The plaintiffs are African-American and Caucasian voters residing in Hamilton, Montgomery, Sandusky, and Summit Counties in Ohio. They filed their complaint on October 11, 2002 alleging that: (1) the use of unreliable, deficient voting equipment, including the punch card ballot, in some Ohio counties but not other counties violates the Equal Protection Clause of the Fourteenth Amendment; (2) the use of error prone voting equipment deprives voters of their due process right to have their votes counted accurately; and (3) the use of punch card voting systems in Hamilton, Montgomery, and Summit Counties has a disparate impact on African-American voters in violation of Section 2 of the Voting Rights Act of 1965. The plaintiffs sought declaratory and injunctive relief prohibiting the defendants from: (1) continuing to allow the use of "non-notice" and deficient punch card and optical scan voting equipment in some Ohio counties while using more reliable voting equipment in other counties; (2) using non-notice punch card voting equipment in Hamilton, Montgomery, and Summit Counties; and (3) using non-notice optical scan voting systems in Sandusky County.

On December 14, 2004, the district court rejected the plaintiffs' claims and granted summary judgment in favor of the defendants. Some commentators have suggested that these types of voting rights challenges are taking us into a brave new world. Others suggest that they are simply variations of old challenges. Regardless of the proper characterization, we find ourselves bound by Supreme Court precedent, and therefore, with regard to the plaintiffs' claim under the Equal Protection Clause, we REVERSE the district court's judgment. With regard to the plaintiffs' claim under the Voting Rights Act, we VACATE the district court's judgment and REMAND for proceedings consistent with this opinion.

A. Background Information on Voting Technology

Ohio law empowers the Secretary of State to certify voting equipment. Ohio Rev.Code § 3506.15.1 The Secretary has certified two general types of equipment: (1) "Notice" equipment such as Digital Recording Electronic (DRE) and precinct-count optical scan equipment that prevent overvotes (when a voter votes for more than the permissible number of candidates for a given office) and warn voters when they are casting undervotes (when a voter does not vote in a particular race or votes for fewer candidates than is permissible for a given office) — together, overvotes and undervotes are referred to as "residual votes"; and (2) "Non-notice" equipment such as punch card and central-count optical scan equipment that do not provide notice of and the opportunity to correct residual votes. In the 2000 general election, approximately 72.5% of Ohio voters used non-notice equipment and 27.5% used notice equipment.

In the 2000 general election, the most frequently used equipment in Ohio was the Votomatic punch card, a non-notice system that relies on a ballot card with pre-scored, square perforations or "chad" that correspond to the names of the candidates listed in an accompanying booklet. Names of candidates or other identifying information for ballot measures do not appear on the actual ballot. The punch card system does not provide independent notice of an overvote or undervote. A vote is recorded by the machine when light passes through the detached holes. Problems with the machines can cause "hanging chad" that remain attached to the ballot by one, two, or three corners; "pierced chad" that are penetrated by the stylus but not dislodged from the ballot; and "dimpled chad" that are dented but not penetrated or dislodged. Because of these inherent chad problems, light often cannot pass through the holes and a vote is not recorded. Problems inherent in the punch card machines are sometimes caused by the build up of chads which may make it difficult or impossible to cleanly punch the card and record a vote.

Optical scan systems resemble answer sheets used in standardized testing. The voter is given a ballot listing the names of all candidates and ballot initiatives and either uses a pencil to darken the circle next to the preferred candidate or draws a straight line connecting two parts of an arrow. Optical scan systems can be either precinct-count systems, which enable voters to scan the ballot at the polling place thereby providing independent notice of and an opportunity to correct residual votes, or central-count systems, which do not provide independent notice or the opportunity to correct mistakes.

Electronic DRE machines come in several varieties, but most often resemble automated teller machines or ATMs used at banks. Voters either touch the name of the preferred candidate on the screen or press a button that corresponds to the preferred candidate. All forms of DRE technology currently used in Ohio make it impossible to overvote for the same office or ballot initiative. DRE systems can also be programmed to warn voters if their ballots contain undervotes. DRE systems (like precinct-count optical scan systems), therefore, provide independent notice of residual votes.

In the 2000 general election, sixty-nine of eighty-eight Ohio counties used punch card ballots. Eleven counties used optical scan equipment, six used electronic equipment, and two used automatic or "lever" voting machines. These systems utilize different methods of reading and counting votes. Some of the systems allow voters to check their ballots for residual votes. For example, one county and part of another county utilized precinct-count optical scan equipment, and six others use electronic voting equipment that allows a voter to verify their ballot on a screen before the final ballot is cast. Most systems, however, including the ones operated by the four county defendants, scan and count ballots at a central location after the polls have closed. Thus, in total, eighty-one of eighty-eight Ohio counties used non-independent-notice equipment — voting technology that does not provide a voter with notice from the voting device that a problem might exist before the ballot is finally cast — in the 2000 general election.

Only three counties collected statistics on overvotes — Hamilton County, which had 2,916 overvotes, Summit County, which had 1,470 overvotes, and Montgomery County, which had 2,469 overvotes. This is a total of 6,855 overvotes in those three counties, which represents approximately 34% of the total residual votes cast in those counties. Franklin County used notice technology and there were zero overvotes.

The plaintiffs' expert, Dr. Martha Kropf, testified regarding estimates of intentional and unintentional undervoting based on data collected by National Elections Studies and the Voters News Survey in exit polls and surveys in presidential elections between 1980 and 1996. Kropf testified that intentional undervoting in presidential elections is a relatively rare event that is estimated to involve between .23% and.75% of all residual votes. Dr. Kropf concluded that when levels of undervoting exceed this threshold and vary by equipment it is probable that they resulted from unintentional undervoting that is associated with problems of the punch card ballot. She also found no difference between African-American and non African-American voters in levels of intentional undervotes. Kropf measured the performance of voting equipment by examining presidential and U.S. Senate races at the top of the ballot because these are statewide elections where all voters face the same candidates, and media coverage, levels of candidate competition, and voter mobilization are relatively uniform. Kropf reported an overall statewide residual vote rate of 2.29% for punch card systems and 2.14% for central-count optical scans. That is, voters in punch card counties are approximately four times as likely not to have their votes counted as a voter using reliable electronic voting equipment. In some counties specific precincts encountered more severe problems with residual voting. In Akron City Precinct 3-F the residual vote rate was 15% and in Dayton City's 14th Ward Precinct C the residual vote rate was 17%.2 In addition, the counties in Ohio experiencing the highest percentage of residual votes in the 2000 presidential election were those in which voters used punch card technology while the counties experiencing the lowest percentage of residual votes used other technology. The twenty-nine counties in Ohio with the highest residual vote percentages were all counties that used punch card machines; the seven counties with the lowest residual vote percentages were all counties that did not use punch card machines as their primary voting system.

Roy Saltman, formerly of the National Bureau of Standards and the author of two federal studies on the use of computers in vote tallying testified that his studies "demonstrate that punch cards are inherently fragile, and they become less stable when ballots are handled or manipulated or sent through a reader, resulting in overvotes, undervotes, and inconsistent vote tabulations." Saltman explained that lost votes are not attributable solely to voters' failure to follow instructions. According to Saltman, "[w]hen the ballot is then handled or manipulated or sent through a reader, it is more likely that additional chads will be dislodged and fall out. And if that happens, the votes indicated on the ballot are changed because the presence of holes indicates votes." Dana Walch, Director of Election Reform for the Ohio Secretary of State's Office, confirmed the plaintiffs' evidence that there is "a higher residual vote rate in punch card counties than in ... counties with other types of voting technology." Walch further testified that problems with the punch card ballots "were the result of some physical failure of the ballot or voter error." The record also includes letters from Tim Burke, the Chairman of the Board of Elections in Hamilton County, which state that:

While I continue to believe that the punch card system functions reasonably well for the price, it does have faults that the electronic machines do not have. I believe that the chad problem is a very small part of our problem here in Hamilton County. It does exist though. In fact, it is an inherent part of the punch card system. On the other hand, chads are totally eliminated with electronic systems.

Another letter authored by Burke to U.S. Senator Mike DeWine stated that:

Having looked closely at the punch card system of voting we use here in Hamilton County, I am convinced that this outdated technology is having a disparate impact in depriving a significant number of voters of having their electoral choices given.... Newer technology, particularly the touch screen voting systems, provide both a more accurate count and prevent a voter from [overvoting].

Finally, a third letter from Burke stated that:

I am not as confident that our punch card voting system tells us with precise accuracy the number of votes any particular individual received or should have received if the intent of the voter had been properly accounted for.... In the punch card system you can and it is objectively provable that hundreds and hundreds of people lose their vote each year.

In response to this evidence, the district court acknowledged that "running the punch card ballots repeated times through the counting machinery will result in different results."

The Caltech MIT Voting Technology Project report — a joint venture between the two institutions to study, in part, the reliability of existing voting equipment — which is referenced throughout the record, is also informative. As the report notes, that "[i]f voting equipment has no effect on the ability of voters to express their preferences, then the residual vote should be unrelated to machine types." See Caltech-MIT Voting Technology Project, Residual Votes Attributable to Technology: An Assessment of the Reliability of Existing Voting Equipment (Version 2: March 30, 2001), available at http://www.vote.caltech.edu (last accessed April 1, 2006). The report concluded that the error rate from punch cards is 50 percent higher than other technologies and that the pattern holds up when "holding constant turnout, income, racial composition of counties, age distribution of counties, literacy rates, the year of a shift in technology, the number of offices and candidates on the ballot, and other factors that operate in a county or in a particular year." Report at 22. In conclusion, the report stated that "[t]he incidence of such residual votes with punch card methods ... is forty to seventy percent higher than the incidence of residual votes with the other technologies," Report at 17, and cautioned that "[i]f election administrators wish to avoid catastrophic failures, they may heed th[is] warning ... Stop using punch cards," Report at 11.

Defendants' expert Dr. John Lott examined the performance of voting technology across three election cycles, 1992, 1996, and 2000, in the presidential, Congressional, Ohio Senate, and Ohio House elections. His findings for the presidential and U.S. Senate elections mirror Kropf's in the 2000 election. Lott reported an overall statewide residual vote rate of 2.4% for punch card systems.

Dr. Lott's report stresses findings from the down-ballot contests — contests listed below the presidential candidates farther down the ballot — that were not uniform across the state and included non-competitive and uncontested elections. The experts on both sides appear to agree that the most likely explanations for the fall-off in down-ballot voting are that voters deliberately choose not to vote or that elections were not competitive. Lott's report does not include specific findings on overvotes. Further, the defendants' experts made no effort to distinguish intentional from unintentional undervoting in the down-ballot elections. Lott also admitted that a review of the literature revealed that no other major scholars have agreed with his findings regarding the relevance of analyzing non-uniform down-ballot races.

