act of affiliation with the Democratic Party"). The fact that voting is episodic does not, in my judgment, undermine its associational significance; it simply reflects the special character of the electoral process, which allows citizens to join together at regular intervals to shape government through the choice of public officials.

As to the question of dual associations, I fail to see why registration with one party should negate a voter's First Amendment interest in associating with a second party. We surely would not say, for instance, that a registered Republican or Democrat has no protected interest in associating with the Libertarian Party by attending meetings or making political contributions. The validity of voters' and parties' interests in dual associations seems particularly clear where minor parties are concerned. For example, a voter may have a longstanding affiliation with a major party that she wishes to maintain, but she may nevertheless have a substantial interest in associating with a minor party during particular election cycles or in elections for particular offices. The voter's refusal to disaffiliate from the major party may reflect her abiding commitment to that party (which is not necessarily inconsistent with her desire to associate with a second party), the objective costs of disaffiliation, see, e. g., infra, at 606-607, or both. The minor party, for its part, may have a significant interest in augmenting its voice in the political process by associating with sympathetic members of the major parties.

None of this is to suggest that the State does not have a superseding interest in restricting certain forms of association. We have never questioned, for example, the States' *602 authority to restrict voters' public registration to a single party or to limit each voter to participating in a single party's primary. But the fact that a State's regulatory authority may ultimately trump voters' or parties' associational interests in a particular context is no reason to dismiss the validity of those interests. As a more general matter, I question whether judicial inquiry into the genuineness, intensity, or duration of a given voter's association with a given party is a fruitful way to approach constitutional challenges to regulations like the one at issue here. Primary voting is an episodic and sometimes isolated act of association, but it is a vitally important one and should be entitled to some level of constitutional protection. Accordingly, where a party invites a voter to participate in its primary and the voter seeks to do so, we should begin with the premise that there are significant associational interests at stake. From this starting point, we can then ask to what extent and in what manner the State may justifiably restrict those interests.

II

As to the remainder of the constitutional analysis, I am substantially in accord with the Court's reasoning. Our constitutional system assigns the States broad authority to regulate the electoral process, and we have recognized that, "as a practical matter, there must be a substantial regulation of elections if they are to be fair and honest and if some sort of order, rather than chaos, is to accompany the democratic processes," Storer v. Brown, 415 U. S. 724, 730 (1974). We have sought to balance the associational interests of parties and voters against the States' regulatory interests through the flexible standard of review reaffirmed by the Court today. See ante, at 586-587. Under that standard, "the rigorousness of our inquiry into the propriety of a state election law depends upon the extent to which a challenged regulation burdens First and Fourteenth Amendment rights." Burdick v. Takushi, 504 U. S. 428, 434 (1992). Regulations *603 imposing severe burdens on associational rights must be narrowly tailored to advance a compelling government interest. Timmons v. Twin Cities Area New Party, 520 U. S. 351, 358 (1997). Regulations imposing lesser burdens are subject to less intensive scrutiny, and reasonable, nondiscriminatory restrictions ordinarily will be sustained if they serve important regulatory interests. Ibid.

This regime reflects the limited but important role of courts in reviewing electoral regulation. Although the State has a legitimate — and indeed critical — role to play in regulating elections, it must be recognized that it is not a wholly independent or neutral arbiter. Rather, the State is itself controlled by the political party or parties in power, which presumably have an incentive to shape the rules of the electoral game to their own benefit. Recognition of that basic reality need not render suspect most electoral regulations. Where the State imposes only reasonable and genuinely neutral restrictions on associational rights, there is no threat to the integrity of the electoral process and no apparent reason for judicial intervention. As such restrictions become more severe, however, and particularly where they have discriminatory effects, there is increasing cause for concern that those in power may be using electoral rules to erect barriers to electoral competition. In such cases, applying heightened scrutiny helps to ensure that such limitations are truly justified and that the State's asserted interests are not merely a pretext for exclusionary or anti-competitive restrictions.

Throughout the proceedings in the lower courts, respondents framed their suit as a facial challenge to Oklahoma's semiclosed primary law. The sum of their argument was that, by requiring voters to register either as Libertarians or Independents in order to participate in the LPO's primary, state law imposes a severe and unjustified burden on the LPO's and Oklahoma voters' associational rights. For the reasons explained by the Court, ante, at 597-598, that is the *604 only claim properly before us. Assuming (as I believe we must under the circumstances) that Oklahoma provides reasonable avenues for voters to reregister as Independents or Libertarians, I agree with the Court that the semiclosed primary law imposes only a modest and politically neutral burden on associational rights. The burden is not altogether trivial: A voter with a significant commitment to a major party (for example) must forfeit registration with that party in order to participate in the LPO primary in any given election cycle, and the LPO cannot define the bounds of the association as broadly as it would like. See post, at 610, and n. 1 (STEVENS, J., dissenting); see also supra, at 601 (discussing the interest in dual associations). But neither is it severe or discriminatory.

