FLETCHER, Circuit Judge:
In 1996 California voters passed Proposition 198, which converted the State’s primary election from a closed to a blanket primary in which voters may vote for any candidate regardless of the voter’s or candidate’s party affiliation. Appellants, the California Democratic, Republican, Libertarian, and Peace and Freedom Parties, together with other individuals, opponents of Proposition 198, seek reversal of the district court’s judgment upholding its constitutionality. California Democratic Party v. Jones, 984 F.Supp. 1288 (E.D.Cal.1997). We have jurisdiction under 28 U.S.C. § 1291, and we affirm.
STANDARD OF REVIEW
We review the constitutionality of a statute de novo. Crawford v. Lungren, 96 F.3d 380, 384 (9th Cir.1996), cert. denied, 620 U.S. 1117, 117 S.Ct. 1249, 137 L.Ed.2d 330 (1997). A district court’s determinations on questions of law and on mixed questions of law and fact that implicate constitutional rights are also reviewed de novo. Neal v. Shimoda, 131. F.3d 818, 823 (9th Cir.1997); Jacobsen v. United States Postal Serv., 993 F.2d 649, 663 (9th Cir.1992). We have undertaken that review.
CONCLUSION
We have reviewed the briefs to this court, the record made before the district court, and the careful, detailed, and eloquent opinion of the district court. Because we concur in it in every respect, we have elected to adopt it as the opinion of our court. We attach it as an appendix. It also may be found at 984 F.Supp. 1288 (E.D.Cal.1997) sub nom. California Democratic Party v. Jones. [648]*648The judgment of the district court is AFFIRMED.
ATTACHMENT
APPENDIX
CALIFORNIA DEMOCRATIC PARTY, Art Torres, Kathy Bowler, Paul Jorjo-rian, Peace and Freedom Party, C.T. Weber, Libertarian Party of California and Gail Lightfoot, Plaintiffs, v. Bill JONES, Secretary of State of the State of California, Defendant. California Republican Party, Michael Schroeder, Shawn Steel and Donna Shalansky, Intervenors/Plaintiffs, Californians for an Open Primary, Inter-venor/Defendant.
No. CIV. S-96-2038 DFL.
United States District Court, E.D. California.
Nov. 17, 1997.
As Corrected Nov. 24, 1997.
MEMORANDUM OF OPINION AND FINDINGS OF FACT AND CONCLUSIONS OF LAW
LEVI, District Judge.
In March of 1996, the people of the State of California, by a wide margin, adopted Proposition 198, an initiative statute known as the Open Primary Act. Proposition 198 converts the State’s primary election from a closed to an open or blanket primary in which voters may vote for any candidate regardless of the candidate’s or the voter’s party affiliation. In a blanket primary under Proposition 198, one ballot will be prepared at the primary election for all voters just as in the general election. Because a blanket primary permits voters to vote in the primary election of a party without being registered in that party, the political parties and party officials who bring this action contend that Proposition 198 violates their right of association guaranteed by the First Amendment of the Constitution.
The court concludes that Proposition 198 withstands this constitutional challenge.
I.
Under Proposition 198 “All persons entitled to vote, including those not affiliated with any political party, shall have the right to vote ... at any election in which they are qualified to vote, for any candidate regardless of the candidate’s political affiliation.” 1 By this language and the conforming amendments to the State Election Code, Proposition 198 changed the California primary from a closed to an open or blanket system.2 As the Legislative Analyst explained in the ballot pamphlet distributed to voters before the election, under the closed system “[i]n order to vote in primary elections for partisan offices, a voter must have identified a political party affiliation when registering to vote and can vote only for candidates of that party.” Defs.’ Ex. A at 5. Thus, for example, in the closed primary a registered Democratic Party voter receives a ballot that includes only candidates competing for the Democratic Party nomination. Only registered Democrats could vote for these candidates, and the winner of the primary would be the Democratic Party’s candidate at the general election.3 By contrast, the Legislative Analyst [649]*649explained, in the blanket primary instituted by Proposition 198:
[A]ll persons who are entitled to vote in primary elections, including those not affiliated with a political party [are allowed to] vote for any candidate regardless of the candidate’s political party affiliation. Thus, voters in primary elections would be allowed to vote for candidates across political party lines. Furthermore, the initiative provides that county elections officials prepare only one ballot for all voters. The candidates for an office would be listed randomly on the ballot and not grouped by political party affiliation. The candidate of each political party who receives the most votes for a state elective office becomes the nominee of that party at the next general election.
Defs.’ Ex. A at 5. Therefore, in the blanket primary after Proposition 198, a registered Democratic Party voter, for example, could vote for a Republican Party candidate for Governor and a Democratic Party candidate for Assembly, and so on down the ballot. An independent voter could do the same.4 Through such cross-over voting, it is possible under the blanket primary for independent voters and voters registered in one party to participate in the primary election of another party.
