DAKOTA NELSON; BELINDA BIAFORE, individually and as Chairperson of the West Virginia Democratic Party; ELAINE A. HARRIS, individually and as Chairperson of the Kanawha County Democratic Executive Committee; WEST VIRGINIA DEMOCRATIC PARTY; and WEST VIRGINIA HOUSE LEGISLATIVE COMMITTEE, Plaintiffs, v. MAC WARNER in his official capacity as West Virginia Secretary of State; and VERA MCCORMICK, in her official capacity as Clerk of Kanawha County, West Virginia, and all ballot commissioners for the state of West Virginia, Defendants.
CIVIL ACTION NO. 3:19-0898
IN THE UNITED STATES DISTRICT COURT FOR THE SOUTHERN DISTRICT OF WEST VIRGINIA HUNTINGTON DIVISION
August 10, 2020
This suit challenges the constitutionality of West Virginia‘s Ballot Order Statute, which mandates that ballots for partisan offices list first the party whose presidential candidate received the most votes in the last election. Following the parties’ bench trial, the Court must now weigh the extent to which the Statute burdens the plaintiffs’ First and Fourteenth Amendment rights against the state‘s asserted interests. Finding the state‘s interests insufficient to justify the burdens created by the Statute, the Court now declares the Ballot Order Statute unconstitutional, enjoins the defendants from enforcing it, and orders the defendants to implement a constitutional ballot ordering system for future elections, including the November 2020 general election.
The West Virginia legislature enacted the state‘s Ballot Order Statute in 1991. Stips. ¶ 22, ECF No. 106-2. The Statute provides:
The party whose candidate for president received the highest number of votes at the last preceding presidential election is to be placed in the left, or first column, row or page, as is appropriate to the voting system. The party which received the second highest vote is to be next and so on. Any groups or third parties which did not have a candidate for president on the ballot in the previous presidential election are to be placed in the sequence in which the final certificates of nomination by petition were filed.
The plaintiffs, all of whom are affiliated with the Democratic Party, argue candidates listed first on a ballot benefit from a human tendency to choose the first candidate in a list of names. Am. Compl. 3, 25, ECF No. 7. Because of this “primacy effect,” the plaintiffs argue the Ballot Order Statute gives certain candidates a significant advantage over others based solely on partisan affiliation.1 Am. Compl. ¶¶ 1–3. Count One alleges the Ballot Order Statute is an undue burden on the right to vote in violation of the First and Fourteenth Amendments because the Statute dilutes votes for candidates whose political party is not favored by the Statute. Am. Compl. ¶¶ 36–42. Count Two alleges the Statute constitutes disparate treatment in violation of the Fourteenth Amendment‘s Equal Protection Clause because the Statute treats one major political party (and its candidates, members, constituencies, and supportive voters and organizations) differently from the other major party by granting an electoral advantage based solely on the party‘s performance in
The first individual plaintiff is Dakota Nelson, a resident of Huntington, Cabell County, West Virginia. Stips. ¶ 1; Trial Tr. 89, ECF Nos. 119–22. He is registered to vote as a Democrat and is a member of the Democratic Party. Stips. ¶ 1; Tr. 89. Nelson was a Democratic candidate in the 2018 general election for District 16 of the West Virginia House of Delegates, which is a three-delegate district.2 Tr. 92; Pls.’ Ex. 8, at 4, ECF No. 104-9. The ballot listed him last out of six candidates, and he lost the election. Tr. 90; Pls.’ Ex. 8, at 4. Nelson ran again for District 16 in the June 2020 primary and was nominated. Stips. ¶ 2. He will appear on the November 2020 ballot with three Republicans and two other Democrats who are also running for District 16. Stips. ¶ 4; Tr. 89-90.
The second individual plaintiff is Belinda Biafore. She is a registered Democrat and resident of Fairmont, Marion County, West Virginia. Stips. ¶ 6. She has been a member of the Democratic Party in West Virginia for over forty-four years and an active supporter of the Democratic Party for over fifty years. Stips. ¶ 6. She is the Chairperson of the West Virginia Democratic Party and previously served as Vice Chairperson, as a member of the Executive Committee, and in several leadership roles for the West Virginia Federation of Democratic
Elaine Harris is the third individual plaintiff. She did not testify at trial, but the parties stipulated to the following. She is a registered Democrat and resident of Kanawha County, West Virginia. Stips. ¶ 7. She has been a member and active supporter of the Democratic Party in West Virginia for many years. Stips. ¶ 7. She currently serves as Chairperson of the Kanawha County Democratic Executive Committee. Stips. ¶ 7. Harris regularly supports Democratic candidates in West Virginia elections and intends to vote for Democratic Party candidates in the upcoming general election. Stips. ¶ 7.
The plaintiffs also include two organizations. The West Virginia Democratic Party is a state political party as defined by
The Court has adjudicated this case on an expedited schedule because the ballot order for the November election must be set on August 25.
II. STANDING
Before arriving at the case‘s merits, the Court must answer the defendants’ contention that the plaintiffs lack standing to bring their claims. The defendants have unsuccessfully challenged the plaintiffs’ standing at every stage of this litigation. In denying the defendants’ motions for summary judgment, the Court held that Nelson, the Democratic Party, and the Legislative Committee demonstrated direct standing based on harm to their electoral prospects.4 Mem. Op. 4–10, ECF No. 87. The Court also held the Democratic Party demonstrated associational standing
A. Legal Standard
B. Nelson, the Democratic Party, and the Legislative Committee have direct standing.
1. Injury-in-Fact
a. Nelson proved an injury based on his harmed electoral prospects.