Oklahoma's semiclosed primary law simply requires that voters wishing to participate in the LPO's primary do what they would have to do in order to participate in any other party's primary. By providing a reasonably fixed party-related electoral base from the close of registration until the date of the vote, this requirement facilitates campaign planning. And assuming the availability of reasonable reregistration procedures, a party's inability to persuade a voter to disaffiliate from a rival party would suggest not the presence of anticompetitive regulatory restrictions, but rather the party's failure to win the voter's allegiance. The semiclosed primary law, standing alone, does not impose a significant obstacle to participation in the LPO's primary, nor does it indicate partisan self dealing or a lockup of the political process that would warrant heightened judicial scrutiny.

For essentially the reasons explained by the Court, see ante, at 593-597, I agree that Oklahoma has a legitimate interest in requiring voters to disaffiliate from one party before participating in another party's primary. On the record before us, I also agree that the State's regulatory interests are adequate to justify the limited burden the semiclosed primary law imposes on respondents' freedom of association. *605 And finally, I agree that this case is distinguishable from Tashjian. See ante, at 591-593. I joined the dissent in that case, and I think the Court's application of strict scrutiny there is difficult to square with the flexible standard of review articulated in our more recent cases, see supra, at 602-603. But Tashjian is entitled to respect under principles of stare decisis, and it can be fairly distinguished on the grounds that the closed primary law in that case imposed a greater burden on associational interests than does Oklahoma's semiclosed primary law, see ante, at 592, while the State's regulatory interests in Tashjian were weaker than they are here, compare ante, at 593-597, with Tashjian, 479 U. S., at 217-225.

III

In briefing and oral argument before this Court, respondents raise for the first time the claim that Oklahoma's semiclosed primary law severely burdens their associational rights not through the law's own operation, but rather because other state laws make it quite difficult for voters to reregister as Independents or Libertarians so as to participate in the LPO primary. See Brief for Respondents 12-24. Respondents characterize Oklahoma's regulatory scheme as follows.

Partisan primaries in Oklahoma are held on the last Tuesday in July of each even-numbered year. Okla. Stat. Ann., Tit. 26, § 1-102 (West Supp. 2005). To field a party candidate in an election, the LPO must obtain "recognized" party status. See ibid.; see also §§ 1-107, 5-104 (West 1997 and Supp. 2005). This requires it to submit, no later than May 1 of any even-numbered year (i. e., any election year), a petition with the signatures of registered voters equal to at least five percent of the total votes cast in the most recent gubernatorial or Presidential election. § 1-108 (West Supp. 2005). The State Election Board then has 30 days to determine whether the petition is sufficient. § 1-108(3). The LPO has attained recognized party status in this fashion in every Presidential *606 election year since 1980. However, unless the party's candidate receives at least 10 percent of the total votes cast for Governor or President in the general election (which no minor party has been able to do in any State in recent history), it loses recognized party status. § 1-109. To regain party status, the group must go through the petition process again. Ibid.

When a party loses its recognized status, as the LPO has after every general election in which it has participated, the affiliation of any voter registered with the party is changed to Independent. § 1-110. As the District Court noted, "it is highly likely that the ranks of independents, and, indeed, of registered Republicans and Democrats, contain numerous voters who sympathize with the LPO but who simply do not wish to go through the motions of re-registering every time they are purged from the rolls." Memorandum Opinion, Case No. CIV-00-1071-F (WD Okla., Jan. 24, 2003), App. to Pet. for Cert. A-48. And the Republican and Democratic Parties in Oklahoma, as it turns out, do not permit voters registered as Independents to participate in their primaries.

Most importantly, according to respondents, the deadline for changing party affiliation makes it quite difficult for the LPO to invite voters to reregister in order to participate in its primary. Assuming the LPO submits its petition for recognized party status on the May 1 deadline, the State has until May 31 to determine whether party status will be conferred. See Okla. Stat. Ann., Tit. 26, § 1-108 (West Supp. 2005). But in order to participate in the LPO primary, a voter registered with another party must change her party affiliation to Independent or Libertarian no later than June 1. See § 4-119. Moreover, no candidate for office is permitted officially to declare her candidacy with the State Election Board until the period between the first Monday in June and the next succeeding Wednesday. § 5-110.