In the statement of support included in the ballot pamphlet, proponents of Proposition 198 argued that the closed primary “favors the election of party hard-liners, contributes to legislative gridlock, and stacks the deck against more moderate problem-solvers.” The open primary would permit voters to vote “for the best candidate for each office, regardless of party affiliation” thereby giving voters greater choice. In addition, according to proponents, the blanket primary would increase voter participation in the primary election, restore healthy competition particularly in “safe” legislative districts in which one party clearly dominates, make elected officials more responsive to voters as opposed to party officials, reduce the power of special interest groups, and strengthen the political parties by assuring the nomination of candidates with broader bases of support. Opponents of Proposition 198 contended that it would work an unconstitutional interference with the right of the parties to choose their nominees: “Allowing members of one party a large voice in choosing another party’s nominee — which Proposition 198 would do — is like letting UCLA’s football team choose USC’s head coach!”5 According to the opponents, the existing closed primary provided a “real choice” among candidates of different parties while a blanket primary would be “an invitation to political mischief’ by special interests and political consultants. Defs.’ Ex. A at 6-7.6
At the election, Proposition 198 was adopted by a convincing margin of those voting — 59.51 percent (3,340,642 votes) to 40.49 percent (2,273,064 votes). According to reliable exit polls, Proposition 198 was supported by 61 percent of Democrats, 57 percent of Republicans, and 69 percent of Independents. The measure was supported by a majority of every demographic subgroup, including groupings by sex, age, race, education, political ideology, income, religion and party identification. Defs.’ Exs. B & D. Proposition 198 commanded majorities in every county in California and in 75 of the 80 Assembly districts. Defs.’ Ex. I. Although the turnout by California’s six minor parties was too small to sample, given the broad public support for Proposition 198 among all groups of every sort, it is likely that a majority of the members of these parties also favored the initiative. R.T. at 847 (Testimo[650]*650ny of Thomas Quinn). Moreover, the vote on Proposition 198 reflects the long-standing support for an open primary among California voters. According to Mervin Field, an eminent California political polling expert, Californians have favored an open primary by wide margins since at least 1983. See Defs.’ Exs. L-2 & L-5.7
By passing Proposition 198, California joined a small group of states, consisting of Alaska, Louisiana, and Washington, who also have blanket or nonpartisan primary elections. The other states have either a closed or an open primary, or some variant of the closed or open primary.8 In the classic closed primary, such as California had prior to Proposition 198, participation in a party’s primary election is limited to voters who have registered as members of that party a specified period of time prior to the primary election. Fifteen states have closed primaries of this kind. See Defs.’ Ex. E, Gerber Report, at 2-4. Eight additional states have a “semi-closed” primary: participation in a party’s primary is limited to those who have registered as party members and independents. Id. Participation of registered members of other parties is not permitted.
In an open primary, a registered voter may request, on election day, the ballot for any party’s primary in which the voter intends to vote, whether or not the voter previously has registered as a member of that party; however, the voter may only vote in one party’s primary election. Twenty-one states have an open primary in this sense of the term. M9 Finally, four states have blanket or nonpartisan primaries. The blanket primary is a version of the open primary in which the voter is not limited to the ballot of any single party. All voters receive the same ballot, and a voter is not limited to the candidates of any single party but may vote, as to each office contested, for any candidate regardless of party affiliation. This is the primary system in Washington and Alaska, and is the system adopted by Proposition 198. Alaska first adopted a blanket primary in 1947 and has had a blanket primary since that time with the exception of the period from 1960 to 1966 when it adopted a traditional open primary. Washington has had a blanket primary continuously since 1935. Finally, Louisiana has a version of a blanket primary, a nonpartisan primary open to all registered voters. Similar to the blanket primary in the sense that a voter can vote for any candidate without regard to party affiliation and is not limited to one party’s nominees, the nonpartisan primary differs because “[t]he two top vote receivers, regardless of party, meet in a subsequent (runoff or general) election.” Id. at 3.
“The relative merits of closed and open primaries have been the subject of substantial debate since the beginning of this century and no consensus has as yet emerged.” Tashjian v. Republican Party of Conn., 479 U.S. 208, 222, 107 S.Ct. 544, 552, 93 L.Ed.2d [651]*651514 (1986). It is apparent from this very case that the States continue to debate the question and to try new approaches.10
All of the four political party plaintiffs in this case have bylaws that prohibit persons who are not registered in their parties from voting in their parties’ primary. Article X, § 8 of the By-Laws & Rules of the California Democratic Party states that “No person shall be entitled to vote for a Democratic candidate at a partisan primary election unless he or she is a registered Democrat.” Pis.’ Ex. 1 at 36. The Republican Party also disputes the legitimacy of an open or blanket primary: “The [State Central] Committee shall recognize only those partisan nomination processes under California law that limit the electorate for partisan nomination elections to registered Republican voters, and which do not impose upon the Committee, without its express concurrence, a nomination process open to voters registered in other parties or as decline to state.” Defs.’ Ex. S at 2.1 The Peace & Freedom Party State Central Committee By-Laws and the Libertarian Party of California Bylaws have similar provisions. See Pis.’ Ex. 3 at 6, and Ex. 4 at 13.11
The California Democratic Party, the Peace and Freedom Party, and the Libertarian Party of California and certain of their officers brought this action against the Secretary of State of the State of California seeking a declaration that Proposition 198 is unconstitutional and requesting an injunction against its implementation.12 The California Republic Party and certain of its officers intervened as plaintiffs. Californians For An Open Primary intervened as a defendant.13 The court heard testimony over four days and has received numerous exhibits, including the reports of various experts.