The State's Help America Vote Act report also provides evidence of the problem of deficient electoral technology. At the outset, the report notes that "[p]ublic confidence in the accuracy of punch card voting systems has been seriously undermined." Thus, "Boards of election should upgrade their voting systems to new, more trustworthy technology." Additionally, "[t]hese goals demand immediate attention, or our state runs the risk of repeating the problems of our nation's most recent presidential election — and suffering irreparable damage to the most important and basic concepts of democracy." Regarding the deficient technology currently in use throughout the State, the report stated that:

"While the Secretary of State notes that punch-card voting is not explicitly prohibited under the Help America Vote Act, other requirements of the Act make it impractical to use punch card voting as a primary voting device in the state. In a study of `over' and `under' voting in Ohio, it was clearly demonstrated that punch-card voting was unreliable to the extent [that] votes cast by thousands of Ohioans were not being counted in the final election tabulation."

(Emphasis added). Additionally, the report stated that:

As election officials, if we know voters are disenfranchised and that legitimately cast ballots are being discounted, we have not only a moral obligation to immediately embrace a solution, but a legal obligation to find a remedy and enact measures to prevent that from happening. If even one voter is denied the right to vote, we are obligated, by law, to determine the cause and forge a solution. The evidence is overwhelming that thousands of Ohio voters have been disenfranchised by antiquated voting equipment and that many thousands have lost confidence in the reliability and accuracy of voting devices currently in use in most of Ohio's 88 counties.

(Emphasis added).

Finally, in response to the legislature's slow response to the electoral problems, the Secretary of State wrote a letter stating that "the possibility of a close election with punch cards as the state's primary voting device invites a Florida-like calamity."

On their Voting Rights Act claim, the plaintiffs alleged that the punch card system used in Hamilton, Montgomery, and Summit Counties produces a higher residual vote rate for African-American voters than for white voters. The plaintiffs presented regression analysis that the correlation between overvoting and the percentage of African-American voters in a given precinct in Hamilton County was .517 and in Summit County it was .682.3 The plaintiffs' experts characterized these correlations as "strong." In Montgomery County, where only data of overvotes mixed with undervotes was available at the precinct level, there was a smaller, but nevertheless "strong" .440 correlation.

The plaintiffs' expert, Dr. Richard Engstrom, analyzed the data based on methods of statistical analysis approved by the Supreme Court and other federal courts in voting rights cases. See e.g., Thornburg v. Gingles, 478 U.S. 30, 106 S. Ct. 2752, 92 L. Ed. 2d 25 (1986); Mallory v. Ohio, 173 F.3d 377 (6th Cir.1999). The three methods are homogeneous precinct analysis, ecological regression, and ecological inference. Dr. Engstrom used all three methods and triangulated among them to verify that his findings moved consistently in the same direction; the defendants' expert did not use any of these methods. Dr. Engstrom concluded that: (1) African-Americans in Hamilton County overvoted at a rate seven times higher than non African-Americans; (2) in Summit County, African-Americans overvoted at a rate nine times higher than non African-Americans; and (3) in Montgomery County (where only combined over and undervote statistics are available on a precinct basis), African-Americans had a residual voting rate 2.5 times that of non African-Americans. In contrast to the three punch card counties, Franklin County had no overvotes because it used DRE machines that prevent overvoting.

Based on this information, Dr. Engstrom testified that punch card equipment interacts with socioeconomic conditions, resulting in statistically significant disparities between the levels of residual voting among African-American and non African-American voters.

The defendants contend that the plaintiffs did not provide factual evidence to prove a violation of the Constitution or Voting Rights Act. With heavy emphasis on the plaintiffs' stipulation that they were not denied physical access to the polls, the defendants seem to allege that the plaintiffs were not denied the right to vote. State Defendant's Br. at 8. The defendants attack the plaintiffs' expert, Martha Kropf, who testified that intentional undervoting is rare, based on the fact that she relied exclusively on exit polls commissioned by the National Election Survey and the Voters News Service. They further point out that in the 1988 election, the National Election Survey determined that between 3-5% of those who claimed to have voted had no record of actually voting in the election. Thus, according to the defendants, the data the plaintiffs relied on consistently over-reports voter turnout — thereby inflating the residual vote statistics. The defendants also criticize the plaintiffs for not examining the voting results of the entire state and instead focusing on Hamilton, Sandusky, Summit, and Montgomery Counties. The defendants further claim to poke holes in the plaintiffs' theory by pointing out that the plaintiffs' own study shows that African-Americans using punch card technology in Hamilton County had a lower residual vote rate than non African-Americans using the punch card technology in Summit County.

The defendants also point to their expert, Dr. John Lott, who examined the 1992, 1996, and 2000 presidential elections and down-ballot races such as U.S. Senate, U.S. Congress, state legislature races, and local races. Dr. Lott found that punch cards outperformed electronic machines in congressional races, that punch cards reliability improved for down-ballot races relative to other technologies, and that punch cards produced fewer non-voted ballots for 1992, 1996, and 2000 races than either electronic machines or lever machines and produced virtually the same results as optical scan machines.

The defendants also rely on Dr. Herb Asher's findings that precincts with a higher concentration of poverty had a residual vote rate higher than average in Ohio. Thus, Dr. Asher pointed to race, education, and poverty as the cause. Dr. Asher also pointed to Appalachian counties where the population is less than one percent African-American and concluded that because there is a residual vote rate in those counties, education and income levels are contributing factors and one factor alone — race — cannot explain everything.

Prior to the trial, the district court denied the plaintiffs' motion for class certification. They appeal this decision. The district court then conducted a four-day bench trial and found in favor of the defendants. At the outset, the district court laid out the parties' arguments — it noted that according to the plaintiffs' expert, Dr. Kropf, the residual vote rate was 2.29% for punch cards, 1.15% for precinct-count optical scan, 1.04% for lever technology, and 0.94% for DRE technology. The defendants' expert produced similar results: 2.4% for punch cards, 2.0% for optical scan, 1.4% for level technology, and 1.0% for DRE technology. Based on these figures, the plaintiffs' alleged that punch cards and central-count optical scan technology (but not lever, precinct-count optical scan, or DRE technology) violate the Equal Protection and Due Process clauses.

Citing this Court's decision in Mixon v. State of Ohio, 193 F.3d 389 (6th Cir.1999), the district court determined, in a footnote, that rational basis is the appropriate standard for evaluating the voting systems. The footnote concluded by stating: "However, if the Court were to apply strict scrutiny, the Court's ruling would be the same." The court then stated that "[t]he primary thrust of this litigation is an attempt to federalize elections by judicial rule or fiat via the invitation to this Court to declare a certain voting technology unconstitutional and then fashion a remedy. This Court declines the invitation." According to the district court, determination of the voting process has always been left to the legislative branch, and "subject to constitutional amendment, that is where the determination should remain."

The court went on to state that "[t]he use of the punch card voting technology is neither confusing nor difficult to operate."4 Then, taking estimates for intentional undervoting into account and multiplying it by the residual vote rate across the state, the district court concluded that:

Viewing the plaintiffs' case in a light most favorable to them, leads to the conclusion that seven to thirteen voters out of 1000 [i.e., 0.7% to 1.3%] using punch card technology accidentally failed to record a vote in the year 2000 in the presidential election. Such a de minimis conclusion, assuming arguendo that it is justified, fails to prove a constitutional violation, either on a Due Process or Equal Protection analysis.5

Regarding the African-American plaintiffs' Voting Rights Act claim, the district court found that it "fail[s] because their alleged injury does not amount to a vote denial under § 2 of the Voting Rights Act." The court noted that plaintiffs can bring vote denial or vote dilution claims under the Act, but found that the plaintiffs brought only a vote denial claim and that the plaintiffs admitted that they were not denied physical access to the polls. Because, according to the district court, "[a] vote denial arises when a state or municipality employs a `practice or procedure' that results in the `actual' denial of the right to vote on account of race," see 42 U.S.C. § 1973(a), the plaintiffs did not bring a cognizable Voting Rights Act claim.6 The court stated that: "When coupled with the previously referenced de minimis affects of the punch card ballots, these facts do not allow this Court to conclude that an `actual' denial of the right to vote on account of race occurs."

Finally, the court concluded that "the operation of different voting systems by different counties within the same state does not amount to a violation of the Equal Protection Clause." Relying heavily on Justice Souter's dissenting opinion in Bush v. Gore, 531 U.S. 98, 134, 121 S. Ct. 525, 148 L. Ed. 2d 388 (2000) (Souter, J., dissenting), the district court concluded that states may freely use a variety of different voting technologies without running afoul of the Equal Protection Clause. Further, the court concluded that the defendants have a rational basis for continuing to utilize punch card technology in that it is cost effective and there are security concerns with electronic technology.

Before turning to the merits of the case, we address the defendants' argument that the plaintiffs lack standing. The district court stated that it "is of the view that the defendants have the better argument on the issue of standing, but declines the invitation to dismiss the case on standing." The district court erred in not determining whether the plaintiffs had standing. Steel Co. v. Citizens for a Better Environment, 523 U.S. 83, 94-95, 118 S. Ct. 1003, 140 L. Ed. 2d 210 (1998) (courts have an obligation to satisfy jurisdiction, including standing, even though the parties are prepared to concede it). The defendants argue that: (1) because one of the four plaintiffs did not vote in the previous election, she suffered no injury in fact, and (2) no plaintiff has standing to bring suit against Hamilton County, because none of the four voters have ever lived or voted in Hamilton County.

We reject the defendants' arguments and conclude that the plaintiffs have standing to bring suit against the State of Ohio and each of the four counties. It has been stipulated that at least one of the plaintiffs, all of whom are registered voters, resides in each defendant-county. Furthermore, the plaintiffs' standing does not depend on any injury suffered in the previous election, but rather on the probability that their votes will be miscounted in upcoming elections.

In Bryant v. Yellen, 447 U.S. 352, 100 S. Ct. 2232, 65 L. Ed. 2d 184 (1980), the Supreme Court held that a group of farm workers had suffered an injury sufficient to confer standing based on an increased likelihood that land would be available for sale if a section of the Boulder Canyon Project Act had been applied to 160 acres of land in Imperial Valley, California. The Supreme Court concluded that the circuit court was correct in finding standing on this basis, even though the plaintiffs could not establish that they would be able to purchase land if the section of the Act were applied. Id. at 366-67, 100 S. Ct. 2232. The farm workers did in fact have standing, because it was unlikely that any of the land would be available for sale without the applicability of the section of the Act, and likely that the land would be available at less than market prices after the application of the Act. Id. at 368, 100 S. Ct. 2232. Bryant indicates that the increased probability of a future injury is sufficient to confer Article III standing.