If this characterization of state law is accurate, a registered Democrat or Republican sympathetic to the LPO or to *607 an LPO candidate in a given election year would seem to face a genuine dilemma. On the one hand, she may stick with her major party registration and forfeit the opportunity to participate in the LPO primary. Alternatively, she may reregister as a Libertarian or Independent, thus forfeiting her opportunity to participate in the major party primary, though no candidate will have officially declared yet and the voter may not yet know whether the LPO will even be permitted to conduct a primary. Moreover, she must make this choice roughly eight weeks before the primaries, at a time when most voters have not yet even tuned in to the election, much less decided upon a candidate. See California Democratic Party, 530 U. S., at 586 (KENNEDY, J., concurring). That might pose a special difficulty for voters attracted to minor party candidates, for whom support may not coalesce until comparatively late in the election cycle. See Anderson v. Celebrezze, 460 U. S. 780, 791-792 (1983) (discussing emergence of independent candidacies late in the election cycle).

Throughout the proceedings in the lower courts, which included a full bench trial before the District Court, respondents made no attempt to challenge these other electoral requirements or to argue that they were relevant to respondents' challenge to the semiclosed primary law. The lower courts, accordingly, gave little or no consideration to how these various regulations interrelate or operate in practice, nor did the State seek to justify them. Given this posture, I agree with the Court that it would be neither proper nor prudent for us to rule on the reformulated claim that respondents now urge. See ante, at 597-598.

Nevertheless, respondents' allegations are troubling, and, if they had been properly raised, the Court would want to examine the cumulative burdens imposed by the overall scheme of electoral regulations upon the rights of voters and parties to associate through primary elections. A panoply of regulations, each apparently defensible when considered alone, may nevertheless have the combined effect of severely *608 restricting participation and competition. Even if each part of a regulatory regime might be upheld if challenged separately, one or another of these parts might have to fall if the overall scheme unreasonably curtails associational freedoms. Oklahoma's requirement that a voter register as an Independent or a Libertarian in order to participate in the LPO's primary is not itself unduly onerous; but that is true only to the extent that the State provides reasonable avenues through which a voter can change her registration status. The State's regulations governing changes in party affiliation are not properly before us now. But if it were shown, in an appropriate case, that such regulations imposed a weighty or discriminatory restriction on voters' ability to participate in the LPO's or some other party's primary, then more probing scrutiny of the State's justifications would be required.

JUSTICE STEVENS, with whom JUSTICE GINSBURG joins, and with whom JUSTICE SOUTER joins as to Parts I, II, and III, dissenting.

The Court's decision today diminishes the value of two important rights protected by the First Amendment: the individual citizen's right to vote for the candidate of her choice and a political party's right to define its own mission. No one would contend that a citizen's membership in either the Republican or the Democratic Party could disqualify her from attending political functions sponsored by another party, or from voting for a third party's candidate in a general election. If a third party invites her to participate in its primary election, her right to support the candidate of her choice merits constitutional protection, whether she elects to make a speech, to donate funds, or to cast a ballot. The importance of vindicating that individual right far outweighs any public interest in punishing registered Republicans or Democrats for acts of disloyalty. The balance becomes even more lopsided when the individual right is *609 reinforced by the right of the Libertarian Party of Oklahoma (LPO) to associate with willing voters.

In concluding that the State's interests override those important values, the Court focuses on interests that are not legitimate. States do not have a valid interest in manipulating the outcome of elections, in protecting the major parties from competition, or in stunting the growth of new parties. While States do have a valid interest in conducting orderly elections and in encouraging the maximum participation of voters, neither of these interests overrides (or, indeed, even conflicts with) the valid interests of both the LPO and the voters who wish to participate in its primary.

In the final analysis, this case is simple. Occasionally, a political party's interest in defining its platform and its procedures for selecting and supporting its candidates conflicts with the voters' interest in participating in the selection of their elected representatives. If those values do conflict, we may be faced with difficult choices. But when, as in this case, those values reinforce one another a decision should be easy. Oklahoma has enacted a statute that impairs both; it denies a party the right to invite willing voters to participate in its primary elections. I would therefore affirm the Court of Appeals' judgment.

I

In rejecting the individual respondents' claims, the majority focuses on their associational interests. While the voters in this case certainly have an interest in associating with the LPO, they are primarily interested in voting for a particular candidate, who happens to be in the LPO. Indeed, I think we have lost sight of the principal purpose of a primary: to nominate a candidate for office. Cf. Burdick v. Takushi, 504 U. S. 428, 445 (1992) (KENNEDY, J., dissenting) ("[T]he purpose of casting, counting, and recording votes is to elect public officials, not to serve as a general forum for political expression").