II.
This case presents two competing views as to the function of the direct primary and, to some extent, as to the function of the political parties. From the parties’ perspective, the parties are autonomous organizations and the primary is their opportunity to select their leaders and to define their positions on political questions. From the defendants’ perspective, the primaries are the first step by which the electorate as a whole, regardless of party affiliation, chooses its leaders, and the political parties, as they operate to frame the choice of candidates, are a part of a highly regulated governmental activity — the election process. Both positions find support in the Constitution and the relevant case law. The “freedom of association protected by the First and Fourteenth Amendments includes partisan political organization.” Tashjian v. Republican Party of Connecticut, 479 U.S. 208, 214, 107 S.Ct. 544, 548, 93 L.Ed.2d 514 (1986). “As a result, political parties’ government, structure, and [652]*652activities enjoy constitutional protection.” Timmons v. Twin Cities Area New Party, 620 U.S. 351, 357, 117 S.Ct. 1364, 1369, 137 L.Ed.2d 589 (1997). However, the States also have a strong interest in the conduct of elections. “[T]he Constitution grants to the States a 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, 479 U.S. at 217, 107 S.Ct. at 549. It is the clash of these two constitutionally rooted interests that must be resolved in this case.
The Supreme Court has never directly considered the constitutionality of a blanket or open primary. In Democratic Party of United States v. Wisconsin ex rel. LaFollette, 450 U.S. 107, 101 S.Ct. 1010, 67 L.Ed.2d 82 (1981), the Court had the opportunity to decide the constitutionality of Wisconsin’s open primary but decided the case on a different basis, merely noting that the State Supreme Court had concluded that the “open primary serves [a] compelling state interest” and that “[ujpon this issue, the Wisconsin Supreme Court may well be correct.” Id. at 121, 101 S.Ct. at 1018.14 Perhaps the closest case is Tashjian v. Republican Party of Connecticut, 479 U.S. 208, 107 S.Ct. 544, 93 L.Ed.2d 514 (1986), but in Tashjian the Court addressed the reverse of the situation presented here; in Tashjian the state election law provided for a closed primary and the political party desired to open its primary to independent voters.15 Thus, the Court had no occasion to address either the State’s interests in an open or blanket primary or the burdens imposed on a political party by such a primary system. Two State Supreme Courts — the Supreme Courts of Alaska and Washington — have examined the constitutionality of the blanket primary and both have concluded that a blanket primary withstands a party’s First Amendment challenge. See O’Callaghan v. Alaska, 914 P.2d 1250 (Alaska 1996), cert. denied, 520 U.S. 1209, 117 S.Ct. 1690, 137 L.Ed.2d 818 (1997); Heavey v. Chapman, 93 Wash.2d 700, 611 P.2d 1256 (1980); Anderson v. Millikin, 186 Wash. 602, 59 P.2d 295 (1936). As noted above, the Supreme Court of Wisconsin came to the same conclusion concerning the Wisconsin open primary in the face of the Democratic Party’s claim that by permitting non-party members to vote in its primary, including Republicans and adherents of other parties, the State’s open primary violated the Party’s associational rights. See Wisconsin ex rel. La Follette v. Democratic Party, 93 Wis.2d 473, 287 N.W.2d 519 (1980).
The Supreme Court recently has explained the standard and manner of review that should be applied when a state election law is challenged by a political party on the basis of the First Amendment:
When deciding whether a state election law violated First and Fourteenth Amendment associational rights, we weigh the ‘character and magnitude’ of the burden the State’s rule imposes on those rights against the interests the State contends justify that burden, and consider the extent to which the State’s concerns make the burden necessary. Regulations imposing severe burdens on plaintiffs’ rights must be narrowly tailored and advance a [653]*653compelling state interest. Lesser burdens, however, trigger less-exacting review, and a State’s “important regulatory interests” will usually be enough to justify “reasonable nondiscriminatory restrictions.” No bright line separates permissible election-related regulation from unconstitutional infringements on First Amendment freedoms.