“The inability to compete on an equal footing due to the application of allegedly biased criteria has been recognized in many contexts as an injury in fact sufficient to support constitutional standing.” Nat. Law Party of U.S. v. Fed. Elec. Comm‘n, 111 F. Supp. 2d 33, 44 (D.D.C. 2000) (collecting cases); e.g. Ne. Fla. Chapter of Associated Gen. Contractors of Am. v. City of Jacksonville, 508 U.S. 656, 666 (1993) (holding contractors had standing to challenge city ordinance based on “inability to compete on an equal footing in the bidding process“); see also Fulani v. Brady, 935 F.2d 1324, 1327 (D.C. Cir. 1991) (collecting cases and explaining that courts have recognized “competitor standing” in “circumstances where a defendant‘s actions benefitted a plaintiff‘s competitors, and thereby caused the plaintiff‘s subsequent disadvantage“). The D.C. Circuit explained so-called “competitive standing” this way: “when regulations illegally structure a competitive environment—whether an agency proceeding, a market, or a reelection race—parties defending concrete interests (e.g., retention of elected office) in that environment suffer legal harm under Article III.” Shays v. Fed. Elec. Comm‘n, 414 F.3d 76, 87 (D.C. Cir. 2005).
Several circuits have extended this competitive standing theory to elections, holding that a candidate and his or her party can show an injury-in-fact if the defendant‘s actions harm the candidate‘s chances of winning. See, e.g., Pavek v. Donald J. Trump for President, Inc., No. 20-2410, 2020 WL 4381845, at *1 (8th Cir. July 31, 2020) (holding two Democratic Party committees
Contrary to this overwhelming precedent, the defendants have relied on the recent Eleventh Circuit decision in Jacobson v. Florida Secretary of State, 957 F.3d 1193 (11th Cir. 2020) to argue
Both the plaintiffs in Jacobson and here relied on expert witness Dr. Jon A. Krosnick, one of the nation‘s foremost political psychologists. In both cases, Dr. Krosnick submitted extensive statistical research and concluded that listing a candidate first “almost always” accords that person an advantage. Pls.’ Ex. 2, at 1, ECF No. 104-2; Sec‘y‘s Ex. 2, at 2, ECF No. 105-2. His research includes a meta-analysis of over 1,000 primacy effect tests, of which 84 percent manifested differences in the direction of primacy a statistically significant pattern constituting “near unanimity.” Pls.’ Ex. 2, at 21–23; Sec‘y‘s Ex. 2, at 34–36. Due either to a misunderstanding of the evidence or an unclear presentation of it by the plaintiffs, the Eleventh Circuit failed to grasp that Dr. Krosnick‘s conclusions on the prevalence of the primacy effect apply to all elections. To prevent this Court from making the same interpretative error, Dr. Krosnick repeatedly testified that his research establishes an 80 to 85 percent probability of the primacy effect occurring in each and
As explained below in the merits analysis, the Court finds Dr. Krosnick‘s conclusion that the primacy effect is 80 to 85 percent likely to manifest in each West Virginia election, including multi-candidate elections, to be reasonable, reliable, and credible. Nelson therefore proved harm to his electoral prospects based on the extreme likelihood that the primacy effect manifested in his 2018 race and will manifest in his 2020 race. The 15 to 20 percent probability of the primacy effect not occurring in an election does not undermine Nelson‘s standing because standing does not “uniformly require plaintiffs to demonstrate that it is literally certain that the harms they identify will come about.” Clapper, 568 U.S. at 414 n.5. Instead, proving a “substantial risk” of injury is sufficient. Id.; Susan B. Anthony List, 573 U.S. at 158. An 80 to 85 percent probability of the primacy effect benefiting a candidate‘s opposition is undoubtedly a “substantial risk” of injury.6 See also Pavek v. Simon, No. 19-CV-3000 (SRN/DTS), 2020 WL 3183249, at *13–14 (D. Minn. June 15, 2020) (holding plaintiffs demonstrated an injury-in-fact to challenge Minnesota‘s ballot order statute based on a substantial risk that the primacy effect will occur in the 2020 general
b. The Democratic Party and the Legislative Committee also proved an injury based on harm to their candidates’ electoral prospects.
As noted above, the unlawful structuring of a competitive election not only injures a disadvantaged candidate, but also the political party supporting that candidate. See Pavek, 2020 WL 4381845, at *1 (holding two Democratic committees demonstrated standing to challenge ballot order statute based on the statute unequally favoring supporters of other political parties); Green Party of Tenn., 767 F.3d at 538 (holding third parties had standing to challenge state‘s ballot access and ballot order statutes); Drake, 664 F.3d at 783 (explaining that the “potential loss of an election” can be a sufficient injury-in-fact for party officials); Tex. Democratic Party, 459 F.3d at 587–88 (holding party demonstrated injury of “harm to its election prospects” and “threatened loss of political power“); Fulani, 917 F.2d at 1030 (holding third party faced injury of “increased competition” when defendants allegedly improperly placed major-party candidates on the ballot); Schulz, 44 F.3d at 53 (holding third-party representative had standing because the party stood to suffer from “competition on the ballot from candidates that . . . were able to ‘avoid complying with the Election Laws’ and a resulting loss of votes“); Hollander, 566 F. Supp. 2d at 68 (explaining that a “political party has standing to challenge the inclusion of an allegedly ineligible rival on the ballot, on the theory that doing so hurts the . . . party‘s own chances of prevailing in the election“). The Fifth Circuit explained why in Texas Democratic Party v. Benkiser:
[A] political party‘s interest in a candidate‘s success is not merely an ideological interest. Political victory accedes power to the winning party, enabling it to better direct the machinery of government toward the party‘s interests. While power may be less tangible than money, threatened loss of that power is still a concrete and particularized injury sufficient for standing purposes.