*610 Because our recent cases have focused on the associational interest of voters, rather than the right to vote itself, it is important to identify three basic precepts. First, it is clear that the right to vote includes the right to vote in a primary election. See United States v. Classic, 313 U. S. 299, 318 (1941); Terry v. Adams, 345 U.S. 461 (1953). When the State makes the primary an "integral part of the procedure of choice," every eligible citizen's right to vote should receive the same protection as in the general election. Classic, 313 U. S., at 318; see also, e. g., Gray v. Sanders, 372 U. S. 368 (1963) (invalidating primary system that diluted individual's vote in a primary). Second, the right to vote, whether in the primary or the general election, is the right to vote "for the candidate of one's choice." Reynolds v. Sims, 377 U. S. 533, 555 (1964). Finally, in assessing burdens on that right — burdens that are not limited to absolute denial of the right — we should focus on the realities of the situation, not on empty formalism. See Classic, 313 U. S., at 313 (identifying "the practical operation of the primary law"); Terry, 345 U. S., at 469-470 (noting that the Jaybird primary is "the only effective part" of the election process and examining "[t]he effect of the whole procedure" in determining whether the scheme violated the Fifteenth Amendment).

Here, the impact of the Oklahoma statute on the voters' right to vote for the candidate of their choosing is not a mere "burden"; it is a prohibition.^[1] By virtue of the fact that their preferred candidate is a member of a different party, respondents are absolutely precluded from voting for him or her in the primary election. It is not an answer that the *611 voters could participate in another primary (i. e., the primary for the party with which they are registered) since the individual for whom they wish to vote is not a candidate in that primary. If the so-called "white primary" cases make anything clear, it is that the denial of the right to vote cannot be cured by the ability to participate in a subsequent or different election. Just as the "only election that has counted" in Terry, 345 U. S., at 469, was the Jaybird primary, since it was there that the public official was selected in any meaningful sense, the only primary that counts here is the one in which the candidate respondents want to vote for is actually running. See Burdick, 504 U. S., at 442 (KENNEDY, J., dissenting) ("Because [petitioner] could not write in the name of a candidate he preferred, he had no way to cast a meaningful vote").

This is not to say that voters have an absolute right to participate in whatever primary they desire. For instance, the parties themselves have a strong associational interest in determining which individuals may vote in their primaries, and that interest will normally outweigh the interest of the uninvited voter.^[2] But in the ordinary case the State simply has no interest in classifying voters by their political party and in limiting the elections in which voters may participate as a result of that classification. Just as we held in Reynolds that all voters of a State stand in the same relation to the State regardless of where they live, and that the State must thus not make their vote count more or less depending *612 upon that factor, 377 U. S., at 565, so too do citizens stand in the same relation to the State regardless of the political party to which they belong. The State may thus not deny them participation in a primary of a party that seeks their participation absent a state interest of overriding importance.

II

In addition to burdening the individual respondent's right to vote, the Oklahoma scheme places a heavy burden on the LPO's associational rights. While Oklahoma permits independent voters to participate in the LPO's primary elections, it refuses to allow registered Republicans or Democrats to do so. That refusal has a direct impact on the LPO's selection of candidates for public office, the importance of which cannot be overstated. A primary election plays a critical role in enabling a party to disseminate its message to the public. California Democratic Party v. Jones, 530 U. S. 567, 575 (2000). It is through its candidates that a party is able to give voice to its political views, to engage other candidates on important issues of the day, and to affect change in the government of our society. Our cases "vigorously affirm the special place the First Amendment reserves for, and the special protection it accords, the process by which a political party `select[s] a standard bearer who best represents the party's ideologies and preferences.'" Ibid. (quoting Eu v. San Francisco County Democratic Central Comm., 489 U. S. 214, 224 (1989)).

The Oklahoma statute prohibits the LPO from associating with all of the voters it believes will best enable it to select a viable candidate. The ability to select those individuals with whom to associate is, of course, at the core of the First Amendment and goes to the heart of the associational interest itself. "Freedom of association means not only that an individual voter has the right to associate with the political party of her choice, but also that a political party has a right to identify the people who constitute the association. . . ." *613 Ibid. (internal quotation marks and citations omitted). See also Democratic Party of United States v. Wisconsin ex rel. La Follette, 450 U. S. 107, 122 (1981). While Libertarians can undoubtedly associate with Democrats and Republicans in other ways and at other times, the Oklahoma statute "limits the Party's associational opportunities at the crucial juncture at which the appeal to common principles may be translated into concerted action, and hence to political power in the community." Tashjian v. Republican Party of Conn., 479 U. S. 208, 216 (1986).