The Supreme Court restated the elements of standing in Lujan v. Defenders of Wildlife, 504 U.S. 555, 112 S. Ct. 2130, 119 L. Ed. 2d 351 (1992). There, the Court emphasized that in order to have standing, a plaintiff must have suffered an injury in fact, which must be concrete and particularized, and actual or imminent, as opposed to conjectural or hypothetical. Id. at 560. Lujan, however, does not overrule Bryant. In fact, after Lujan, courts have continued to recognize that the increased risk of harm constitutes an injury sufficient to support standing. For example, although ultimately denying him relief, the Supreme Court considered the merits of the claims brought by a plaintiff who had been exposed to asbestos, but had not yet manifested any symptoms of asbestos-related disease. Metro-North Commuter R. Co. v. Buckley, 521 U.S. 424, 117 S. Ct. 2113, 138 L. Ed. 2d 560 (1997). This Court has also recognized in the medical context that a risk of injury in the future can be grounds for standing. See Sutton v. St. Jude S.C., Inc., 419 F.3d 568, 573-74 (6th Cir.2005) (finding standing based on an increased risk of harm requiring medical monitoring); but see Natural Resources Def. Council v. EPA, 440 F.3d 476 (D.C.Cir.2006).

In the voting context, this Court and others have recognized that voters can have standing based on an increased risk that their votes will be improperly discounted. In Sandusky County Democratic Party v. Blackwell, 387 F.3d 565, 574 (6th Cir.2004) (per curiam), this Court held that the plaintiffs had standing to bring a claim on behalf of voters alleging that the Secretary of State's issuance of provisional ballots in Ohio elections violated the Help America Vote Act. The Act allowed voters to cast provisional ballots in those instances where their names could not be located on the list of qualified voters. Id. at 569. The Secretary of State issued a directive that would prohibit voters from casting provisional votes unless the poll worker was able to confirm that the voter was eligible to vote in that specific precinct. Id. at 571. On behalf of their members, i.e. voters, plaintiffs alleged that the directive violated the Act because the directive would allow "poll workers to withhold a provisional ballot from anyone who is not — according to the poll worker's on-the-spot determination at the polling place — a resident of the precinct in which the would-be voter desires to cast a provisional ballot." Id.

This Court held that the plaintiffs had standing, even though they were unable to name specific voters who would seek to vote at polling places that would be deemed to be wrong by voters:

Appellees have not identified specific voters who will seek to vote at a polling place that will be deemed wrong by election workers, but this is understandable; by their nature, mistakes cannot be specifically identified in advance. Thus, a voter cannot know in advance that his or her name will be dropped from the rolls, or listed in an incorrect precinct, or listed correctly but subject to a human error by an election worker who mistakenly believes the voter is at the wrong polling place. It is inevitable, however, that there will be such mistakes. The issues Appellees raise are not speculative or remote; they are real and imminent.

Id. at 574. So too here, the plaintiffs are unable to articulate which voter will be harmed in the future by deficient equipment.7 It is inevitable, however, that errors have been made and will be made in the future. As the district court found, "[a] flaw in the punch card ballot is its fragile nature and the fact that running the punch card ballots repeated times through the counting machinery will result in different results." The claims of the plaintiffs here are not speculative or remote, but real and imminent.

The plaintiffs here have alleged an injury in fact sufficient to confer Article III standing. The increased probability that their votes will be improperly counted based on punch-card and central-count optical scan technology is neither speculative nor remote.

The defendants also claim that there is no case or controversy because of the Help America Vote Act of 2002. Pub.L. No. 107-252, 116 Stat. 1666 (codified at 42 U.S.C. § 15304). Section 102 of the Act requires States receiving federal funds under the Act to discontinue the use of punch card systems in time for the first federal election of 2006. The Act also contains a section on voting system standards. 42 U.S.C. § 15481. The Act requires independent notice technology, the opportunity to change the ballot before it is cast, and an FEC compliant error rate. Ohio has declared its intentions to comply with the Act and thereby will receive federal funding under the Act. Id.

The mootness doctrine is grounded in Article III's "case or controversy" requirement. Thus, as mootness is a jurisdictional requirement, courts lack judicial power to entertain and decide moot cases. See Los Angeles County v. Davis, 440 U.S. 625, 631, 99 S. Ct. 1379, 59 L. Ed. 2d 642 (1979) ("[A] case is moot when the issues presented are no longer `live' or the parties lack a legally cognizable interest in the outcome.") (quoting Powell v. McCormack, 395 U.S. 486, 496, 89 S. Ct. 1944, 23 L. Ed. 2d 491 (1969)). The standard for determining whether a case has been mooted by the defendant's voluntary conduct "is stringent: `A case might become moot if subsequent events made it absolutely clear that the allegedly wrongful behavior could not reasonably be expected to recur.' The `heavy burden of persua[ding]' the court that the challenged conduct cannot reasonably be expected to start up again lies with the party asserting mootness." Laidlaw, 528 U.S. at 189, 120 S. Ct. 693 (quoting United States v. Concentrated Phosphate Export Ass'n, 393 U.S. 199, 203, 89 S. Ct. 361, 21 L. Ed. 2d 344 (1968)) (emphasis added). The State has asserted that its voluntary compliance with HAVA is sufficient to moot the case. We disagree. Ohio's voluntary compliance with the Act is just that — voluntary. The State, should it wish to opt out of the Act's demands, may do so at any time and the only penalty is the return of a prorated amount of the federal funds in proportion to the precincts not converted to acceptable technology. This is not the type of voluntary action that moots a case, and in fact, is just the type of action that we have found does not carry the "heavy burden" of demonstrating mootness. See e.g., Am. Canoe Ass'n v. City of Louisa Water & Sewer Comm'n, 389 F.3d 536, 543 (6th Cir.2004) (defendants voluntary conduct failed to meet "stringent" test for mootness); Johnson v. City of Cincinnati, 310 F.3d 484, 490 (6th Cir.2002) (city's assurance that it no longer enforced the challenged ordinance deemed insufficient to moot case); Deja Vu of Nashville v. Metro. Gov't of Nashville & Davidson Co., 274 F.3d 377, 387 (6th Cir.2001) (amendment of challenged ordinance did not moot case). We now turn to the merits of the case.

Voting is a fundamental right. See e.g., Reynolds v. Sims, 377 U.S. 533, 561-62, 84 S. Ct. 1362, 12 L. Ed. 2d 506 (1964) ("Undoubtedly, the right of suffrage is a fundamental matter in a free and democratic society."). For more than a century the Supreme Court has acknowledged the fundamental nature of the right to vote. See Yick Wo v. Hopkins, 118 U.S. 356, 371, 6 S. Ct. 1064, 30 L. Ed. 220 (1886) (acknowledging "the political franchise of voting" as "a fundamental political right"). "Especially since the right to exercise the franchise in a free and unimpaired manner is preservative of other basic civil and political rights, any alleged infringement of the right of citizens to vote must be carefully and meticulously scrutinized." Reynolds, 377 U.S. at 562, 84 S. Ct. 1362. Careful and meticulous scrutiny is necessary because even minor infringements on the franchise can have reverberations in other contexts and throughout democratic society. "A consistent line of decisions by [the Supreme] Court in cases involving attempts to deny or restrict the right of suffrage has made this indelibly clear." Id. at 554, 84 S. Ct. 1362 ("Undeniably the Constitution of the United States protects the right of all qualified citizens to vote, in state as well as in federal elections."); see also U.S. Const. amends. XV, XVII, XIX, XXIV, XXVI. For example, all qualified voters have a constitutionally protected right to vote. Ex parte Yarbrough, 110 U.S. 651, 656, 4 S. Ct. 152, 28 L. Ed. 274 (1884). Citizens cannot be denied the right to vote outright. Guinn v. United States, 238 U.S. 347, 362, 35 S. Ct. 926, 59 L. Ed. 1340 (1915); Lane v. Wilson, 307 U.S. 268, 274-75, 59 S. Ct. 872, 83 L. Ed. 1281 ("The Fifteenth Amendment secures freedom from discrimination on account of race in matters affecting the franchise," and "nullifies sophisticated as well as simple-minded modes of discrimination."). Dilution of the right to vote may not be accomplished by stuffing the ballot-boxes. Ex parte Siebold, 100 U.S. 371, 25 L. Ed. 717 (1879). Nor may the right to vote be diluted by alteration of ballots or improper counting of ballots. United States v. Classic, 313 U.S. 299, 315, 61 S. Ct. 1031, 85 L. Ed. 1368 (1941). In Classic, the Court stated that: "Obviously included within the right to choose, secured by the Constitution, is the right of qualified voters within a state to cast their ballots and have them counted . . . This Court has consistently held that this is a right secured by the Constitution." Id. (emphasis added). Dilution of the right to vote through various techniques, including racial gerrymandering, Gomillion v. Lightfoot, 364 U.S. 339, 345, 81 S. Ct. 125, 5 L. Ed. 2d 110 (1960), and conducting white primaries, Terry v. Adams, 345 U.S. 461, 469, 73 S. Ct. 809, 97 L. Ed. 1152 (1953), likewise violate the Constitution due to the effect of denying some citizens the right to vote. What is clear from all of the Supreme Court's voting rights cases is that "[t]he right to vote freely for the candidate of one's choice is of the essence of a democratic society, and any restrictions on that right strike at the heart of representative government." Reynolds, 377 U.S. at 555, 84 S. Ct. 1362. The Supreme Court's cases have also made another principle unmistakable: "[T]he right of suffrage can be denied by a debasement or dilution of the weight of a citizen's vote just as effectively as by wholly prohibiting the free exercise of the franchise." Id.

Nearly a year before Reynolds, in Gray v. Sanders, 372 U.S. 368, 381, 83 S. Ct. 801, 9 L. Ed. 2d 821 (1963), the Supreme Court held unconstitutional Georgia's county unit system in statewide primary elections. The Court found the system to be unconstitutional because it diluted the weight of votes cast by certain Georgia residents based on where they lived. Id. at 379-80, 83 S. Ct. 801. The Court stated:

Once the geographical unit for which a representative is to be chosen is designated, all who participate in the election are to have an equal vote — whatever their race, whatever their sex, whatever their occupation, whatever their income, and wherever their home may be in that geographical unit. This is required by the Equal Protection Clause of the Fourteenth Amendment. The concept of `we the people' under the Constitution visualizes no preferred class of voters but equality among those who meet the basic qualifications. The idea that every voter is equal to every other voter in his State, when he casts his ballot in favor of one of several competing candidates, underlies many of our decisions.

Id. at 379-80, 83 S. Ct. 801. The Court continued and concluded by stating that: "The conception of political equality from the Declaration of Independence, to Lincoln's Gettysburg Address, to the Fifteenth, Seventeenth, and Nineteenth Amendments can mean only one thing — one person, one vote." Id. at 381, 83 S. Ct. 801.