Following the Supreme Court‘s recent decision in Gill v. Whitford, 138 S. Ct. 1916 (2018), the Eleventh Circuit held in Jacobson that an organization‘s “general interest in its preferred candidates winning as many elections as possible” is only a “generalized partisan preference” that federal courts are “not responsible for vindicating.” Jacobson, 957 F.3d at 1206 (quoting Gill v. Whitford, 138 S. Ct. 1916, 1933 (2018)). The organizational plaintiffs in Jacobson thus failed to establish standing because their claimed injury was a “systemic disadvantage to [their] party relative to other political parties” rather than harm to “a particular candidate‘s prospects in a future election.” Id. The Eleventh Circuit, however, expressly declined to decide “whether a political party would have standing to challenge an electoral practice that harmed one of its candidate‘s electoral prospects in a particular election.” Id. (emphasis in original). Jacobson is therefore inapplicable here because the Democratic Party and the Legislative Committee base their standing on the Ballot Order Statute‘s harm to their candidates (Nelson and the twenty-eight Committee members previously identified) in the upcoming November election. See Tex. Democratic Party, 459 F.3d at 586 (holding party had standing to challenge an action that would reduce “its congressional candidate‘s chances of victory” in an upcoming election); Pavek, 2020 WL 3183249, at *14 n.13 (holding Jacobson did not apply to standing analysis for organizational plaintiffs because the plaintiffs’ focused their injury on harm to the plaintiffs’ candidates in the upcoming election), stayed on other grounds pending appeal, 2020 WL 4381845, at *1 (affirming the plaintiffs had standing). As already discussed, the plaintiffs proved through Dr. Krosnick that a “substantial risk” exists that
2. Traceability
Having concluded that Nelson, the Democratic Party, and the Legislative Committee proved an injury-in-fact, the Court now turns to traceability. This element of standing requires the injury be “fairly traceable to the challenged action of the defendant, and not the result of the independent action of some third party not before the court.” Lujan, 504 U.S. at 560 (citation omitted). As previously discussed, the plaintiffs showed through Dr. Krosnick that the Ballot Order Statute causes the primacy effect to disproportionately benefit Republican candidates, which injures Democratic candidates’ electoral prospects. This injury is fairly traceable to the defendant class of ballot commissioners because their statutory duties include preparing the ballots “necessary for conducting every election for public officers in which the voters of the county participate,” including all “general, special, and primary elections held in the county or any magisterial district thereof.”
3. Redressability
Lastly, the plaintiffs must show it is “likely, as opposed to merely speculative, that the injury will be redressed by a favorable decision.” Lujan, 504 U.S. at 561 (citation and internal quotation marks omitted). The plaintiffs ask the Court to declare the Ballot Order Statute unconstitutional, enjoin the defendants from enforcing it, and require the defendants to use a ballot ordering system that gives similarly situated major-party candidates an equal opportunity for the ballot to list them first. Am. Compl. 17. As explained regarding traceability, the Secretary of State and ballot commissioners are responsible under state law for enforcing the Statute, so declaratory and injunctive relief against them would effectively stop implementation of the Statute across West Virginia. Cf. Jacobson, 957 F.3d at 1208 (explaining that enjoining the Florida Secretary of State from following Florida‘s ballot order statute would not provide redress because the Secretary does not enforce the statute and the county supervisors who do enforce it were not joined as parties). If the Court enjoins enforcement of the Statute, the primacy effect will no longer disproportionately harm candidates’ electoral prospects based on their party affiliation. And the parties have stipulated that the kind of randomized or rotational ballot ordering that the plaintiffs seek to cure the harm of the Ballot Order Statute is possible with West Virginia‘s current voting systems. Stip. Regard. Voting Sys., ECF No. 106-1. The Court therefore finds that granting the requested relief would redress the plaintiffs’ injury.
In addition to having standing on their own behalf, the Democratic Party and the Legislative Committee have associational standing on behalf of members who are candidates in contested races in the November election. See Tex. Democratic Party, 459 F.3d at 587–88 (holding the Texas Democratic Party had associational standing on behalf of its candidate). Associational standing for an organization exists when “(a) its members would otherwise have standing to sue in their own right; (b) the interests it seeks to protect are germane to the organization‘s purpose; and (c) neither the claim asserted nor the relief requested requires the participation of individual members in the lawsuit.” Hunt v. Wash. State Apple Advert. Comm‘n, 432 U.S. 333, 343 (1977). The Democratic Party and the Legislative Committee meet all three requirements.
First, the Court has already determined that Nelson, a Democratic Party member, has standing to sue based on his harmed electoral prospects in the 2018 general election and the upcoming November election. See Retail Indus. Leaders Ass‘n v. Fielder, 475 F.3d 180, 186 (4th Cir. 2007) (citation omitted) (“Associational standing may exist even when just one of the association‘s members would have standing.“). For the same reasons, the twenty-eight Legislative Committee members who are running for reelection in November against Republican challengers have standing. See Stips. ¶ 13; Tr. 53–55. As with Nelson, the Statute‘s placement of these Democratic candidates below their Republican opposition creates a substantial risk of harm to the candidates’ electoral prospects because it is extremely likely that the primacy effect will benefit the candidates listed first.