Timmons v. Twin Cities Area New Party, 520 U.S. 351, 358, 117 S.Ct. 1364, 1370, 137 L.Ed.2d 589 (1997) (citations omitted). In describing the “flexible” balancing test now used by the Court, the Ninth Circuit has cautioned that even when “strict scrutiny,” the highest standard of review, has been invoked, “its invocation in election law cases has not preordained their outcome.” Lightfoot v. Eu, 964 F.2d 865, 869 (9th Cir.1992). Moreover, the Court does not require “elaborate, empirical verification of the weightiness of the State’s asserted justifications.” Timmons, 520 U.S. at 363, 117 S.Ct. at 1372.
In evaluating the constitutionality of Proposition 198, the court therefore must first determine the character and magnitude of the burdens imposed by the blanket primary on the associational rights of the political parties. The court then must consider the State’s interests and weigh the strength of those interests against the particular burdens placed on the parties’ associational rights. The greater the intrusion on the parties’ associational rights, the greater must be the State’s interests and the more narrowly tailored must be the Proposition to advance those interests.16
III.
The parties’ main argument is that a blanket primary permits voters who are not members of a particular party — who may indeed be hostile to that party’s electoral interests — to vote in that party’s primary and thus to participate in the process by which the party defines its positions on a variety of issues and by which it chooses its standard bearers. In Tashjian the Supreme Court recognized the importance of these interests in invalidating a Connecticut statute that forbade the parties from permitting nonmembers from voting in their primaries:
[T]he freedom to join together in furtherance of common political beliefs ‘necessarily presupposes the freedom to identify the people who constitute the association.’ ... The statute here places limits upon the group of registered voters whom the Party may invite to participate in the basic function’ of selecting the Party’s candidates. The State thus 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, 479 U.S. at 215-16, 107 S.Ct. at 548-49 (citations omitted). Accord Timmons v. Twin Cities Area New Party, 520 U.S. 351, 359, 117 S.Ct. 1364, 1370, 137 L.Ed.2d 589 (1997) (“the New Party, and not someone [654]*654else, has the right to select the New Party’s standard bearer’ ”).
A.
The parties contend that their interest in limiting candidate selection to their own members is so great that no matter what the State’s interest and no matter how Proposition 198 will work in practice, the blanket primary must be unconstitutional on its face. Yet while the interest in determining who will vote in a party’s primary is one protected by the First Amendment, and is undoubtedly an important interest, this is not the end of the inquiry. If it were, then open primaries would also be unconstitutional upon any party’s objection: an open primary, every bit as much as a blanket primary, permits voters who are not registered in a party to vote in that party’s primary. But many of the States have some version of an open primary and to invalidate them all on this theoretical ground would be an extraordinary intrusion into the complex and changing election laws of the States and would remove from the American political system a method for candidate selection that many States consider beneficial and which in the uncertain future could take on new appeal and importance.17
Similarly, were the parties’ power to determine membership and to select their nominees unfettered, then the States would also lack power to prescribe candidate selection through a primary election or to require that all voters who have met a State’s registration requirements be permitted to vote in the parties’ primaries notwithstanding any other membership requirement that a party might impose.18 However, the States have considerable regulatory authority over the process of candidate selection and may require that the parties select their candidates by a primary election open to registered voters. See Lightfoot v. Eu, 964 F.2d 865, 871 (9th Cir.1992) (“political parties’ rights to nominate whomever they want, however they want, is not sacred”); Timmons v. Twin Cities Area New Party, 520 U.S. 351, 117 S.Ct. 1364, 137 L.Ed.2d 589 (1997) (State can prohibit fusion candidacies). The States’ authority over the political parties stems in part from the Constitution, Art. I, 4, cl. 1, and in part from the public function that the political parties perform in the candidate nominating and political processes.
Moreover, the analogy on which the parties rely so heavily, the analogy between the political parties and the myriad of private political, charitable, recreational, educational or religious associations at the center of American life, is impeifect at best. The political scientists who testified in this case agree that a political party cannot be described simply as a private club or association; instead, a political party has three dis-[655]*655tinet aspects' — the party in the electorate, the party organization, and the party in government.19 Unlike other private associations, at least in one of their avatars — the party in government — the political parties are very like the government itself. See Lowenstein, supra, at 1757 (“unlike any other private groups, political parties routinely, pervasively, and legitimately exercise their influence from within the government”). And the parties perform functions that are fairly characterized as governmental in nature, such as the nomination of candidates.
Furthermore, at least the major political parties lack the unity of purpose and cohesive membership characteristic of most private organizations. The major political parties in this country “ ‘have been characterized by a fluidity and overlap of philosophy and membership.’ It can hardly be denied that [the parties] generally [have] been composed of various elements reflecting most of the American political spectrum.” Democratic Party, 450 U.S. at 133, 101 S.Ct. at 1024 (Powell, J., dissenting) (citations omitted). And, as discussed above, the term “party member” as applied to registered voters is something of a misnomer. The States determine who will vote at primary elections— registered voters. See Cal. Elections Code §§ 2151, 7164, 7362, 7804, 8001. These voters have not allied themselves to a political party in the customary way that one normally becomes a member of a private association; they have not paid dues, taken an oath, expressed agreement with certain ideals or goals, attended meetings, accepted any obligations, or attended upon any other rite of initiation. The parties do not and could not insist upon more than mere registration as the price of voting in the primary elections.20 Thus, to contend that Proposition 198 permits non-party members to vote in a party’s primary, while accurate, covers over much of the imprecision in the political reality.