Also, in Wesberry v. Sanders, 376 U.S. 1, 84 S. Ct. 526, 11 L. Ed. 2d 481 (1964), the Court addressed the constitutional test for determining the validity of congressional districting schemes and held that Equal Protection requires substantial equality of population amongst the districts established by state legislatures for the election of members to the United States House of Representatives. Id. at 18, 84 S. Ct. 526 ("While it may not be possible to draw congressional districts with mathematical precision, that is no excuse for ignoring our Constitution's plain objective of making equal representation for equal numbers of people the fundamental goal for the House of Representatives. That is the high standard of justice and common sense which the Founders set for us."). The high standard establishes that: "No right is more precious in a free country than that of having a voice in the election of those who make the laws under which, as good citizens, we must live. Other rights, even the most basic, are illusory if the right to vote is undermined." Id. at 17, 84 S. Ct. 526.

The Reynolds Court also favorably cited Justice Douglas's dissent in South v. Peters, 339 U.S. 276, 70 S. Ct. 641, 94 L. Ed. 834 (1950), where he stated that:

There is more to the right to vote than the right to mark a piece of paper and drop it in a box or the right to pull a lever in a voting booth. The right to vote includes the right to have the ballot counted. It also includes the right to have the vote counted at full value without dilution or discount.

Id. at 279, 70 S. Ct. 641 (citations omitted); see also Reynolds, 377 U.S. at 562, 84 S. Ct. 1362 ("And, if a State should provide that the votes of citizens in one part of the State should be given two times, or five times, or 10 times the weight of votes of citizens in another part of the State, it could hardly be contended that the right to vote of those residing in the disfavored areas had not been effectively diluted."). Vote dilution, of course, while just as effective as an outright denial of the franchise, may be accomplished in many ways, both intentionally and unintentionally, in a manner that does not immediately shock the senses as would an outright denial. This is not a reason, however, to be any less cognizant of the Equal Protection Clause implications. "One must be ever aware that the Constitution forbids sophisticated as well as simpleminded modes of discrimination." Reynolds, 377 U.S. at 563, 84 S. Ct. 1362 (internal quotation marks omitted).

Helpful to the analysis in Reynolds and favorably cited by the Court was Justice Black's dissent in Colegrove v. Green, 328 U.S. 549, 66 S. Ct. 1198, 90 L. Ed. 1432 (1946), where he stated that:

No one would deny that the equal protection clause would also prohibit a law that would expressly give certain citizens a half-vote and others a full vote. . . [T]he Constitutionally guaranteed right to vote and the right to have one's vote counted clearly imply the policy that state election systems, no matter what their form, should be designed to give approximately equal weight to each vote cast. . . . Thus, a state legislature cannot deny eligible voters the right to vote for Congressmen and the right to have their vote counted. It can no more destroy the effectiveness of their vote in part and no more accomplish this in the name of `apportionment' than under any other name.

Id. at 569-70, 66 S. Ct. 1198 (emphasis added).

Additional cases are relevant to our decision. Harper v. Virginia State Bd. of Elections held invalid a Virginia poll tax and declared that "a State violates the Equal Protection Clause of the Fourteenth Amendment whenever it makes the affluence of the voter of payment of any fee an electoral standard." 383 U.S. 663, 666, 86 S. Ct. 1079, 16 L. Ed. 2d 169 (1966). To reach this conclusion, the Court noted that: "We have long been mindful that where fundamental rights and liberties are asserted under the Equal Protection Clause, classifications which might invade or restrain them must be closely scrutinized and carefully confined." Id. at 670, 86 S. Ct. 1079.

Next, in Dunn v. Blumstein, 405 U.S. 330, 92 S. Ct. 995, 31 L. Ed. 2d 274 (1972), the Court struck down Tennessee's durational residence statute. Applying strict scrutiny, the Court stated:

There is no need to repeat now the labors undertaken in earlier cases to analyze this right to vote and to explain in detail the judicial role in reviewing state statutes that selectively distribute the franchise. In decision after decision, this Court has made clear that a citizen has a constitutionally protected right to participate in elections on an equal basis with other citizens in the jurisdiction.

Id. at 336, 92 S. Ct. 995 (citations omitted). Thus, the law would be "unconstitutional unless the State can demonstrate that such laws are necessary to promote a compelling governmental interest." Id. at 342, 92 S. Ct. 995 (quotation omitted). Stated another way, "a heavy burden of justification is on the State, and [] the statute will be closely scrutinized in light of its asserted purposes." Id. at 343, 92 S. Ct. 995. The Court further stated that laws "affecting constitutional rights must be drawn with precision . . . and must be tailored to serve their legitimate objectives. And if there are other, reasonable ways to achieve those goals with a lesser burden on constitutionally protected activity, a State may not choose the way of greater interference." Id. (internal quotations and citations omitted).

Somewhat more recently decided is Bush v. Gore, 531 U.S. 98, 121 S. Ct. 525, 148 L. Ed. 2d 388 (2000),8 which reiterated long established Equal Protection principles. Id. at 104, 121 S. Ct. 525 ("When the state legislature vests the right to vote for President in its people, the right to vote as the legislature has prescribed is fundamental; and one source of its fundamental nature lies in the equal weight accorded to each vote and the equal dignity owed to each voter."). Echoing long-revered principles, the Court emphasized that states, after granting the right to vote on equal terms, "may not, by later arbitrary and disparate treatment, value one person's vote over that of another." Id. at 104-05, 121 S. Ct. 525 (citing Harper v. Virginia Bd. of Elections, 383 U.S. 663, 665, 86 S. Ct. 1079, 16 L. Ed. 2d 169 (1966)).9 That is, the right to vote encompasses "more than the initial allocation of the franchise. Equal protection applies as well to the manner of its exercise." Id.

Equal Protection, therefore, requires "minimal procedural safeguards" such that there is "at least some assurance that the rudimentary requirements of equal treatment and fundamental fairness are satisfied." Id. at 109, 121 S. Ct. 525.10 Echoing Dunn's holding that "if there are other, reasonable ways to achieve those goals with a lesser burden on constitutionally protected activity, a State may not choose the way of greater interference," 405 U.S. at 343, 92 S. Ct. 995, the Court held that "[t]he formulation of uniform rules to determine intent based on these recurring circumstances is practicable and, we conclude, necessary," Bush, 531 U.S. at 106, 121 S. Ct. 525. Furthermore, the Court cautioned that "[i]t must be remembered that `the right of suffrage can be denied by a debasement or dilution of the weight of a citizen's vote just as effectively as by wholly prohibiting the free exercise of the franchise.'" Id. at 105, 121 S. Ct. 525 (quoting Reynolds, 377 U.S. at 555, 84 S. Ct. 1362). And finally, the Court expressed its reluctance at having to decide the issue, but asserted that "[w]hen contending parties invoke the process of the courts . . . it becomes our unsought responsibility to resolve the federal and constitutional issues the judicial system has been forced to confront." Id. at 111, 121 S. Ct. 525.

Last, contrary to the dissent's argument, the Court's decision in Burdick v. Takushi, 504 U.S. 428, 112 S. Ct. 2059, 119 L. Ed. 2d 245 (1992) does not require a different approach. In Burdick, the Court considered whether Hawaii's prohibition on write-in voting violated the First and Fourteenth Amendments. Id. at 430, 112 S. Ct. 2059. The Court began by noting that any election law will in some way "impose some burden upon individual voters," and, as a result, applying strict scrutiny to any claim tangentially related to the right to vote would unnecessarily tie the hands of state officials seeking to regulate orderly elections. Id. at 433, 112 S. Ct. 2059. The Court did reassert that the right to vote is fundamental, but noted that from this principle it does not follow that the "the right to associate [with a particular candidate] for political purposes through the ballot [is] absolute." Id. In Burdick, therefore, the Court found the burden to be minimal and affecting only those voters failing to comply with Hawaii's reasonable and non-discriminatory election procedures for qualifying for a place on the ballot. Id. at 434-45, 112 S. Ct. 2059. Thus, unlike the instant case, where the technological burden is not within the control of the voters, any burden in Burdick was the result of the voter's failure to timely identify a candidate and seek to place such person on the ballot. Id. at 436-37, 112 S. Ct. 2059 ("Consequently, any burden on voters' freedom of choice and association is borne only by those who fail to identify their candidate of choice until days before the primary. But in Storer v. Brown, [415 U.S. 724, 736, 94 S. Ct. 1274, 39 L. Ed. 2d 714 (1974)] we gave little weight to `the interest the candidate and his supports may have in making a late rather than an early decision to seek independent ballot status.'"). Finally, noting that compliance with Hawaii's procedures afforded "adequate ballot access," the State's ban on write-in voting was only a "limited burden on voters' rights to make free choices and to associate politically through the vote." Id. at 438-39, 112 S. Ct. 2059. We have little problem dismissing the relevance of Burdick, the context of which is a voter's desire to vote for a write-in candidate after failing to properly comply with the state's procedures for getting that candidate on the ballot in the first place. The voter in Burdick had access to the ballot on equal terms with the rest of the electorate. The Supreme Court viewed the burden in Burdick as little more than a minor burden on those who waited until "the eleventh hour," id. at 439, 112 S. Ct. 2059, to decide whom to vote for or the interest in recording a "protest vote," id. at 438, 112 S. Ct. 2059. In this case, however, through no failure on their parts, Ohio voters facing deficient technology approach the ballot in a position unequal from the portion of the electorate using adequate technology. As the State's HAVA report stated: "The evidence is overwhelming that thousands of Ohio voters have been disenfranchised by antiquated voting equipment." The burden on the franchise in Burdick pales in comparison to the burden in this case.

The dissent's reliance on Burdick, therefore, is misplaced. First, "[i]n Burdick, the plaintiffs' First Amendment rights to ballot access were at issue," Southwest Voter, 344 F.3d at 899, not the plaintiffs' right to vote, which is at issue in this case. Second, the dissent's claim that "[f]ederal courts applying the Burdick framework have evaluated challenges to various state voting regulations under a rational-basis standard of review" Dis. Op. at 35, is correct, but misses the point. Burdick was about a candidate's access to the ballot; it was not a case that addressed a voter being denied an equal chance to have her vote counted. Each of the cases cited by the dissent, employing Burdick's framework, are likewise cases having little to nothing to do with the exercise of the franchise.11 Moreover, the dissent adopts an overly narrow interpretation of the "right to vote" and ignores the Supreme Court's statement in Bush that the right to vote encompasses "more than the initial allocation of the franchise. Equal protection applies as well to the manner of its exercise." Bush, 531 U.S. at 104, 121 S. Ct. 525 (emphasis added).12 We also note that the case before us is more similar to Bush where the Court did not find it necessary to address or cite Burdick when applying a heightened standard of review.