Second, both the Democratic Party and the Legislative Committee undoubtedly seek to protect their organizational interests in this suit. Both organizations’ missions include electing Democratic candidates across West Virginia. Stips. ¶¶ 10, 13; Tr. 52, 70–71. In the words of the
Third, the claims asserted do not require participation by the organizations’ individual members. Even though Nelson is a plaintiff and his candidacy is a basis for standing, the Democratic Party‘s and the Legislative Committee‘s separate standing make Nelson‘s inclusion as a plaintiff unnecessary. And nothing about the plaintiffs’ claims or their requested relief make participation by the Legislative Committee‘s individual members necessary. “Unlike a suit for money damages, which would require examination of each member‘s unique injury, this action seeks a declaratory judgment and injunctive relief, the type of relief for which associational standing was originally recognized.” Retail Indus. Leaders Ass‘n, 475 F.3d at 187. Therefore, having met all three requirements, the Democratic Party and the Legislative Committee have associational standing to bring this suit.
In sum, the Court concludes Nelson, the Democratic Party, and the Legislative Committee have direct standing in this case based on harm to their electoral prospects. The Democratic Party also has associational standing on behalf of Nelson, and the Legislative Committee has associational standing on behalf of the twenty-eight identified members seeking reelection. Having found standing on these grounds, the Court declines to address whether Biafore and Harris have individual standing as voters. See Bostic v. Schaefer, 760 F.3d 352, 370 (4th Cir. 2014) (“[T]he presence of one party with standing is sufficient to satisfy Article III‘s case-or-controversy requirement.“) (quoting Rumsfeld v. Forum for Acad. & Institutional Rights, Inc., 547 U.S. 47, 52 n.2 (2006)). The Court now proceeds to the case‘s merits.
A. Legal Standard
“[V]oting is of the most fundamental significance under our constitutional structure.” Burdick v. Takushi, 504 U.S. 428, 433 (1992) (citation omitted). And 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). Electoral regulations often “implicate substantial voting, associational and expressive rights protected by the First and Fourteenth Amendments.” Libertarian Party of Va. v. Alcorn, 826 F.3d 708, 716 (4th Cir. 2016) (citation omitted). Voting regulations also implicate the Equal Protection Clause because “[h]aving once granted the right to vote on equal terms, the State may not, by later arbitrary and disparate treatment, value one person‘s vote over that of another.” Bush v. Gore, 531 U.S. 98, 104–05 (2000) (citation omitted). In other words, the Equal Protection Clause protects the right “to participate in elections on an equal basis with other citizens in the jurisdiction.” Dunn v. Blumstein, 405 U.S. 330, 336 (1972) (citations omitted). “These rights, however, are not absolute” because “all election laws, including perfectly valid ones, inevitably affect—at least to some degree-the individual‘s right to vote and his right to associate with others for political ends.” Alcorn, 826 F.3d at 716 (cleaned up) (citing Anderson v. Celebrezze, 460 U.S. 780, 788 (1983)).
Both of the plaintiffs’ claims here are cognizable because the Ballot Order Statute implicates the right to vote and the right to equal protection. Regarding Count One, the Statute affects the individual plaintiffs’ right to vote under the First and Fourteenth Amendments because, by awarding the benefit of the primacy effect to one party‘s candidates based on partisan affiliation, the Statute dilutes votes for candidates whose party the Statute disfavors. See, e.g., Pavek v. Donald J. Trump for President, Inc., No. 20-2410, 2020 WL 4381845, at *2 (8th Cir. July 31, 2020)
The Court adjudicates both Counts One and Two using the standard established by the Supreme Court in Anderson v. Celebrezze, 460 U.S. 780 (1983), and Burdick v. Takushi, 504 U.S. 428 (1992). To assess the Ballot Order Statute‘s constitutionality under the First and Fourteenth Amendments:
[The Court] must first consider the character and magnitude of the asserted injury to the rights protected by the First and Fourteenth Amendments that the plaintiff seeks to vindicate. It then must identify and evaluate the precise interests put forward by the State as justifications for the burden imposed by its rule. In passing judgment, the Court must not only determine the legitimacy and strength of each of those interests; it also must consider the extent to which those interests make it
necessary to burden the plaintiff‘s rights. Only after weighing all these factors is the reviewing court in a position to decide whether the challenged provision is unconstitutional.
This inquiry is “flexible” and “depends upon the extent to which a challenged regulation burdens First and Fourteenth Amendment rights.” Burdick, 504 U.S. at 434. Laws imposing only “reasonable, nondiscriminatory restrictions” or “modest” burdens are usually justified by a state‘s “important regulatory interests.” Anderson, 460 U.S. at 788; Alcorn, 826 F.3d at 716–17 (citation omitted). On the other hand, laws imposing “severe” burdens are subject to “strict scrutiny” and must be “narrowly drawn to advance a state interest of compelling importance.” Alcorn, 826 F.3d at 717 (citations omitted). The class of laws facing strict scrutiny is limited because “subjecting too many laws to strict scrutiny would unnecessarily tie the hands of States seeking to assure that elections are operated equitably and efficiently.” Id. (cleaned up) (citing Burdick, 504 U.S. at 433).