In short, it is necessary to look beyond the parties’ claim that their constitutional interest in limiting candidate selection to their respective members necessarily prevails over any State interest and without regard to how the voters may behave under a blanket primary.
B.
Turning to the more practical implications of a blanket primary, the parties contend that a blanket primary process in California will have a destructive effect on all political parties both major and minor. The parties predict that there will be widespread crossover voting, whether benevolent or malicious, and that the consequences of this large crossover vote will be many and adverse to the parties. The parties contend that the cross[656]*656over vote, and the prospect of a large crossover vote, will affect the selection and behavior of party nominees and elected officials; weaken party discipline; increase the costs of primary elections by causing candidates to compete as if it were a general election; dampen the morale of party activists; and disrupt internal party structure and governance.
Much of the testimony at trial was directed to the question of whether voters in blanket primary states tend to cross over in sufficient numbers to influence the outcome of elections and whether they cross over for benevolent or malevolent reasons. A cross-over voter is one who votes for a candidate of a party in which the voter is not registered. Thus, the cross-over voter could be an independent voter or one who is registered to a competing political party.21 A cross-over voter may be actuated by three different purposes. First, the voter’s purpose may be to vote for the voter’s most preferred choice to win in the general election; this kind of cross-over voter is sometimes referred to as engaging in “sincere” voting. Second, the voter may be engaged in “strategic” voting, designed to obtain the best possible outcome, by voting for the voter’s second choice in a setting where the voter’s first choice, within the voter’s own party, either has no apparent chance of victory at the general election or is so certain to win the party’s nomination that the voter perceives that his or her vote is best spent influencing another party’s primary and thus the choice at the general election. Finally, the voter may be a raider, crossing over to vote for a candidate who is perceived as the “weakest” of an opposing party’s candidates in the hopes that this candidate will win the primary only to be trounced at the general election.22
The parties contend that a blanket primary will unleash a torrent of cross-over voting even by comparison to open primaries. The parties reason that unlike an open primary, a blanket primary permits a voter to cross over as to particular contested elections without losing the ability to vote in the remainder of the contests in the voter’s own party. The evidence at trial suggests that the parties are correct that cross-over voting will be more prevalent after Proposition 198. Indeed, defendants can hardly dispute this point while maintaining that the State’s interests are furthered by a blanket primary precisely because it facilitates cross-over voting. Nonetheless, the parties’ characterization of the likely amount or effects of cross-over voting due to a blanket primary is overdrawn.
To begin with, there is little evidence that raiding will be a factor under the blanket primary. On this point there is almost unanimity among the political scientists who were called as experts by the plaintiffs and defendants. Professor Costantini, for plaintiffs, gave his opinion that
[s]urvey evidence tends not to support any fear that raiding (sometimes characterized as “mischievous voting”) will be a common phenomenon under Proposition 198. At least in open presidential primaries (the focus of almost all the relevant studies), it appears that the great majority of crossover voters — from 90% to 95% or more— invade a party’s primary to vote for a preferred candidate and not to vote for a candidate who they will oppose in the general election and who they believe will be easier to defeat in that election.
Pis.’ Ex. 7, Costantini Report, at 7; R.T. at 485-86. Professor Donald Olson of the University of Washington-Seattle, reflecting on the over half century of experience with a blanket primary in Washington, explained that the conditions for successful raiding rarely exist:
The potential for raiding under the blanket primary generates. wide media attention and spawns electioneering lore that is simply unwarranted by actual experience. Technically defined, raiding would occur where partisan voters of party X crossover in sufficient numbers to help select the weakest or most vulnerable candidate in party Y. The requirements for this form of electoral subterfuge are: (1) the absence [657]*657of a meaningful primary contest in party X; (2) provision of sufficient information to identify the weaker and most vulnerable potential opponent in party Y; (3) concerted action by a large enough bloc of voters in party X to influence the outcome in the primary of party Y; and (4) the abandonment of traditional party loyalties by partisan voters to enter the opponent’s primary. In the experience of Washington state, these requirements are simply too demanding and have not been fulfilled in actual practice.
Defs.’ Ex. G, Olson Report, at 6. See Pis.’ Ex. 6, Cain Report, at 6-7; R.T. at 355. Given that the purpose of raiding is to weaken a competitive party’s advantage at the general election, raiding is not a concern for minor parties because their candidates rarely have a chance of prevailing at the general election. Indeed, no conclusive instance of raiding in such numbers as to affect the outcome of any primary was offered by plaintiffs. 23
As to benevolent cross-over voting, the evidence suggests that there will be particular elections in which there will be a substantial amount of cross-over voting — perhaps as high as 25% of the vote for that office — • although the cross-over vote will rarely change the outcome of any election and in the typical contest will not be at significantly higher levels than in open primary states. There will be a small number of elections in which the cross-over vote proves decisive, and there will be other elections in which the possible importance of the cross-over vote will be an influence on the conduct of the primary campaign and the conduct of elected officials.