In concluding that the Oklahoma statute is constitutional, the majority argues that associational interests between the LPO and registered members of other parties are either nonexistent or not heavily burdened by the Oklahoma scheme. The plurality relies on a single footnote in Jones to show that there are no associational interests between the LPO and registered Republicans and Democrats. See ante, at 588-589 (citing 530 U. S., at 573-574, n. 5). In Jones, of course, the political parties did not want voters of other parties participating in their primaries; the putative associational interest in this case, in which the LPO is actively courting voters of other parties, simply did not exist. More importantly, our decision in Tashjian rejected these arguments.

In Tashjian we held that the State could not prohibit Republicans from inviting voters who were not registered with a political party to participate in the Republican primary. We recognized that "[t]he Party's attempt to broaden the base of public participation in and support for its activities is conduct undeniably central to the exercise of the right of association." 479 U. S., at 214. Importantly, we rejected the notion that the associational interest was somehow diminished because the voters the party sought to include were not formally registered as Republicans. Id., at 215 ("[C]onsidered from the standpoint of the Party itself, the act of formal enrollment or public affiliation with the Party is merely one element in the continuum of participation in *614 Party affairs, and need not be in any sense the most important"). We reasoned that a State could not prohibit independents from contributing financial support to a Republican candidate or from participating in the party's events; it would be anomalous if it were able to prohibit participation by independents in the "`basic function'" of the party. Id., at 216. Because of the importance of those interests, we carefully examined the interests asserted by the State, and finding them lacking, struck down the prohibition on independents' participation in the Republican primary.

Virtually identical interests are at stake in this case. It is the LPO's belief that attracting a more diverse group of voters in its primary would enable it to select a more mainstream candidate who would be more viable in the general election. Like the Republicans in Tashjian, the LPO is cognizant of the fact that in order to enjoy success at the voting booth it must have support from voters who identify themselves as independents, Republicans, or Democrats.

The LPO's desire to include Democrats and Republicans is undoubtedly informed by the fact that, given the stringent requirements of Oklahoma law, the LPO ceases to become a formally recognized party after each election cycle, and its members automatically revert to being independents.^[3] Because the LPO routinely loses its status as a recognized party, many voters who might otherwise register as Libertarians instead register as Democrats or Republicans.^[4] Thus, the LPO's interest in inviting registered Republicans *615 and Democrats to participate in the selection of its standard bearer has even greater force than did the Republican Party's desire to invite independents to associate with it in Tashjian.

III

As justification for the State's abridgment of the constitutionally protected interests asserted by the LPO and the voters, the majority relies on countervailing state interests that are either irrelevant or insignificant. Neither separately nor in the aggregate do these interests support the Court's decision.

First, the Court makes the remarkable suggestion that by opening up its primary to Democrats and Republicans, the LPO will be saddled with so many nonlibertarian voters that the ultimate candidate will not be, in any sense, "libertarian." See ante, at 594.^[5] But the LPO is seeking the crossover voting of Republicans and Democrats. Rightly or wrongly, the LPO feels that the best way to produce a viable candidate is to invite voters from other parties to participate in its primary. That may dilute what the Court believes to be the core of the Libertarian philosophy, but it is no business of the State to tell a political party what its message should be, how it should select its candidates, or how it should form coalitions to ensure electoral success. See Jones, 530 U. S., at 581-582 (rejecting state interests in producing candidates that are more centrist than the nominee the party would have selected absent the blanket primary).^[6]

*616 Second, the majority expresses concern that crossover voting may create voter confusion. This paternalistic concern is belied by the District Court's finding that no significant voter confusion would occur. App. to Pet. for Cert. A-43 (noting that "very simple rules for voting eligibility can be posted at polling places when the primary and runoff elections are conducted").

Third, the majority suggests that crossover voting will impair the State's interest in properly classifying candidates and voters. As an empirical matter, a crossover voter may have a lesser commitment to the party with which he is registered if he votes in another party's primary. Nevertheless, the State does not have a valid interest in defining what it means to be a Republican or a Democrat, or in attempting to ensure the political orthodoxy of party members simply for the convenience of those parties. Cf. West Virginia Bd. of Ed. v. Barnette, 319 U. S. 624, 642 (1943) ("If there is any fixed star in our constitutional constellation, it is that no official, high or petty, can prescribe what shall be orthodox in politics, nationalism, religion, or other matters of opinion. . ."). Even if participation in the LPO's primary causes a voter to be a less committed "Democrat" or "Republican" (a proposition I reject^[7]), the dilution of that commitment does not justify abridgment of the fundamental rights at issue in this case. While party identity is important in *617 our political system, it should not be immunized from the risk of change.^[8]

Fourth, the majority argues that opening up the LPO primary to members of the Republican and Democratic Parties might interfere with electioneering and party-building efforts. It is clear, of course, that the majority here is concerned only with the Democratic and Republican Parties, since party building is precisely what the LPO is attempting to accomplish. Nevertheless, that concern is misplaced. Even if, as the majority claims, the Republican and Democratic voter rolls, mailing lists, and phone banks are not as accurate as they would otherwise be,^[9] the administrative inconvenience of the major parties does not outweigh the right to vote or the associational interests of those voters and the LPO. At its core, this argument is based on a fear that the LPO might be successful in convincing Democratic or Republican voters to participate more fully in the LPO. Far from being a compelling interest, it is an impermissible one. Timmons v. Twin Cities Area New Party, 520 U. S. 351, 367 *618 (1997) (State may not "completely insulate the two-party system from minor parties' or independent candidates' competition and influence").