The Fourth Circuit used the Anderson/Burdick standard to evaluate Virginia‘s three-tiered ballot ordering system in Libertarian Party of Virginia v. Alcorn, 826 F.3d 708 (4th Cir. 2016). For a first-tier position, a candidate‘s party must have received at least ten percent of the total votes for any statewide office filled in either of the two preceding statewide general elections. Id. at 712. Only the Republican and Democratic parties met this requirement. Id. The second tier listed other recognized political parties, including the Libertarian Party, and the third tier included independent candidates. Id. The order of candidates in the first and second tiers was set by lot, and each political office on the ballot replicated that order. Id. Candidates in the third tier were listed alphabetically by surname. Id. The appellant, a figure in the Libertarian Party, argued the statute conferred an unconstitutional advantage on Republican and Democratic candidates while disadvantaging minor-party candidates. Id. at 714. The appellant conceded that the burden created by the law did not warrant strict scrutiny, and the court upheld it because the law only imposed a slight burden
While Alcorn is a helpful example for applying the Anderson/Burdick standard, the Fourth Circuit addressed a ballot statute significantly different than West Virginia‘s. The core issue in Alcorn was Virginia‘s privileging of major parties over minor ones, and it is well-established that the “dissimilar treatment of . . . dissimilar groups does not violate the Equal Protection Clause.” Kolbe v. Hogan, 813 F.3d 160, 190 (4th Cir. 2016), on reh‘g en banc, 849 F.3d 114 (4th Cir. 2017). The court also held the statute‘s tiered structure was facially neutral and nondiscriminatory because all parties were “subject to the same requirements” and had “an evenhanded chance at achieving political party status and a first-tier ballot position.” Alcorn, 826 F.3d at 717. The court determined that the appellant exaggerated the difficulty of entering the first tier because first-tier status only required candidates for any office to receive ten percent of the vote in either of the two preceding statewide general elections. Id. Once a party met this requirement, the statute determined its position in the first tier by lot. Id. at 712. Therefore, no party was “automatically elevated to the top of the ballot.” Id. at 717. The Fourth Circuit made the broad observation that “[a]ccess to a preferred position on the ballot so that one has an equal chance of attracting the windfall vote is not a constitutional concern.” Id. at 719 (quoting New Alliance Party v. N.Y. State Bd. of Elections, 861 F. Supp. 282, 295 (S.D.N.Y. 1994)). But this observation must be understood in light of the court‘s finding that Virginia‘s law was politically neutral and nondiscriminatory, except for making the legitimate distinction between dissimilar major and minor parties. While the benefit of an advantageous ballot position is not necessarily a constitutional concern, a discriminatory allocation of that benefit among similarly situated political parties is.
B. Character and Magnitude of the Injury
The Court‘s first step under the Anderson/Burdick standard is to “consider the character and magnitude of the asserted injury to the rights protected by the
As to the injury‘s magnitude, the plaintiffs presented Dr. Jon A. Krosnick. Dr. Krosnick is a renowned political scientist and professor at Stanford University, where he runs a research
Dr. Krosnick presented a comprehensive review of decades of peer-reviewed academic research on the primacy effect. He examined studies of name order effects in general, primary, and presidential elections in the United States, as well as studies from abroad. Tr. 124–37, 139–40; Pls.’ Ex. 2, at 9–20. These studies examine many types of elections across many years using a diversity of research methods and data sets. Tr. 137, 140–42. Included were studies of elections in states that rotate candidates’ names, which provide the strongest evidence for the primacy effect. Tr. 156. For example, Dr. Krosnick reviewed his prior research on the 1992 election in Ohio where candidates’ names were rotated by precinct. Tr. 125–31. Examining 118 races in Ohio‘s 3 largest
Dr. Krosnick offered two basic reasons for why the primacy effect manifests in elections. First, voters often lack information about specific races. The order of candidates’ names can nudge these low-information voters toward choosing the top-listed candidate. Tr. 147–49; Pls.’ Ex. 2, at 24–26. Second, voters may be ambivalent in their choice between candidates. Even though these voters may be highly informed, their ambivalence makes them susceptible to the candidates’ order nudging them toward the top choice. Tr. 149–50; Pls.’ Ex. 2, at 27–28. This explains why the primacy effect still manifests in high information races like the 2016 presidential election between Hillary Clinton and Donald Trump. Tr. 149–50. The defendants did not dispute the literature supporting these two explanations for the primacy effect.