The experience in Washington suggests that as to the election for any particular office, approximately 10% of the voters are cross-over voters, excluding independents, while in some exceptional cases this figure may rise as high as 20-25%. R.T. at 511.24 Typically the cross-over vote in Washington does not change the result of an election but only the margin of victory. R.T. at 513. Moreover, the amount of cross-over voting is greater in a blanket primary than an open primary but not dramatically so. According to Professor Nagler, who co-authored the most comprehensive study of patterns of cross-over voting presented at the trial, the average cross-over rate, excluding independents, is something on the order of 5% of voters in closed primary states,25 10% of voters in open primary states, and 11-15% in blanket primary states. R.T. at 780, 788; Defs.’ Ex. H, Alvarez and Nagler Report, at 11-12. The study concludes that “crossover voting in blanket primaries is not dramatically higher than in other types of open primary systems” and that “there is very little crossover voting in general in primary elections in the United States.” Id. at 35-36.26
[658]*658Although cross-over voting typically does not change the outcome of a primary and may not be much greater in a blanket than in an open primary, it is probable over time that there will be some elections under Proposition 198 that will be decided by the crossover vote. Cross-over voting is most likely to occur when one party’s primary is competitive while the other party’s primary is not. When this condition of “asymmetric competition” occurs, voters of the party with the noncompetitive primary may cross over to the other primary to influence the choice. The incentive to cross over will be particularly high if the winner of the other party’s primary is almost certain to win at the general election. See Pis.’ Ex. 6, Cain Report, at 8. Every two years in California there are 80 Assembly races, 20 State Senate races, and 52 Congressional races. Examining the 1996 Assembly primaries, Professor Cain identified ten races in which there were asymmetric competitive conditions. In some of these races, a cross-over rate of 5% could have changed the outcome of the election had all of the cross-over voters voted for the same candidate. Given that in Washington, cross-over rates occasionally reach levels as high as 20-25%, and that the cross-over rate in California could be higher than Washington,27 it is likely that in the fullness of time some California primary races will be decided by cross-over voters, even if the number of such occasions is not large. Depending on the circumstances, the swing of even one seat could give a legislative majority to one party.
Furthermore, the possibility of a decisive cross-over vote — whether or not that vote materializes — could well affect the conduct of elections and elected officials. Under a blanket or open primary, the primary election becomes similar to the general election. Candidates will perceive a need or opportunity to attract support from the entire electorate, not simply from members of their respective parties. This will affect not only the cost of the primary — which is likely to increase, see R.T. at 270 — but may affect the positions that candidates take on political issues as they seek to attract cross-over voters. Moreover, once elected, candidates may perceive that the cross-over vote makes them less vulnerable at the primary to the wrath of the party organization or party regulars for acts of apostasy. A blanket or open primary weakens the “disciplining effect” of the primary challenge in requiring “the officeholder [to] carry out the party philosophy and support the majority position of the party caucus.” R.T. at 214. The corollary to this loss of discipline is a loss of power to party activists such that the ardor of party volunteers and activists may be dampened in a system that permits nonparty members to exert influence over the party’s selection of a candidate. This has been the experience in Washington and is likely to occur in California. See R.T. at 558; R.T. at 18; R.T. at 989-990.
Finally, the parties argue that their internal governance will be disrupted by the blanket primary and the cross-over vote. Although Proposition 198 specifically exempts party committee positions from the blanket primary, both the Republican and Democratic Parties permit officeholders and nominees to appoint members of their governing bodies, the Democratic State Central Committee and the Republican State Central Committee. Because nominees may have been selected by non-party members, and because the nominees in turn choose members of the Central Committees, the parties contend that their governing body may be adulterated. Further, the parties warn that their respective national organizations may not accept the results of the California presidential primary, if that primary is a blanket primary, with the result that the parties may be forced to hold a nominating convention or caucus. Neither of these contentions is particularly forceful, however. As to membership on the Central Committee, the parties are free to change their selection process if they wish. Moreover, both parties already give more positions on the Central Committee to successful candidates; yet these successful candidates are precisely those who have appealed more broadly across party lines. R.T. at 9, 224. As to the national [659]*659nominating conventions, the California parties may be forced to adopt a different procedure for delegate selection other than the presidential primary or they may receive a waiver from their national organizations such as Wisconsin eventually received from the Democratic Party. R.T. at 123,145. This is a choice for the parties to make, and they can do so as they please without interference from the State.
C.