Finally, the majority warns against the possibility of raiding, ante, at 596, by which voters of another party maliciously vote in a primary in order to change the outcome of the primary, either to nominate a particularly weak candidate, a "sore-loser" candidate, or a candidate who would siphon votes from another party. The District Court, whose factual findings are entitled to substantial deference, found as a factual and legal matter that the State's argument concerning raiding was "unpersuasive." App. to Pet. for Cert. A-61.

Even if raiding were a possibility, however, the state interests are remote. The possibility of harm to the LPO itself is insufficient to overcome the LPO's associational rights. See Eu, 489 U. S., at 227-228 ("[E]ven if a ban on endorsements saves a political party from pursuing self-destructive acts, that would not justify a State substituting its judgment for that of the party"). If the LPO is willing to take the risk that its party may be "hijacked" by individuals who hold views opposite to their own, the State has little interest in second-guessing the LPO's decision.

With respect to the possibility that Democratic or Republican voters might raid the LPO to the detriment of their own or another party, neither the State nor the majority has identified any evidence that voters are sufficiently organized to achieve such a targeted result.^[10] Such speculation is not, in *619 my view, sufficient to override the real and acknowledged interest of the LPO and the voters who wish to participate in its primary. See Timmons, 520 U. S., at 375 (STEVENS, J., dissenting) (citing Eu, 489 U. S., at 226; Anderson v. Celebrezze, 460 U. S. 780, 789 (1983); and Norman v. Reed, 502 U. S. 279, 288-289 (1992)).^[11]

In the end, the balance of interests clearly favors the LPO and those voters who wish to participate in its primary. The associational interests asserted — the right to select a standard bearer that the party thinks has the best chance of success, the ability to associate at the crucial juncture of selecting a candidate, and the desire to reach out to voters of other parties — are substantial and undoubtedly burdened by Oklahoma's statutory scheme. Any doubt about that fact is clearly answered by Tashjian. On the other side, the interests asserted by the State are either entirely speculative or simply protectionist measures that benefit the parties in power. No matter what the standard, they simply do not outweigh the interests of the LPO and its voters.

IV

The Libertarian Party of Oklahoma is not the only loser in this litigation. Other minor parties and voters who have primary allegiance to one party but sometimes switch their support to rival candidates are also harmed by this decision. In my judgment, however, the real losers include all participants in the political market. Decisions that give undue deference *620 to the interest in preserving the two-party system,^[12] like decisions that encourage partisan gerrymandering,^[13] enhance the likelihood that so-called "safe districts" will play an increasingly predominant role in the electoral process. Primary elections are already replacing general elections as the most common method of actually determining the composition of our legislative bodies. The trend can only increase the bitter partisanship that has already poisoned some of those bodies that once provided inspiring examples of courteous adversary debate and deliberation.

The decision in this case, like the misguided decisions in Timmons, 520 U. S. 351, and Jones, 530 U. S. 567, attaches overriding importance to the interest in preserving the two-party system. In my view, there is over a century of experience demonstrating that the two major parties are fully capable of maintaining their own positions of dominance in the political marketplace without any special assistance from the state governments that they dominate or from this Court. Whenever they receive special advantages, the offsetting harm to independent voters may be far more significant than the majority recognizes.

In Anderson, 460 U. S. 780, we considered the impact of early filing dates on small political parties and independent candidates. Commenting on election laws that disadvantage independents, we noted:

"By limiting the opportunities of independent-minded voters to associate in the electoral arena to enhance their political effectiveness as a group, such restrictions threaten to reduce diversity and competition in the marketplace of ideas. Historically political figures outside *621 the two major parties have been fertile sources of new ideas and new programs; many of their challenges to the status quo have in time made their way into the political mainstream. In short, the primary values protected by the First Amendment — `a profound national commitment to the principle that debate on public issues should be uninhibited, robust, and wide-open,' — are served when election campaigns are not monopolized by the existing political parties." Id., at 794 (citations omitted).

Because the Court's holding today has little to support it other than a naked interest in protecting the two major parties, I respectfully dissent.