To assess the susceptibility of West Virginia voters to the primacy effect, Dr. Krosnick compared voters in West Virginia to voters in Ohio, California, North Dakota, and New Hampshire. Tr. 155–57; Pls.’ Ex. 2, at 33–35, 48–64. He chose these four states for comparison because they use rotational ballot ordering systems that allow researchers to produce the strongest evidence for the existence and size of the primacy effect. Tr. 156. Dr. Krosnick compared the states by looking at voters’ sex, age, race, ethnicity, marital status, education, employment status, household income, household size, home ownership, active duty service, metropolitan status, and
Finally, Dr. Krosnick did a far-reaching statistical regression analysis of West Virginia elections from 1960 to 2018.8 Tr. 159–77; Pls.’ Ex. 4, at 8–10, ECF No. 104-4. His data included all general election partisan races for federal offices (President, Senator, and Representative) and high-profile state offices (Governor, Attorney General, Treasurer, Secretary of State, State Auditor, Commissioner of Agriculture, State Senator, State Delegate, Supreme Court Justice, Circuit Court Judge, and Family Court Judge) that included a Democrat and Republican. Tr. 159; Pls.’ Ex. 2, at 36; Pls.’ Ex. 4, at 4–5. Dr. Krosnick only excluded a relatively small number of races in which voters could choose more than one candidate from the same party because he was unable to determine the order of same-party candidates on the ballot. Tr. 161–62. This nearly 60-year collection of elections included 21,196 vote totals, a massive number of data points allowing for a high degree of analytical accuracy. Tr. 175; Pls.’ Ex. 3, ECF No. 104-3. To isolate the impact of the primacy effect, Dr. Krosnick controlled for 33 different variables, including political party, incumbency, gender, the unemployment rate, and party affiliation with the governor. Tr. 166–74; Pls.’ Ex. 4, at 8–10. From his regression, Dr. Krosnick determined the average benefit for a
Putting all of this research together, Dr. Krosnick testified to a reasonable degree of certainty that there is an 80 to 85 percent probability of the primacy effect occurring in every West Virginia election, and the best scientific estimate of that effect in each election is 2.94 percentage points. Tr. 122–24, 200–01, 378, 416, 428. Because one candidate gains votes at the expense of another, the 2.94 figure is doubled to determine the primacy effect‘s ultimate impact on the margin of victory in a race. Tr. 326–28; Pls.’ Ex. 2, at 8–9. In other words, the outcome of an election in which the first-listed candidate won by less than 5.88 percentage points would likely have been different if the other major candidate had been listed first. Pls.’ Ex. 2, at 41–43. At trial, the defendants challenged the applicability of Dr. Krosnick‘s conclusions to Nelson‘s 2018 race because that race involved several candidates from each party and was therefore excluded from Dr. Krosnick‘s data. See Tr. 161–62. However, Dr. Krosnick testified that because his analysis includes such a vast range of elections, his conclusions can be confidently generalized to all partisan races in West Virginia, including multi-candidate district races. Tr. 162–63, 200. In fact, he testified that the name order effects in multi-candidate district races are likely stronger because voters cannot use party affiliation to distinguish between candidates of the same party. Tr. 162–63. Thus, Dr. Krosnick concluded that Nelson faced in 2018, and will face in 2020, an 80 to 85 percent probability of the primacy effect occurring with the best estimate of that effect being 2.94 percentage points. Tr. 404–05, 425.
To challenge Dr. Krosnick‘s conclusions, the defendants called Dr. Jason A. MacDonald as their expert witness. Dr. MacDonald earned his PhD in political science from George Washington University and is now an associate professor of political science at West Virginia University. Tr. 432–33; Sec‘y‘s Ex. 30, at 1–2, ECF No. 105-42. His research focuses on American national political institutions, and he primarily uses quantitative methods, including those used by Dr. Krosnick to study the primacy effect. Sec‘y‘s Ex. 30, at 1. Dr. MacDonald has published his research in peer-reviewed academic journals, and he teaches quantitative political analysis to graduate students at West Virginia University. Tr. 433–37; Sec‘y‘s Ex. 30, at 2–3. He does not, however, have research experience with the primacy effect in elections. Tr. 440; Sec‘y‘s Ex. 30, at 3.
The scope of Dr. MacDonald‘s testimony was limited. He did not evaluate Dr. Krosnick‘s literature review, meta-analysis, or opinions on the psychological bases for the primacy effect. Tr.
After listening to the expert witnesses, it is evident to the Court that accurately measuring the primacy effect requires a high degree of skill and experience given that slight changes in modeling can significantly alter results. It is also evident that the more one studies a phenomenon, the more adept that person becomes at choosing the appropriate method to draw conclusions from a large data set. The Court does not assume Dr. Krosnick‘s credibility based on his qualifications, but his professional achievements demonstrate that his methods for studying the primacy effect have been subjected to the highest academic scrutiny. And while the Court has no doubts regarding Dr. MacDonald‘s competency in quantitative research methods, his inexperience in studying the primacy effect in elections seriously undermines the reliability of his opinions in this case. Dr. MacDonald has never studied name order effects or published a research paper on the topic. Tr. 526. He has not thoroughly reviewed the existing literature on the primacy effect in elections. Tr. 439, 526. And the data set compiled by Dr. Krosnick for this case is the only data set related to the primacy effect and voter behavior that Dr. MacDonald has ever looked at. Tr. 529.
Having considered the criticisms raised by Dr. MacDonald, the Court concludes that none of them convincingly call into question Dr. Krosnick‘s methodological judgments. In his direct examination, Dr. Krosnick anticipated and responded to many of these criticisms. For example, in
For these reasons, the Court finds Dr. Krosnick‘s methods and conclusions to be far more credible, reasonable, and reliable than Dr. MacDonald‘s, and the Court confidently adopts Dr. Krosnick‘s conclusion that first-listed candidates in West Virginia are overwhelmingly likely to receive an approximately 2.94 percentage point advantage. To appreciate the magnitude of the burden created by this effect, the Court must look at its impact on election outcomes. The 2.94 figure is doubled to determine the primacy effect‘s impact on the margin of victory in a race because the votes gained by one candidate are necessarily lost from the other candidate. Tr. 326–28; Pls.’ Ex. 2, at 8–9. Therefore, the Court can reasonably conclude that the outcome of every election in which the first-listed candidate won by a margin of less than 5.88 percentage points would have been different if the other major party had been listed first instead. Pls.’ Ex. 2, at 41–43. Dr. Krosnick identified 105 elections between 1960 and 2018 that would likely have changed if the order of major-party candidates had been flipped. Tr. 197–99, 399–401; Pls.’ Ex. 2, at 43, 69–71. These elections include races for President, U.S. Senator, U.S. Representative, State Senator, State Delegate, Secretary of State, Attorney General, Commissioner of Agriculture, and
In conclusion, the plaintiffs proved that West Virginia‘s Ballot Order Statute burdens their
C. State‘s Interests
The Court‘s second step under the Anderson/Burdick standard is to “identify and evaluate the precise interests put forward by the State as justifications for the burden imposed” by the Ballot
The Secretary of State‘s General Counsel Donald Kersey articulated several state interests served by Ballot Order Statute. Many of these interests overlap, and they can be grouped into two basic categories: the effective administration of elections and voting efficiency. First, the Statute serves the effective administration of elections by creating a simple process for determining ballot order that can be easily implemented across the state. Tr. 265–66. In contrast, a rotational ballot ordering system would be more complicated and expensive to administer, including in the preparation of sample ballots. Tr. 276. Second, the Statute serves voting efficiency by creating party-symmetry on the ballot. Tr. 266–67. Because many voters rely on candidates’ party affiliation when voting, a symmetrical ballot allows voters to quickly identify and mark their preferred candidates. Tr. 266–67. Symmetrical ballots thus reduce confusion and errors and speed up the voting process. Tr. 266–67. According to Kersey, these efficiencies ultimately increase voters’ confidence in the integrity of elections. Tr. 265, 277–79.