Zealous protection of the right to vote has been the norm. Few rights have been so extensively and vigorously protected as the right to vote. Its fundamental nature and the vigilance in its defense, both from the courts, Congress, and through the constitutional amendment process, stem from the recognition that our democratic structure and the preservation of other rights depends to a great extent on the franchise. To that effect, from Yick Wo to Reynolds to Bush, Supreme Court precedent is clear: we must apply strict scrutiny to the challenged voting practice at issue here.

Ending this section on the proper standard of review, we now briefly review other lower courts treatment of similar voting rights challenges in recent years.13

B. Similar Equal Protection Challenges to Voting Technology

Relatively soon after the 2000 presidential election, the first case challenging the use of certain voting technology as violative of the Equal Protection Clause was heard in the Central District of California. The plaintiffs alleged that the use of punch card voting systems amounted to vote denial in violation of the Fourteenth Amendment as well as a claim under the Voting Rights Act. See Common Cause Southern Christian Leadership Conference of Greater Los Angeles v. Jones, 213 F. Supp. 2d 1106, 1107-08 (C.D.Cal.2001). The Secretary of State moved for judgment on the pleadings and the district court denied the motion. Id. at 1108.

Like Ohio's Secretary of State, California's Secretary of State "has the power to publish a list of voting systems from which counties may choose." Id. at 1107. This list included punch card systems which the plaintiffs alleged were less reliable than other listed systems, and therefore those living in counties where the punch card system was used were substantially less likely to have their votes counted in violation of the Fourteenth Amendment. Id. The plaintiffs also alleged that the counties using punch card systems "have high racial minority populations in comparison with counties using other voting systems" and therefore alleged a denial of the right to vote on the basis of race under Section 2 of the Voting Rights Act. Id. at 1107-08.

The district court concluded that the plaintiffs alleged sufficient facts to survive a motion for judgment on the pleadings under both strict scrutiny and rational basis review. Id. at 1109 ("Even if the more lenient standard is ultimately applied by this Court, Plaintiff has alleged facts indicating that the Secretary of State's permission to counties to adopt either punchcard voting procedures or more reliable voting procedures is unreasonable and discriminatory.").

Regarding the Voting Rights Act claim, the court characterized the plaintiffs' claim as racial minorities being "disproportionately denied the right to vote because their votes are uncounted in disproportionate numbers as a result of the voting mechanism that they are supplied." Id. at 1110. The court again denied the defendants' motion to dismiss. Id. Following the decision, the parties agreed to and the court entered a consent decree requiring that for the nine counties in California using punch card technology convert to "other certified voting equipment" by March 2004. See Common Cause v. Jones, No. 01-03470 SVW (RZX), 2002 WL 1766436 (C.D.Cal. Feb.19, 2002), enforced by, No. 01-03470 SVW (RZ), 2002 WL 1766410 (C.D.Cal. May 9, 2002).

In 2002, in a more comprehensive fashion, the district court for the Northern District of Illinois addressed a similar claim. See Black v. McGuffage, 209 F. Supp. 2d 889 (N.D.Ill.2002). The court began by noting that many election problems have come to light in recent years. Id. at 891. The plaintiffs in Black made essentially the same claims now before this Court — that is, claims under the Equal Protection and Due Process Clauses, as well as a claim for the minority plaintiffs under Section 2 of the Voting Rights Act. Id. at 892 ("Plaintiffs allege that African American and Latino voters are disproportionately forced to use — and are disproportionately injured when they use — the challenged voting systems.").

Like Ohio and California (and most other states), the counties in Illinois may choose from several voting systems certified by the Illinois State Board of Elections. Id. Illinois jurisdictions could choose from four technologies — precinct-count optical scan (with error notification), central-count optical scan (without error notification), precinct-count punch card (with error notification) and central-count punch card (without error notification) technologies. Id. Of the 110 local election, boards in Illinois, in the 2000 election, ninety-five boards used central-count punch card systems. Id. Various residual vote rates were listed by the district court:

The average residual vote rate across Illinois, for ballots cast for president in the November, 2000 presidential election, was approximately 3.85%. This rate varied substantially among Illinois jurisdictions. [In][j]urisdictions using optical scan ballots with error notification the average residual vote rate was less than 1%. The rates ranged from 0.32% in McHenry County to 1.07% in Franklin County. In jurisdictions using punch card ballots without error notification the average residual vote rate was more than 4%. For example, in Chicago, the rate was 7.06% citywide, 12.59% in the 12th ward, 12.4% in the 37th ward, and 36.73% in the 48th precinct of the 29th ward. Similarly, the rate was 5.23% in suburban Cook County, 8.8% in the Cicero Township of Cook County and 7.48% in Alexander County. The rate was 3.17% in Whiteside County, 2.48% in Will County, and 2.15% in Sangamon County. In the three election jurisdictions which use optical scan ballots without error notification the average residual vote rate was more than 4%. For example, the rate was 10.88% in all of East St. Louis, but 22.30% in the 20th precinct of East St. Louis.

Id. at 893. As in the instant case, based on various residual vote rates, the plaintiffs argued that "individuals living in punch card jurisdictions have a greater statistical probability of not having their votes counted . . . [and] that the counties with the punch card system have larger populations of minorities than do counties using other voting systems, and thus use of those less accurate machines has a disparate impact on minority voters." Id. at 894.

The district court addressed the claim in the context of a 12(b)(6) motion to dismiss. The court first rejected the defendants' claims based on standing and mootness, id. at 894-96, and also rejected the defendants' 12(b)(6) claim as to the merits.

The court first addressed the plaintiffs' claim under Section 2 of the Voting Rights Act of 1965. The court stated that "the Supreme Court has emphasized [that the Voting Rights Act] `should be interpreted in a manner that provides the broadest possible scope in combating racial discrimination.'" Id. at 896 (quoting Chisom v. Roemer, 501 U.S. 380, 403, 111 S. Ct. 2354, 115 L. Ed. 2d 348 (1991) (internal quotations omitted)). Section 2(a) of the Act prohibits the use of any electoral practice of procedure that "results in a denial or abridgement of the right of any citizen of the United States to vote on account of race or color." 42 U.S.C. § 1973(a). Section 2 broadly protects the franchise and includes "all action necessary to make a vote effective" including "casting a ballot and having such ballot counted properly." 42 U.S.C. § 1973l(c)(1). Section 2(b) of the Act provides that an electoral practice or procedure results in a violation of the franchise under section (a)

if, based on the totality of circumstances, it is shown that the political processes leading to nomination or election in the State or political subdivision are not equally open to participation by members of a class of citizens protected by subsection (a) of this section in that its members have less opportunity than other members of the electorate to participate in the political process and to elect representatives of their choice.

42 U.S.C. § 1973(b).

Based on these statutes, the district court in Black found that "[t]here are two statutory elements to a claim under Section 2:(1) the use of an electoral `standard, practice or procedure,' and (2) a resulting diminution of the opportunity to [minority] voters `to participate in the political process and to elect representatives of their choice.'" Id. at 896 (quoting 42 U.S.C. § 1973(a)-(b)). The court then concluded that the plaintiffs' complaint met this requirement by identifying an electoral practice ("deficient ballot system"), and alleged that because of the deficient system, the minority plaintiffs have less opportunity than other members of the electorate to participate in the process and elect representatives of their choice. Id. at 897. Finding that the facts alleged could demonstrate that the plaintiffs "participation in the political process could be significantly diminished," the district court denied the defendants' motion to dismiss. Id.

The court then turned to the plaintiffs' claim under the Equal Protection Clause. The Equal Protection claim in Black, and in the instant case, is not that a classification has been drawn based on race, but rather that "the counties' use of different voting systems in different parties of the state violates the Equal Protection Clause" by valuing one person's vote more than another's. Id. at 897. Because there was no explicit classification, the defendants argued that rational basis was the standard of review while the plaintiffs argued that strict scrutiny is the appropriate standard because the voting system affects the plaintiffs' fundamental right to vote. Id. The district court did not directly decide the issue and reserved its determination for trial. Id.

The court then turned to Supreme Court precedent and found the plaintiffs' allegations fitting squarely within the Court's Equal Protection admonitions. The district court did find the plaintiffs' claims to be more analogous to Bush v. Gore than any other of the Supreme Court's previous voting rights cases, but the principle behind each of them remained the same. Id. at 898-99. According to the court, as a result of the state's arbitrary and disparate treatment of voters in different counties, "voters in some counties are statistically less likely to have their votes counted than voters in other counties in the same state in the same election for the same office. Similarly situated persons are treated differently in an arbitrary manner." Id. at 899. Thus, because of the voting system used, the state "`value[s] one person's vote over that of another,' . . . [and] does not afford the `equal dignity owed to each voter.'" Id. (quoting Bush v. Gore, 531 U.S. at 104-05, 121 S. Ct. 525).

The third applicable case involved the California gubernatorial recall election. Following the disposition of Common Cause, but prior to the gubernatorial recall election in California in 2003, various plaintiffs filed suit in the same district court as Common Cause, seeking an injunction that would prevent the use of punch card technology in the recall election. See Southwest Voter Registration Education Project v. Shelley, 278 F. Supp. 2d 1131 (C.D.Cal. 2003). The district court denied the request for the injunction, finding that res judicata and laches were viable defenses based on the Common Cause consent decree, that the plaintiffs were unlikely to succeed on the merits of their claims based on the short time until the election, and that the balance of hardships and the public interest weighed in favor of allowing the election to go forward as scheduled. Id. In a fast moving process, the district court issued its ruling on August 20, 2003, and the case was argued before the Ninth Circuit on September 11. The court issued its opinion on September 15 and reversed the district court. Southwest Voter Registration Education Project v. Shelley, 344 F.3d 882 (9th Cir.2003), vacated by 344 F.3d 913, rev'd by 344 F.3d 914 (en banc). The panel stayed its decision and the court decided to hear the case en banc on September 22. The en banc court reversed the panel decision on the following day, which allowed the election to proceed days later.

The panel had concluded that the plaintiffs' "Equal Protection Clause claim mirrors the one recently analyzed by the Supreme Court in Bush v. Gore." Southwest Voter, 344 F.3d at 894. The panel cited Reynolds in which the Supreme Court held that "[i]t has been repeatedly recognized that all qualified voters have a constitutionally protected right to vote, and to have their votes counted." 377 U.S. at 554, 84 S. Ct. 1362 (citations omitted). Citing precedent, the panel wrote that the case before it presented "a classic voting rights equal protection claim." Southwest Voter, 344 F.3d at 895. The panel cited Gray v. Sanders, where the Court held that: "Every voter's vote is entitled to be counted once. It must be correctly counted and reported." 372 U.S. 368, 380, 83 S. Ct. 801, 9 L. Ed. 2d 821 (1963). Also discussed was Reynolds, where the Court held that "the weight of a citizen's vote cannot be made to depend on where he lives." Reynolds, 377 U.S. at 567, 84 S. Ct. 1362.