NOTES

[*] A brief of amici curiae urging reversal was filed for the State of South Dakota et al. by Lawrence E. Long, Attorney General of South Dakota, Craig M. Eichstadt, Deputy Attorney General, and Gene C. Schaerr, and by the Attorneys General for their respective States as follows: J. Joseph Curran, Jr., of Maryland, Thomas F. Reilly of Massachusetts, Kelly A. Ayotte of New Hampshire, Patricia A. Madrid of New Mexico, Roy Cooper of North Carolina, Mark Shurtleff of Utah, and Darrell V. McGraw, Jr., of West Virginia.

A brief of amicus curiae urging affirmance was filed for the Coalition for Free and Open Elections by Richard Shepard.

[1] Ariz. Rev. Stat. Ann. § 16-241(A) (West 1996); Cal. Elec. Code Ann. § 13102 (West 2003); Colo. Rev. Stat. § 1-3-101(1) (Lexis 2004); Conn. Gen. Stat. § 9-431(a) (2005); Del. Code Ann., Tit. 15, § 3110 (Lexis 1999); Fla. Stat. § 101.021 (2003); Iowa Code §§ 43.38, 43.42 (2003); Kan. Stat. Ann. § 25-4502 (2000); Ky. Rev. Stat. Ann. § 116.055 (Lexis 2004); La. Stat. Ann. § 18:1280.25 (West Supp. 2005); Mass. Gen. Laws Ann., ch. 53, § 37 (West Supp. 2005); Neb. Rev. Stat. § 32-312 (2004); Nev. Rev. Stat. § 293.287 (2003); N. H. Rev. Stat. Ann. § 659:14 (West 1996); N. J. Stat. Ann. § 19:23-45.1 (West Supp. 2004); N. M. Stat. Ann. § 1-12-7 (1995); N. Y. Elec. Law Ann. § 1-104.9 (West 2004); N. C. Gen. Stat. § 163-59 (Lexis 2003); Pa. Stat. Ann., Tit. 25, § 292 (Purdon 1994); R. I. Gen. Laws §§ 17-9.1-24, 17-15-24 (Lexis 2003); S. D. Codified Laws § 12-6-26 (West 2004); W. Va. Code § 3-1-35 (Lexis 2002); Wyo. Stat. § 22-5-212 (Lexis 1977-2003).

[2] Respondents argue, for the first time before this Court, that Oklahoma election statutes other than § 1-104 make it difficult for voters to disaffiliate from their parties of first choice and register as Libertarians or Independents (either of which would allow them to vote in the LPO primary). Brief for Respondents 13-19. For reasons we explain fully in Part III, we decline to consider this aspect of respondents' challenge. See infra, at 597-598.

[3] Respondents who are members of the Republican and Democratic Parties alleged before the District Court that they wished to have the right to participate in the 2000 LPO primary. See Amended Complaint 4, Record Doc. 23; Complaint 3, id., Doc. 1. The only evidence respondents submitted on this point was a pair of affidavits from respondents Mary Burnett (a registered Republican) and Floyd Turner (a registered Democrat), asserting that each might have wished to vote in the 2000 LPO primary. See Plaintiffs' Motion for Preliminary Injunction, id., Doc. 9 (attached affidavits). Based on Turner's affidavit, the parties stipulated that there were "a number of voters" "registered in political parties other than the [LPO] who wish[ed] to vote" in the 2000 LPO primary. See Supplemental Joint Stipulations of Fact ¶ 32, id., Doc. 17. Respondents have never claimed that they are prevented from associating with the LPO in any way, except that they are unable to vote in the LPO's primary and run-off elections.

[4] To be most effective, a spurned candidate would have to defect in advance of the primary election. Before a candidate may file for nomination by a political party to any state or county office in Oklahoma, generally the candidate must have been a registered member of the party for six months prior to filing. See Okla. Stat. Ann., Tit. 26, § 5-105(A) (West 1997). However, the registration period is only 15 days for candidates from parties, like the LPO, whose lack of electoral support means that they must regularly petition to be recognized as political parties. Ibid.; see also §§ 1-108, 1-109 (West Supp. 2005) (Oklahoma's ballot access requirements). But even though candidates may defect up to two weeks before the primary, registered Republican and Democratic voters may not change their party affiliation after June 1, roughly eight weeks before the primary. See § 4-119; see also § 1-102 (setting primary on last Tuesday of July).

[1] It is not enough that registered members of other parties may simply change their registration. See ante, at 590-591 (plurality opinion). Changing one's political party is not simply a matter of filing a form with the State; for many individuals it can be a significant decision. A view that party membership is merely a label demeans for many the personal significance of party identification and illustrates what little weight the majority actually gives to the associational interests in this case.