The magnitude of the burdens imposed on the parties by Proposition 198 are not fairly characterized either as “severe” or as “negligible,” the two labels most often applied in the caselaw at this juncture of the analysis. The burdens are not negligible. State law requires each party to select its candidate through a primary, and the winner of the primary will be the party’s nominee. The blanket primary wrests a measure of control of the primary process from the party and its members. This is its purpose. The party may still endorse a candidate and may throw financial support to a particular candidate. And, in the typical election, the cross-over vote will not be decisive. Yet by permitting non-party members to vote in a party’s primary-even granting the uncertainty of the label “party member” — Proposition 198 is an intrusion on a party’s traditional right to select its own candidate. In this respect, the blanket primary is no different from an open primary although the practical effect of a blanket primary on the selection process may be somewhat greater than an open primary. In addition, a blanket primary, like an open primary, may depress the vigor and morale of party activists and may affect the ability of the party to discipline errant legislators.
On the other hand, the burdens on the parties due to Proposition 198 are not likely to prove severe or debilitating. The Washington experience, as well as that of the many open primary states, suggests that Proposition 198 will not diminish the efficacy or strength of the political parties in California by any substantial degree.28 Indeed, a study of the political parties in the 1970’s and 1980’s in 13 western states, including Washington and California, concluded that the political parties in Washington were the most competitive with one another of all 13 states. R.T. at 524; see Tashjian, 479 U.S. at 223 n. 12, 107 S.Ct. at 553 n. 12 (“dire” predictions as to decay of parties not borne out by the 29 States that have open primary features). Both the Democratic and Republican Parties continue to be strong political parties in Washington, and can be expected to operate with little diminution in their effectiveness in California.29 The parties do not show that the blanket primary is significantly different theoretically or practically from the open primary. Nor do they show or even contend that Proposition 198 has any discriminatory purpose or effect. Thus, the burden imposed by a blanket primary falls evenly on all of the parties whatever their size or ideology. Finally, Proposition 198 leaves untouched the wide field of party endeavor that is not directed specifically to the primary, activities such as developing the party platform, registering new voters, and getting out the vote on election day. Defs.’ Ex. I, Quinn Report at 18.
In light of all of these considerations, the court finds that the parties have succeeded in showing that the blanket primary imposes a significant but not severe burden on their associational rights. To survive the challenge mounted by plaintiffs, Proposition 198 [660]*660must be supported by interests that are sufficient to outweigh the burdens identified above. Those interests need not be compelling, given that the burdens are not crushing, but they must be important.
IV.
The States have broad regulatory authority over the conduct of elections including primaries.30 The primary election is the critical first stage in the electoral process: “The direct primary in California is not merely an exercise or warm-up for the general election but an integral part of the entire election process, the initial stage in a two-stage process by which the people choose their public officers.” Storer v. Brown, 415 U.S. 724, 735, 94 S.Ct. 1274, 1281, 39 L.Ed.2d 714 (1974). “As a practical matter, the ultimate choice of the mass of voters is predetermined when the nominations [by the political parties] have been made.” Morse v. Republican Party, 517 U.S. 186, 205-06, 116 S.Ct. 1186, 1199, 134 L.Ed.2d 347 (1996) (citation omitted).
Proposition 198 is the latest development in a history of political reform measures that began in the Progressive Era. “The Progressives believed democracy should be something greater than competition between political parties.” Lightfoot v. Eu, 964 F.2d at 872. The Progressives sought to diminish the importance of the parties and “party bosses” by establishing nonpartisan local elections, a comprehensive civil service sys-te'm, direct legislation through the initiative and referendum, and the cross-filing system by which members of one party could run in the primary of the other party. See Defs.’ Ex. I, Quinn Report, at 7 (quoting J. Owens and E. Costantini, California Politics and Parties (1970)).31 The cross-filing system is of particular note because in practice it bore some resemblance to a blanket or nonpartisan primary system.32
Defendants advance a number of State interests furthered by Proposition 198. All of the various interests, however, reduce to one fundamental contention: according to defendants, Proposition 198, like other Progressive Era reforms, enhances the democratic nature of the election process and the representativeness of elected officials.
Defendants stress that a blanket primary opens up the candidate selection process to two types of arguably “disenfranchised” voters — independents and minority party voters in safe districts. In a blanket primary, independent voters will be able to participate in framing the choice that they will ultimately have to make in the general election. Approximately 11% of the electorate in Califor[661]*661nia are registered as independents, and the proportion of independent voters is growing rapidly. R.T. at 940. Moreover, many California voters reside in “safe districts,” districts in which one party is dominant.33 Voters in these districts who belong to the minority party have no influence or voice because the general election is decided as a practical matter in the primary of the dominant party. Of 15.6 million registered voters in California, approximately one quarter are “disenfranchised” in this way. R.T. at 854. Defendants suggest that the inability of these voters to influence the outcome of the election in their districts leads to the election of representatives who see no need to represent these voters or to pay heed to their interests. Of course, such voters could register in the majority party, but this would require them to give up their current party or independent registration.