According to the panel, all that the Constitution requires is some assurance that the rudimentary requirements of equal treatment and fundamental fairness are satisfied. Id. at 896, 84 S. Ct. 1362 (citations omitted). The panel also concluded that strict scrutiny is ordinarily the appropriate standard in a situation akin to the one presented because the use of deficient technology affects the fundamental right to have one's vote counted. Id. at 899, 84 S. Ct. 1362. Nevertheless, the court found that it need not reach the question because the plaintiffs demonstrated a likelihood of success on the merits regardless of the standard of review employed. Id. at 900. "Plaintiffs have tendered sufficient evidence to demonstrate a likelihood of success in establishing that there is no rational basis for using voting systems that have been decertified as `unacceptable' in some counties and not others." Id.

The case was then heard en banc and the panel decision was reversed. Southwest Voter Registration Education Project v. Shelley, 344 F.3d 914 (9th Cir.2003) (en banc). The court, in a per curiam opinion with no dissents, concluded that the district court did not abuse its discretion in denying the preliminary injunction. Id. at 918. The court, however, did not reject outright the Equal Protection argument, noting its similarity to the claims advanced in Bush v. Gore, nor did it reject outright the plaintiffs' Voting Rights Act claim.

The court repeatedly emphasized that abuse of discretion is the standard for reviewing a district court's denial of a preliminary injunction. Shelley, 344 F.3d at 917 ("The [abuse of discretion] standard of review is important to our resolution of this case."); id. at 918 ("We review the district court's decision to grant or deny a preliminary injunction for abuse of discretion. Our review is limited and deferential.") (internal citation omitted). The court then stated that: "We are met with legal authority on both sides of the contest. There is no doubt that the right to vote is fundamental, but a federal court cannot lightly interfere with or enjoin a state election." Id. at 918 (citation omitted). In a short one paragraph discussion of the Equal Protection claim, the en banc court stated: "We have not previously had occasion to consider the precise equal protection claim raised here. That a panel of this court unanimously concluded the claim had merit provides evidence that the argument is one over which reasonable jurists may differ." Id. Nevertheless, with great heed paid to the principle that a federal court should not lightly enjoin a state election, the en banc court held that the "district court did not abuse its discretion in holding that the plaintiffs have not established a clear probability of success on the merits of their equal protection claim." Id. (emphasis added).

The court then turned to the plaintiffs' Voting Rights Act claim, which it referred to as "stronger." Id. Nevertheless, noting a "significant dispute in the record, however, as to the degree and significance of the [residual vote] disparity," the court concluded that "although plaintiffs have shown a possibility of success on the merits, we cannot say that at this stage they have shown a strong likelihood." Id. at 918-19.14

A review of the lower court cases demonstrates consistent application of the principles established by the Supreme Court. Each of the cases discussed above has recognized the equal protection and Voting Rights Act implications of the use of non-notice and disparate technology and we find much of their reasoning persuasive. We now turn to this Court's treatment of right to vote cases.

This Circuit has yet to review a claim of this precise ilk. The district court below and the defendants point to Mixon v. Ohio, 193 F.3d 389 (6th Cir.1999), as establishing the relevant framework for this Court's analysis. Mixon addressed the constitutionality of a state allowing the mayor to appoint a new school board for the Cleveland School District. This Court cited Supreme Court precedent acknowledging that voting is a fundamental right, and then discussed the differing standards of review applicable to equal protection claims. Id. at 402. According to this Court, "[i]f the challenged legislation grants the right to vote to some residents while denying the vote to others, then we must subject the legislation to strict scrutiny and determine whether the exclusions are necessary to promote a compelling state interest." Id. (citing Dunn v. Blumstein, 405 U.S. 330, 337, 92 S. Ct. 995, 31 L. Ed. 2d 274 (1972)). "If the legislation, however, does not infringe on the right to vote [and instead simply regulates a tangential aspect of the franchise], we examine the challenged statute under the rational basis standard." Id.; see also Craigmiles v. Giles, 312 F.3d 220, 223 (6th Cir.2002) ("When a statute regulates certain `fundamental rights' (e.g. voting or abortion) or distinguishes between people on the basis of certain `suspect characteristics' (e.g. race or national origin), the statute is subject to `strict scrutiny.'"). The Mixon court proceeded to analyze the claim presented under rational basis because the statute did not infringe upon any right to vote — that is, "[a]lthough Plaintiffs have a fundamental right to vote in elections before them, there is no fundamental right to elect an administrative body such as a school board, even if other cities in the state may do so." Mixon, 193 F.3d at 403 (citations omitted).

"On appeal from a judgment entered following a bench trial, we review the district court's factual findings for clear error and its legal conclusions de novo." Pressman v. Franklin Nat. Bank, 384 F.3d 182, 185 (6th Cir.2004).

Supreme Court precedent and our own Mixon framework instructs that if the Ohio statute permitting localities to use deficient voting technology "infringe[s] on the right to vote," then strict scrutiny applies; if the statute does not "infringe on the right to vote," and merely regulates some tangential aspect of the franchise, then rational basis review applies. Mixon, 193 F.3d at 402. This begs the question of what the "right to vote" encompasses. We easily conclude that the right to have one's vote counted on equal terms is part of the right to vote. No other conclusion is possible from the case law and thus, strict scrutiny applies.15

All of the precedent indicates that having one's vote properly counted is fundamental to the franchise. See e.g., Bush, 531 U.S. at 104-05, 121 S. Ct. 525 (indicating that having one's vote counted on equal terms with others in the relevant jurisdiction is the quintessential "right to vote" case); Reynolds, 377 U.S. at 555, 84 S. Ct. 1362 ("[T]he right of suffrage can be denied by a debasement or dilution of the weight of a citizen's vote just as effectively as by wholly prohibiting the free exercise of the franchise."); South, 339 U.S. at 279, 70 S. Ct. 641 (Douglas, J., dissenting) ("There is more to the right to vote than the right to mark a piece of paper and drop it in a box or the right to pull a lever in a voting booth. The right includes the right to have the ballot counted. It also includes the right to have the vote counted at full value without dilution or discount.") (citations omitted); Colegrove, 328 U.S. at 569-71, 66 S. Ct. 1198 (Black, J., dissenting) ("The Constitutionally guaranteed right to vote and the right to have one's vote counted clearly imply the policy that state election systems, no matter what their form, should be designed to give approximately equal weight to each vote cast."); Classic, 313 U.S. at 315, 61 S. Ct. 1031 ("Obviously included within the right to choose, secured by the Constitution, is the right of qualified voters within a state to cast their ballots and have them counted."); 42 U.S.C.1973l(c)(1) (defining the right to vote to include "having such ballot counted properly and included in the appropriate total of votes cast"). Case law, statutory definitions, and common sense indicate that the "right to vote" is infringed in this case by the use of the two deficient technologies challenged by the plaintiffs. The distinction in Mixon between elections where persons are entitled to vote (strict scrutiny) and a statute authorizing appointment of administrators instead of having an election (rational basis) further confirms our reasoning here.

Strict scrutiny requires us to determine whether the use of the two challenged technologies in some jurisdictions but not others is a practice "narrowly tailored to further compelling governmental interests." Grutter v. Bollinger, 539 U.S. 306, 326, 123 S. Ct. 2325, 156 L. Ed. 2d 304 (2003). Under this standard, the State's proffered justifications of cost and training are wholly insufficient to sustain its continued certification of the technologies. Administrative convenience is simply not a compelling justification in light of the fundamental nature of the right. See Frontiero v. Richardson, 411 U.S. 677, 690, 93 S. Ct. 1764, 36 L. Ed. 2d 583 (1973) (plurality) ("[W]hen we enter the realm of `strict judicial scrutiny' there can be no doubt that `administrative convenience' is no shibboleth, the mere recitation of which dictates constitutionality."). Moreover, Ohio's reliance on the cost of upgrading the technology fails in light of the monies already devoted to the process. See Shapiro v. Thompson, 394 U.S. 618, 633, 89 S. Ct. 1322, 22 L. Ed. 2d 600 (1969) ("[C]osts cannot justify an otherwise invidious classification."); see also Dunn, 405 U.S. at 343, 92 S. Ct. 995 ("And if there are other, reasonable ways to achieve those goals with a lesser burden on constitutionally protected activity, a State may not choose the way of greater interference."). Additionally, Ohio has used cost as if it were a silver bullet. Any change from the status quo necessarily involves some cost. The State has failed to put forth any evidence indicating that it cannot manage the costs and instead, the evidence indicates that the State has either budgeted for the transition from its own funds or through funds provided by the federal government. The mere fact that there is some cost involved does not make that factor compelling. Further, the State's alleged concern with voter fraud, likewise, in this case, is not compelling in light of the Secretary of State's report concluding that the technology can securely be implemented. See also HB 262, 125 Gen. Assem., Reg. Sess. (Ohio 2004). Finding no compelling reason supporting the State's continued certification of the deficient non-notice technology, i.e., punch-card system and central-count optical scan systems,16 and in light of the fundamental nature of the right to vote, the State's actions violate the Equal Protection Clause. The continued certification of this technology by the Secretary of State does not provide the minimal adequate procedural safeguards to prevent the unconstitutional dilution of votes based on where a voter resides. See Bush, 531 U.S. at 109, 121 S. Ct. 525.17 Unequal treatment and unfairness are perpetrated by allowing this technology to remain certified and the State's reasons for maintaining this disparate system are far from compelling.

Our decision is consistent with binding Supreme Court precedent. In Harper, the Supreme Court held that using wealth or ability to pay as a factor in the power of the franchise "is to introduce a capricious or irrelevant factor." Harper, 383 U.S. at 668, 86 S. Ct. 1079. Likewise, the maintenance of disparate technologies — that the State itself has recognized is inherently flawed and that "[t]he evidence is overwhelming that thousands of Ohio voters have been disenfranchised by antiquated voting equipment" — debases and devalues citizens' votes and is arbitrary and capricious. The technology provided to a voter by the State, like that voter's wealth, has no relation to voting qualifications or the value of that vote. "[T]he right to vote is too precious, too fundamental to be so burdened or conditioned." Harper, 383 U.S. at 670, 86 S. Ct. 1079; see also Hasen, Bush v. Gore and the Future of Equal Protection Law in Elections, 29 Fla. St. U.L.Rev. at 395 (concluding that the disparate treatment that results from "different voting systems with different error rates. . . is all the more disturbing to the extent that it correlates with wealth, looking functionally like the poll tax the Court struck down in Harper").

Overweighing and overvaluation of the votes of those living [in a county with adequate technology] has the certain effect of dilution and undervaluation of the voters of those living [in a county with deficient technology]. The resulting discrimination against those individual voters living in disfavored areas is easily demonstrable mathematically. Their right to vote is simply not the same right to vote as that of those living in a favored part of the State . . . Weighting the votes of citizens differently, by any method or means, merely because of where they happen to reside, hardly seems justifiable.