The Fourth Circuit affirmed the importance of these interests in Libertarian Party of Virginia v. Alcorn. 826 F.3d at 719–21. The court explained that ordering candidates by party “makes the ballot more easily decipherable” and allows voters to “quickly find their preferred choice for a given office, especially when party loyalties influence many voters’ decisions.” Id. Random ordering, on the other hand, can create confusion and “risks requiring voters to decipher lengthy multi-office, multi-candidate ballots in order to find their preferred candidates.” Id. at 719. And “[f]or each extra minute that a voter spends deciphering his ballot in the voting booth, dozens
While the defendants here assert these same interests, there are two important differences between Alcorn and this case. First, the Fourth Circuit held in Alcorn that Virginia‘s ballot order statute imposed “only the most modest burdens” on the appellant‘s constitutional rights, so the state faced a very low bar in justifying the statute. Id. at 717. The constitutional burden imposed here by the Ballot Order Statute is far more significant and requires a weightier set of justifications. Second, in addition to the state‘s more general interests involving voting efficiency, the defendants in Alcorn articulated another state interest—the preservation of political stability—that specifically justified the division between major and minor parties that gave rise to the alleged constitutional burden. Id. at 720. The court held this division served Virginia‘s interest in political stability because, by enacting a regulation that favored the traditional two-party system, Virginia could “temper the destabilizing effects of party-splintering and excessive factionalism.” Id. (citation omitted).
The Eighth Circuit‘s recent decision in Pavek v. Donald J. Trump for President, Inc. is also a helpful example of a state articulating interests that justify the specific feature of the statute producing the constitutional burden. 2020 WL 4381845, at *2–3. At issue in Pavek was Minnesota‘s ballot order statute that lists major-party candidates in reverse order of the parties’ average vote count in the last general election. Id. at *1. The district court granted a preliminary injunction against the statute, but the Eighth Circuit stayed the injunction pending appeal. Id. at *3. The court explained that the statute was likely constitutional because the state‘s interests of encouraging political diversity, countering the advantages of incumbency, and discouraging sustained single-party rule all reasonably supported listing candidates in reverse order of
In contrast to both Alcorn and Pavek, the state‘s interests here do not justify the specific feature of the Ballot Order Statute challenged by the plaintiffs. The burden on the plaintiffs’ constitutional rights stems from the Statute assigning ballot order based on candidates’ partisan affiliation with the presidential candidate who received the most votes in the last election. While all of the state‘s asserted interests are important, none of them actually support using the Statute‘s partisan criteria over any other. The state‘s interests would be just as well served by determining candidates’ order by lot or by using some other nondiscriminatory criteria to create a statewide and party-symmetrical ballot order. For this reason, the state‘s asserted interests do not “make it necessary to burden the plaintiff‘s rights” in any sense of the phrase. Anderson, 460 U.S. at 789. Therefore, while the state‘s interests are important and served generally by the statute, the Court finds them to be universally weak as justifications for the actual burden the Statute imposes.
D. Weighing the Injury Against the State‘s Interests
As its final step under the Anderson/Burdick standard, the Court must determine which level of scrutiny to apply and weigh the burdens on the plaintiffs’ rights against the defendants’ asserted justifications. Anderson, 460 U.S. at 789; Burdick, 504 U.S. at 433–34. Laws imposing “severe” burdens are subject to “strict scrutiny” and must be “narrowly drawn to advance a state interest of compelling importance.” Alcorn, 826 F.3d at 717 (citations omitted). On the other hand, laws imposing only “reasonable, nondiscriminatory restrictions” or “modest” burdens are usually justified by a state‘s “important regulatory interests.” Anderson, 460 U.S. at 788; Alcorn, 826 F.3d at 716–17 (citation omitted). These two forms of scrutiny, however, are only two markers under the “flexible” standard. Burdick, 504 U.S. at 434. The “rigorousness” of the inquiry ultimately
Strict scrutiny does not apply here. As the Fourth Circuit explained, “the class of laws facing [strict scrutiny] is limited” because “[s]ubjecting too many laws to strict scrutiny would unnecessarily tie the hands of States seeking to assure that elections are operated equitably and efficiently.” Alcorn, 826 F.3d at 717 (cleaned up) (quoting Burdick, 504 U.S. at 433). The Ballot Order Statute does not belong to this limited class because it denies “neither the right to vote, nor the right to appear on the ballot, nor the right to form or associate in a political organization.” Id.; see also Green Party of Tenn. v. Hargett, 767 F.3d 533, 547 (6th Cir. 2014) (citation omitted) (explaining a restriction is typically not severe if it “does not ‘affect a political party‘s ability to perform its primary functions,’ such as organizing, recruiting members, and choosing and promoting a candidate“). The Statute does not create an insurmountable burden, and the state has legitimate interests in administering orderly elections to which the Court owes deference.