Apart from the exclusion of independents and minority voters in safe districts, defendants contend that the closed primary gives undue power to party members and loyalists, including incumbent elected officials, such that the voters, in the general election are presented with candidates who are more partisan and who take positions on the policy issues of the day that are not representative of the majority of the electorate. See R.T. at 243, 269, 299, 857. Defendants note that turnout is low at primary elections, and thus that a minority of more partisan party members, influenced by the endorsements of party leaders, determine the choice for the much larger group of all voters at the coming-general election. Defendants suggest that this skewing of the process toward partisan candidates leads- to legislative impasse and thus alienation of the electorate from the political system. Defendants contend that candidates who would command majorities at the general election are unable to prevail at the primary such that the voters at the general election are forced to pick among candidates none of whom is the first choice of the majority of voters. R.T. at 255-256. Although these assertions are not readily susceptible of proof, many are not contested,34 or are supported by persuasive opinion evidence from defendants’ experts. See, e.g., R.T: at 954-963, 974-975. Furthermore, these contentions are clearly stated in the ballot pamphlet, and it is a fair inference from the margin of support that the people of California agree that a blanket primary system will lead to the election of more representative “problem-solvers” who are less beholden to party officials and “special interest groups.” While political scientists may disagree on the question of whether the political system is benefited when parties are more or less distinct or polarized, it is apparent that the voters have sided with less partisanship in adopting Proposition 198.35
Defendants also argue that representativeness and democratic participation are bene-fitted -by the higher voter participation expected in a blanket primary election. Voter participation in primary and general elections in California has declined markedly over the past two decades. R.T. at 658. [662]*662Turnout rates in blanket primary states have been higher than in California, and there is reason to expect that by providing more choice, the blanket primary may encourage more voters to participate at the primary election. Similarly, defendants contend that elections will be more competitive under the blanket primary, again leading to an increase in voter interest and participation. Although the empirical case for the assertion that blanket primaries are more competitive than other primaries is not conclusive, there is reason to expect some increase in the number of candidates, the closeness of elections, and, therefore, voter turnout.
Finally, defendants point to one interest— voter privacy — that does not fit neatly under the representativeness or participation headings, although it is possible that greater privacy could lead to greater participation. Defendants note that a blanket primary protects voter privacy by permitting voters to participate in the primary election without publicly declaring a party affiliation as is the case in closed primary states and in some open primary states.
V.
Taken as a whole the State’s interests that support the blanket primary are substantial, indeed compelling. There is a “substantial state interest in encouraging compromise and political stability, in attempting to ensure that the election winner will represent a majority of the community.” Storer v. Brown, 415 U.S. 724, 729, 94 S.Ct. 1274, 1278, 39 L.Ed.2d 714 (1974). In attempting to reduce partisan strife, Proposition 198 is similar to other measures adopted by the States to deal with the problem of factionalism, a problem that deeply troubled the framers of the Constitution. See The Federalist No. 10 (Madison). A State’s interest in ameliorating “unrestrained factionalism” is a compelling one. Increasing the representativeness of elected officials, .giving voters greater choice, and increasing voter turnout and participation in the primary process are all interests of the first order. See Lightfoot v. Eu, 964 F.2d 865, 872 (9th Cir.1992) (State’s interest in a direct primary is compelling). The voters’ belief in the fairness of the electoral process is also a substantial state interest; it is apparent that the voters of California believe that the system is fairer when all voters, including independents and regardless of party affiliation, may participate in framing the choice of candidates at the general election. And while the parties argue that there are other ways to increase voter turnout, the fundamental goal of enhancing representativeness by providing all voters with a choice that is not predetermined by party members alone can only be advanced by the blanket primary. Unlike Tashjian in which the Democratic Party controlled the legislature and attempted to tell the Republican party who could vote in its primary, and where the State’s purported interest in a closed primary was to protect the parties from disruption from without, Proposition 198 is a non-discriminatory measure that was adopted by a clear majority of voters, with a convincing margin from both major parties, and which advances interests that are uniquely those of the State and its electorate as opposed to the parties. These interests outweigh the also substantial interest that the political parties have in controlling who votes in the primary.
In reaching this conclusion, the court does not decide whether a blanket primary is a good idea; it may prove to be a bad idea, in which case the people of the State presumably will act to reform the system in the future as they have in the past. Nor does the court decide whether the voters were correct in concluding that a more representative and participatory system is desirable at this time in the State’s history or whether it is wise to weaken the parties by some degree to further the goals of representativeness and participation. These important questions are for the people of the State to decide.
Each of the States has its own experiment in democratic government. The history of election law is one of change and adaptation as the States have responded to the play of different political forces and circumstances. There is little reason to expect that the future will be any different. The court concludes that a blanket primary serves suffi[663]*663ciently weighty interests that the State of California should not be precluded from ever trying such a procedure.
Judgment is awarded to defendants.
IT IS SO ORDERED.