377 U.S. at 563, 84 S. Ct. 1362 (emphasis added). By maintaining a system in which these two technologies are utilized, voters in Ohio vote under two separate standards. Although voters approach the polls with the opportunity to vote in the same elections for the same candidates, once they step into the voting booth, they have an unequal chance of their vote being counted, not as a result of any action on the part of the voter, but because of the different technology utilized. Voters able to utilize notice technology choose candidates and before their vote is turned in and counted, the technology notifies them of any errors that would result in the vote being disregarded. Those voters forced to use non-notice technology are not notified of any errors in their ballot and, should errors exist, their votes are disregarded; moreover, voters using the two challenged technologies have an additional likelihood of disenfranchisement due to the inherent deficiencies of the punch-card and central-count optical scan.18

The plaintiffs have alleged vote dilution due to disparate use of certain voting technologies. They have demonstrated their allegations mathematically and from the testimony of witnesses on both sides of the contests. The State's own documents and public statements support the plaintiffs' claims. The State of Ohio concedes that punch-card technology is substandard. See, e.g., Secretary of State's Help America Vote Act Report ("Boards of election should upgrade their voting systems to new, more trustworthy technology."); id. ("These goals demand immediate attention, or our state runs the risk of repeating the problems of our nation's most recent presidential election — and suffering irreparable damage to the most important and basic concepts of democracy."); id. ("The evidence is overwhelming that thousands of Ohio voters have been disenfranchised by antiquated voting equipment. . . ."). The use of the punch card and central-count optical scan technologies result in a greater likelihood that one's vote will not be counted on the same terms as the vote of someone in a notice county.19 The State's continued use of this technology results in the weighting of votes differently and this is hardly justifiable. Technology, as a method or means to dilute voting strength, is no less a violation than any other invidious method. According to the data presented below, and allowing for 0.75% of intentional residual votes, approximately 55,000 presidential votes were lost in 2000 by those using punch card technology.20 This is also a conservative estimate. Ten counties using punch cards had residual vote rates in excess of 3% in the 2000 election.

Although we apply strict scrutiny, we note that the use of this technology would also fail under rational basis review.21 This is true under rational basis review both searching and deferential. See Lawrence v. Texas, 539 U.S. 558, 579, 123 S. Ct. 2472, 156 L. Ed. 2d 508 (2003) (O'Connor, J., concurring in judgment). An individual's vote is the lifeblood of a democracy. To that extent, we find it difficult to conjure up what the State's legitimate interest is by the use of technology that dilutes the right to vote. The State says that its interests in failing to decertify non-notice and substandard technology is based on the cost of replacement and training workers on the new machines. We fail to see how the interest is justified or rational at the expense of tens of thousands of votes.

Under FCC v. Beach Communications, Inc., 508 U.S. 307, 313, 113 S. Ct. 2096, 124 L. Ed. 2d 211 (1993), legislation "must be upheld against equal protection challenge if there is any reasonably conceivable state of facts that could provide a rational basis" for the challenged action. This does not mean, however, that state action survives if the State's proffered justifications are belied by the actual facts of the case or are simply arbitrary justifications. Such is the case here; no facts provide a rational basis for Ohio's continued certification of the deficient technologies. Governments almost always attempt to justify their conduct based on cost and administrative convenience, but the State's reliance on these factors is not necessarily rational, and certainly is not in this case. The State has failed to put forth any evidence indicating that it cannot manage the costs and instead, the evidence indicates that the State has either budgeted for the transition from its own funds or through funds provided by the federal government The loss of so many votes because of the continued use of machines that the State admits are substandard is arbitrary and cannot be considered rational in light of so-claimed, but unsupported, cost concerns. Moreover, the experience of the counties in Ohio who have stopped using the non-notice substandard technology weigh heavily against the State's argument. None of these counties have encountered significant technological difficulties or undue financial burdens. The Secretary of State's report and Ohio legislation concludes that a switch to electronic machines can be securely implemented and there have been no instances of voter fraud in Ohio with the use of electronic voting machines. Moreover, the evidence demonstrates that more accurate technologies are readily available. And finally, while the State's intentions to comply with HAVA are insufficient to moot the case, they do cast considerable doubt on its claims that it is not economically feasible for it to decertify the technology. In fact, the State's HAVA report sufficiently belies each of the State's proffered justifications for failing to decertify the substandard technology. Thus, while we hold that strict scrutiny is the appropriate standard of review, we also determine that the State's failure to decertify the use of non-notice and technologically unsound punch card and central count optical scan machines would also fail rational basis review. Decertification of the technology "is practicable and, we conclude, necessary." Bush, 531 U.S. at 106, 121 S. Ct. 525.

Now, a few words in response to the dissent. To start, the dissent misunderstands the nature of the plaintiffs' claim. According to the dissent, no equal protection problem arises because "properly marked ballots are counted and accorded the same value everywhere in the state." Dis. Op. at 34. The dissent attributes any mistakes to the voters themselves and concludes that the only problem is that "voters in some counties have a reduced chance of turning in a properly marked ballot that will actually be counted because certain voting methods do not catch the voters' own inadvertent mistakes." Id. This fundamentally mischaracterizes the problem with the substandard voting technology. The evidence cited above demonstrates that the problem with the punch card ballot in its design is that a vast number of properly marked ballots are not counted due to machine error. Thus, the issue is not as the dissent states, but rather, whether a state "may allow the use of different types of voting equipment with substantially different levels of accuracy." See Black v. McGuffage, 209 F. Supp. 2d 889, 898 (N.D.Ill.2002) (emphasis added).

Next, the main basis for the dissent is some "question [over] the precedential value of Bush v. Gore." Dis. Op. at 881. The dissent answers the question by concluding that Bush v. Gore is "a case whose precedential value . . . is at best questionable." Dis. Op. at 886.22 The dissent endeavors to explain three "reasons for doubting Bush v. Gore's precedential value: `[T]he limiting language in the opinion, the lack of seriousness with which the Court undertook its own analysis, and the inconsistency with other jurisprudence by this majority of Justices23 all point in the direction of assuming that Bush v. Gore is not good precedent for an expansive reading of equal protection law in elections.'" Dis. Op. at 887 (quoting Hasen, Bush v. Gore and the Future of Equal Protection Law in Elections, 29 Fla. St. U.L.Rev. at 391). The dissent then focuses on the first two reasons: the lack of the apparent seriousness with which the Court decided the case and the inconsistency with other jurisprudence. It is curious that the dissent bases its analysis on the belief that the Supreme Court decided Bush v. Gore, a case that decided the 2000 presidential election, with a "lack of seriousness," and we choose to disagree with that approach. Further, even if the Court was playing fast and loose with the law, we, as an inferior court, are not in a position to disregard Supreme Court precedent because we think they got it wrong. Presumably, if in the dissent's view Bush v. Gore had "precedential value" — as we think all Supreme Court opinions do — he would be compelled to join us. In fact, early this Term, the Supreme Court discussed the proper approach when a lower court finds itself bound by precedent with which it disagrees. See Eberhart v. United States, ___ U.S. ___, ___, 126 S. Ct. 403, 407, 163 L. Ed. 2d 14 (2005) (per curiam) ("We finally add a word about the approach taken by the Court of Appeals. Although we find its disposition to have been in error, we fully appreciate that it is an error shared among the circuits, and that it was caused in large part by imprecision in our prior cases. Our repetition of the phrase "mandatory and jurisdictional" has understandably led the lower courts to err on the side of caution by giving the limitations in Rules 33 and 45 the force of subject-matter jurisdiction. Convinced, therefore, that Robinson and Smith governed this case, the Seventh Circuit felt bound to apply them, even though it expressed grave doubts in light of Kontrick. This was a prudent course. It neither forced the issue by upsetting what the Court of Appeals took to be our settled precedents, nor buried the issue by proceeding in a summary fashion. By adhering to its understanding of precedent, yet plainly expressing its doubts, it facilitated our review.").

Finally, we reject the dissent's claim that Professor Hasen's article has overruled the Supreme Court's decision in Bush v. Gore because "the Supreme Court has had ample opportunity to prove him wrong by explaining, or even citing to, its decision in Bush v. Gore." Dis. Op. at 887. If the dissent were correct, perhaps the same must be said about the hundreds of other law review articles discussing the decision's legitimacy and applicability to other factual scenarios. The Court has not sought to debate these other law professors and commentators and under the dissent's reasoning, those articles must also be correct. In fact, taken to its logical extreme, the dissent's reasoning means that every claim in a law review article that the Supreme Court has failed to refute is actually the law of the land. Thus, we find the dissent's claim that Bush v. Gore has no precedential value because the Court has not sought to "prove [Professor Hasen] wrong" empty.

Moreover, the dissent bases its analysis entirely upon Professor Hasen's suggestion that Bush v. Gore is not serious, but fails to acknowledge the second half of Hasen's article where he concludes that: "In sum, if Bush v. Gore indeed has precedential value, it clearly should apply to prevent the use of these different voting systems in the same election." Hasen, Bush v. Gore, 29 Fla. St. U.L.Rev. at 395; see also id. at 379, 121 S. Ct. 525 (stating that his article "concludes that, if the case were taken seriously, Bush v. Gore should have great precedential value in changing a host of voting procedures and mechanisms, particularly when those procedures and mechanisms are challenged prospectively").24 Thus, the dissent's argument is easily deconstructed. It is premised solely on Professor Hasen's article suggesting that Bush v. Gore should not be given precedential value. Because the dissent concludes that the decision should not be given precedential value (because the Court did not take the case seriously and an apparent inconsistency with other precedent) it does not mention the fact that Professor Hasen goes on to conclude that if Bush v. Gore were followed, it would dictate the result we reach here.25 Unfortunately for the dissent, inferior courts do not have the luxury of suggesting that a Supreme Court decision simply should not be followed without some tenable legal basis. Thus, because the dissent has not endeavored to provide any legitimate basis or principled manner of distinguishing Bush v. Gore — and presumably has not adopted Hasen's argument that "[e]mbarrassment provides the only hope that the case will have precedential value," Hasen, Bush v. Gore, 29 Fla. St. U.L.Rev. at 391 — his argument that we simply should not follow the case does not give us any pause.

In the end, the dissent's reasoning ultimately flounders. The dissent concludes that our decision is "persuasive only to the extent that Bush v. Gore is controlling. Neither [our decision or Shelley I], in my view, successfully refutes the compelling reasons supplied by Professor Hasen for refusing to `take Bush v. Gore's equal protection holding seriously.'" Dis. Op. at 891 (citing Hasen, 29 Fla. St. U.L.Rev. at 380). The dissent, however, fails to mention Professor Hasen's ultimate conclusion that "if Bush v. Gore indeed has precedential value, it clearly should apply to prevent the use of these different voting systems in the same election." Hasen, Bush v. Gore, 29 Fla. St. U.L.Rev. at 395. Without the luxury or the power to decide which Supreme Court decisions we want to follow, we find Professor Hasen's ultimate conclusion, that the reasoning of Bush v. Gore applies here, to be sound.