[2] The voters' interest may still prevail if, as was the case in Terry v. Adams, 345 U. S. 461 (1953), and Smith v. Allwright, 321 U. S. 649 (1944), the party primary is the de facto election. In part because of this Court's refusal to intervene in political gerrymandering cases, Davis v. Bandemer, 478 U. S. 109 (1986), an increasing number of districts are becoming "safe districts" in which one party effectively controls the outcome of the election. See, e. g., Courtney, Redistricting: What the United States Can Learn from Canada, 3 Election L. J. 488 (2004) (concluding that 400 of the 435 Members of the House of Representatives were elected in safe districts in the 2002 election, 81 of whom ran unopposed).

[3] See Okla. Stat. Ann., Tit. 26, § 1-109 (West Supp. 2005) (requiring that a party's nominee for Governor, President, or Vice President receive 10% of the vote in a general election for the party to maintain its status).

[4] See App. to Pet. for Cert. A-48 (District Court recognizing that "it is highly likely that the ranks of independents and, indeed, of registered Republicans and Democrats, contain numerous voters who sympathize with the LPO but who simply do not wish to go through the motions of re-registering every time they are purged from the rolls").

[5] Of course, as the majority recognizes, ante, at 594, since the number of independent voters overwhelms the number of registered-LPO voters, that is already the case.

[6] See also Democratic Party of United States v. Wisconsin ex rel. La Follette, 450 U. S. 107, 123-124 (1981) (State may not substitute its own judgment for that of the party); Jones, 530 U. S., at 587 (KENNEDY, J., concurring) ("A political party might be better served by allowing blanket primaries as a means of nominating candidates with broader appeal. Under the First Amendment's guarantee of speech through free association, however, this is an issue for the party to resolve, not for the State" (emphasis added)). Such coalition building, and reaching out to other groups to ensure a candidate gets elected, is a vital part of the political process. Cf. Colorado Republican Federal Campaign Comm. v. Federal Election Comm'n, 518 U. S. 604, 622-623 (1996) (citing W. Keefe, Parties, Politics, and Public Policy in America 59-74 (5th ed. 1988)).

[7] Allowing a potential crossover voter to vote in the LPO primary would not change the level of commitment he has toward his party of registration; it would simply give him an outlet to express the views he already holds.

[8] If, of course, States were able to protect the incumbent parties in the name of protecting the stability of the two-party system in general, we might still have the Federalists, the Anti-federalists, or the Whigs. See generally J. Aldrich, Why Parties? The Origin and Transformation of Political Parties in America (1995). In any event, we would not have the evolution of thought or policies that are occasioned through the change of political parties. While no such change has occurred in recent memory, that is no reason to ossify the status quo.

[9] The majority's argument is that voters who would otherwise vote in the Republican or Democratic primaries would vote in the LPO primary, and that the Democratic and Republican lists would not be an accurate indicator of who is likely to vote in those primaries, and of which voters to spend party resources on. First, I find it doubtful that those voters who vote in the LPO primary would have voted in the Democratic or Republican primary; rather, they probably would not have been sufficiently motivated to vote at all. Further, this would actually give Republicans and Democrats additional information as to which of their voters have Libertarian leanings.

[10] To change the outcome of an election in a way that would benefit their own party, voters would have to be relatively certain that their preferred candidate in their own primary would win that primary and to vote in the LPO primary for a previously agreed-on candidate who is opposed to their own ideological preferences. Given that voters typically do not focus on an election until several days or weeks before an election, this prospect is unlikely. See California Democratic Party v. Jones, 530 U. S. 567, 586 (2000) (KENNEDY, J., concurring). Further, one would have expected to see some evidence of this in States where it is relatively easy to switch parties close to a primary.

[11] The flimsy character of the state interests in this case confirms my view that today's decision rests primarily on a desire to protect the two-party system. In Jones, the Court concluded that the associational interests of the parties trumped state interests that were much more compelling than those asserted in this case. Here, by contrast, where the associational interests are being asserted by a minor party rather than by one of the dominant parties, the Court has reversed course and rejected those associational interests as insubstantial compared to the interests asserted by the State.

[12] Examples are cases permitting lengthy registration periods, Rosario v. Rockefeller, 410 U. S. 752 (1973), and cases approving bans on fusion candidates, Timmons v. Twin Cities Area New Party, 520 U. S. 351 (1997); and write-in candidates, Burdick v. Takushi, 504 U. S. 428 (1992).

[13] See, e. g., Vieth v. Jubelirer, 541 U. S. 267 (2004); Davis v. Bandemer, 478 U. S. 109 (1986).