At the same time, a low form of scrutiny akin to rational basis is also inappropriate. As previously explained, the Statute is not neutral. See Fusaro, 930 F.3d at 263 (“[W]hen a court evaluates an electoral regulation, political neutrality is a touchstone for assessing the burden imposed thereby.“). It elevates candidates to the top of the ballot to receive the benefit of the primacy effect based solely on partisan affiliation. By doing so, the Statute burdens the plaintiffs’ rights in a politically discriminatory way. And while a primacy effect of 2.94 percent is not a large share of total votes, it has been outcome-determinative in dozens of elections, significantly shaping West Virginia politics. The burden created by the primacy effect, therefore, cannot be dismissed as merely “modest.” Alcorn, 826 F.3d at 716–17 (citation omitted).
While the state‘s asserted interests are important, they are extremely weak justifications for the burden imposed by the Statute. The effective administration of elections and voting efficiency are served by a statewide and party-symmetrical ballot order, but they have no relationship to the Statute‘s specific use of partisan criteria to order the ballot. The state‘s interests therefore are not directly relevant to the feature of the Ballot Order Statute causing the
IV. PERMANENT INJUNCTION
Having concluded the Ballot Order Statute is unconstitutional, the Court now considers the plaintiffs’ request to permanently enjoin the defendants from enforcing it and to adopt a ballot ordering system that gives similarly situated major-party candidates an equal opportunity for the ballot to list them first. Am. Compl. 17. To obtain a permanent injunction, the plaintiffs must demonstrate: (1) they suffered an irreparable injury; (2) their remedies at law are inadequate; (3) the balance of hardships weighs in their favor; and (4) a permanent injunction would not disserve
These elements are easily satisfied. First, “[t]he loss of First Amendment freedoms, for even minimal periods of time, unquestionably constitutes irreparable injury.” Elrod v. Burns, 427 U.S. 347, 373 (1976) (plurality opinion) (citation omitted). And “[c]ourts routinely deem restrictions on fundamental voting rights irreparable injury” because “once the election occurs, there can be no do-over and no redress.” League of Women Voters of N.C. v. North Carolina, 769 F.3d 224, 247 (4th Cir. 2014) (citations omitted). Second, monetary damages cannot remedy the violations of the plaintiffs’ constitutional rights. See Legend Night Club v. Miller, 637 F.3d 291, 302 (4th Cir. 2011) (citation omitted) (“[M]onetary damages are inadequate to compensate for the loss of First Amendment freedoms.“). Third, the balance of hardships weighs in the plaintiffs’ favor because the defendants are “in no way harmed by issuance of an injunction that prevents the state from enforcing unconstitutional restrictions.” Id. at 302–03 (citation omitted). And fourth, “upholding constitutional rights is in the public interest.” Id. at 303 (citation omitted); see also Newsom ex rel. Newsom v. Albemarle Cty. Sch. Bd., 354 F.3d 249, 261 (4th Cir. 2003) (“Surely, upholding constitutional rights serves the public interest.“).
The more complicated issue is deciding the injunction‘s terms. “It is well established . . . that a federal district court has wide discretion to fashion appropriate injunctive relief in a particular case.” Richmond Tenants Org., Inc. v. Kemp, 956 F.2d 1300, 1308 (4th Cir. 1992) (citing Lemon v. Kurtzman, 411 U.S. 192, 200 (1973)). Here, however, the Court must consider both the need to remedy the plaintiffs’ injury as well as the West Virginia legislature‘s authority to regulate elections. As the Fourth Circuit explained in Libertarian Party of Virginia v. Alcorn, extensive debate among the Framers of the Constitution produced a federal system that grants each state‘s
In light of these principles, the Court concludes that a permanent injunction prohibiting enforcement of the Ballot Order Statute is appropriate but requiring the defendants to adopt a particular alternative is not. Many ballot ordering schemes exist that would not systematically award the advantage of the primacy effect based on party affiliation or impose significant burdens on the state. To mandate one of the many possibilities would be too great of an intrusion into the sovereignty of West Virginia‘s political branches. See Sangmeister v. Woodard, 565 F.2d 460, 467–68 (7th Cir. 1977) (holding the district court abused its discretion in ordering county clerks to adopt a specific rotational system for ballot placement); Gould v. Grubb, 536 P.2d 1337, 1347 (Cal. 1975) (“[W]e do not believe that it would be appropriate for this court to mandate a single form of procedure that must be followed in every election.“). The Court therefore limits its injunction to prohibiting the defendants from implementing the Ballot Order Statute. By virtue of
The Court must also decide whether to give the injunction immediate effect considering the imminency of the general election. See Reynolds v. Sims, 377 U.S. 533, 585 (1964) (explaining that “under certain circumstances, such as where an impending election is imminent and a State‘s election machinery is already in progress, equitable considerations might justify a court in withholding the granting of immediately effective relief“). At trial, the Court expressed its inclination to withhold relief until after the election because a short period of time remains before the ballot order is set on August 25.
V. CONCLUSION
In accordance with the reasoning above, the Court DECLARES that
The Court DIRECTS the Clerk to send a copy of this Memorandum Opinion and Order to all counsel of record.
ENTER: August 10, 2020
ROBERT C. CHAMBERS
UNITED STATES DISTRICT JUDGE