Finally, the dissent criticizes us for not articulating a precise mathematical formula for determining when voting technology is constitutional and when it is not. Precise mathematical formulas, however, have never been a part of voting rights cases or cases involving strict judicial scrutiny. A judicially imposed mathematical formula for evaluating voting rights cases would be purely arbitrary. We simply cannot say that x% error rate raises constitutional concerns but y% error rate does not. Nor has the Supreme Court done so or required us to do so in any case applying strict scrutiny.26 Rather, when confronted with difficult constitutional questions, a reasoned analytical approach is necessary. Once the plaintiffs establish that strict scrutiny applies, the State must put forth a compelling reason for its infringement of a fundamental right and we, as a reviewing court, must evaluate that reasons against the plaintiffs' claims. This analysis necessary includes consideration of various factors, including the legal, financial, and practical. See e.g., Dunn, 405 U.S. at 343, 92 S. Ct. 995 ("And if there are other, reasonable ways to achieve those goals with a lesser burden on constitutionally protected activity, a State may not choose the way of greater interference.").

We do not review such claims in a vacuum but do so with a focus on the real world justifications and implications of Ohio's decision. Future claims of this sort will be evaluated with these various factors in mind. If, in the dissent's hypothetical example, where one technology yields a 0.1% error rate and another technology has a 0.01% error rate, and a lawsuit is filed similar to this one, the same questions will have to be asked and the context of the challenge studied and evaluated. If switching to the new technology would bankrupt the State, or the technology was in its development phase, or the potential for fraud was high, or for any number of reasons the State did not believe that the technology could properly be implemented, then, even in light of our decision today, the claim may fail. If, however, switching to the new technology provided only minimal inconvenience, then the State's reasons might be found to be less than compelling. In this case, we do not seek to constitutionalize local election procedures, nor do we seek to mandate uniformity and absolute equality in voting procedures, nor do we hold that notice technology is constitutionally mandated; we merely evaluate the right to vote in light of the State's proffered justifications for using the two challenged technologies at issue and hold that the State's reasons are not compelling. Our inquiry here is aided by Ohio's decision to adopt a litigation position that is undermined by its actions and public words. We therefore hold that the two technologies challenged here fail to satisfy "rudimentary requirements of equal treatment and fundamental fairness," Bush, 531 U.S. at 109, 121 S. Ct. 525, and that based on the evidence presented below, the plaintiffs established a violation of the Equal Protection Clause.27

The African-American plaintiffs also advance a claim under Section 2 of the Voting Rights Act against Hamilton, Montgomery, and Summit Counties based on their use of punch card technology. After the bench trial, the district court concluded that the plaintiffs' "Voting Rights Act claim fails because their alleged injury does not amount to a vote denial under § 2 of the Voting Rights Act." This conclusion was based on the court's legal interpretation of the Voting Rights Act that a vote denial claim is cognizable only "when a state or municipality employs a `practice or procedure' that results in the `actual' denial of the right to vote on account of race." Finding that "none of the plaintiffs . . . claim that they have been denied access to the polls," the court held that "these facts do not allow this Court to conclude that an `actual' denial of the right to vote on account of race occurs." Further, the district court concluded that the contention that "punch card ballots subject [the African-American plaintiffs] to a greater probability that their votes will not be counted than whites," is insufficient to constitute a denial of the right to vote. Thus, the district court held "that the plaintiffs have not established their vote denial claim."

The Supreme Court has said that in interpreting this Act, we should read it in a "a manner that provides the `broadest possible scope' in combating racial discrimination." Chisom, 501 U.S. at 403, 111 S. Ct. 2354 (internal quotation omitted). Moreover, under the 1982 amendments to section 2 of the Act, we are concerned with the results of a practice, not the government's intent. Chisom, 501 U.S. at 394, 111 S. Ct. 2354; 42 U.S.C. § 1973(a).

Section 2(a) of the Act prohibits the use of any electoral practice or procedure that "results in a denial or abridgement of the right of any citizen of the United States to vote on account of race or color." We have noted that

Section 2, unlike other federal legislation that prohibits racial discrimination, does not require proof of discriminatory intent. Instead, a plaintiff need show only that the challenged action or requirement has a discriminatory effect on members of a protected group: A violation of subsection (a) of this section is established if, based on the totality of the circumstances, it is shown that the political processes leading to nomination or election in the State or political subdivision are not equally open to participation by members of a class of citizens protected by subsection (a) of this section in that its members have less opportunity than other members of the electorate to participate in the political process and to elect representatives of their choice.

Moore v. Detroit School Reform Bd., 293 F.3d 352, 363 (6th Cir.2002). The right to vote is defined in 42 U.S.C. § 1973l(c)(1) to include "all action necessary to make a vote effective" including "casting a ballot and having such ballot counted properly." The language of the Act clearly encompasses within the "right of any citizen of the United States to vote" the right to "hav[e] such ballot counted properly." Because the African-American plaintiffs claim that they are disproportionately denied the right to have their ballots counted properly, the district court erred in concluding that the plaintiffs did not state a claim for a violation of the right to vote under the Voting Rights Act.

Finding that the plaintiffs' evidence supports a denial of the right to vote, the next inquiry is under section 2(b) of the Act to determine whether the evidence establishes a violation of the Act. See Chisom, 501 U.S. at 394, 111 S. Ct. 2354 ("Congress not only incorporated the results test in the paragraph that formerly constituted the entire § 2, but also designated that paragraph as subsection (a) and added a new subsection (b) to make clear that an application of the results test requires an inquiry into `the totality of the circumstances.'"). Section 2(b) states that an electoral practice or procedure violates the franchise under section (a)

if, based on the totality of circumstances, it is shown that the political processes leading to nomination or election in the State or political subdivision are not equally open to participation by members of a class of citizens protected by subsection (a) of this section in that its members have less opportunity than other members of the electorate to participate in the political process and to elect representatives of their choice.

42 U.S.C. § 1973(b). Consistent with the results test, this Court has noted that unlike race based claims under the Equal Protection Clause of Fifteenth Amendment, where evidence of discriminatory intent is necessary, "Section 2 of the Voting Rights Act requires only a showing of discriminatory effect." Mixon, 193 F.3d at 407 (citing Thornburg, 478 U.S. at 70-74, 106 S. Ct. 2752). Moreover, claims under the Voting Rights Act require "an intensely local appraisal of the design and impact" of the challenged electoral practice. Thornburg, 478 U.S. at 78, 106 S. Ct. 2752 (internal quotation omitted).

The plaintiffs here alleged that punch cards in these three counties have a discriminatory effect and presented evidence to support their claim. Akin to Southwest Voter, the plaintiffs presented evidence demonstrating that "minority voters disproportionately reside in punch-card counties and that, even within those counties, punch-card machines discard minority votes at a higher rate." 344 F.3d at 918. According to the Ninth Circuit en banc court, "[t]o establish a Section 2 violation, plaintiffs need only demonstrate `a causal connection between the challenged voting practice and [a] prohibited discriminatory result.'" Id. (quoting Smith v. Salt River Project Agric. Improvement & Power Dist., 109 F.3d 586, 597 (9th Cir.1997)).

On this part of the analysis, the district court made very few findings based on its determination under section 2(a). These finding included the finding that there are many residual votes in Appalachian counties with virtually non-existent minority populations, and that undervoting in Franklin County (which was not sued under the Act) is higher than in Hamilton County (which was sued under the Act). Essentially, the court found that because white voters in other counties with minimal minority populations suffer from a similar likelihood that their votes will not be counted, the plaintiffs here have not made a sufficient claim under the Act. The district court did not make any findings under section 2(b) of the Act and did not make the ultimate determination as to whether the plaintiffs proved their claim under the totality of the circumstances. Thus, the appropriate course for this Court is to remand to allow the district court to make detailed findings in the first instance. To aid in the analysis below, however, we do note that Johnson v. De Grandy rejects the argument that vote dilution in one part of the state can be remedied by another part of the state. 512 U.S. 997, 1019, 114 S. Ct. 2647, 129 L. Ed. 2d 775 (1994) (rejecting the premise that "the rights of some minority voters under § 2 may be traded off against the rights of other members of the same minority class"). Thus, what occurred in Franklin County is not necessarily relevant to the success of the plaintiffs' claims in Hamilton, Summit, and Montgomery Counties.

Because the plaintiffs properly stated a claim under the Act, the question then for the district court, under Section 2(b), is whether under the totality of circumstances, the evidence demonstrates a discriminatory result. On remand, the district court will consider the voluminous amount of the plaintiffs' evidence, including the regression analysis showing the correlation between overvoting and the percentage of African-American voters in a given precinct. In Hamilton County the regression coefficient was .517, in Summit County it was .682, and in Montgomery County it was .440. The plaintiffs' expert described these correlations as "strong." The plaintiffs claim that the evidence put forth demonstrates that: (1) African-Americans in Hamilton County overvoted at a rate seven times higher than non African-Americans; (2) in Summit County, African-Americans overvoted at a rate nine times higher than non African-Americans; and (3) in Montgomery County, African-Americans had a residual voting rate 2.5 times that of non African-Americans. The defendants did not present contradictory evidence and did not dispute that racial disparities exist, but instead asserted that education and socioeconomic conditions, not race, account for the disparities. We note, however, that "[t]he essence of a § 2 claim is that a certain electoral law, practice, or structure interacts with social and historical conditions to cause an inequality in the opportunities enjoyed by black and white voters to elect their preferred representatives." Thornburg, 478 U.S. at 47, 106 S. Ct. 2752. With these considerations in mind, we vacate the district court's ruling in favor of the defendants on the Voting Rights Act claim and remand for further proceedings consistent with this opinion.

We also reverse the district court's judgment denying class and sub-class certification. The district court's nine-line order states that the plaintiffs reside in only four of the eighty-eight Ohio counties and therefore they "are not proper representatives for voters in other counties." The court also concluded that although the Secretary of State is the chief elections officer, each county has its own board of elections with substantial powers.

The plaintiffs sought to certify two classes. The first plaintiff class would consist of: "All persons who vote, or who are registered to vote in jurisdictions within the State of Ohio which do not employ ballot marking or other voting machinery with in-precinct error notification technology." The second class, for purposes of the Voting Rights Act claim consists of: "All African-American persons who vote, or who are registered to vote in jurisdictions within the State of Ohio which do not employ ballot marking or other voting machinery with in-precinct error notification technology." The first step in the class certification process requires the plaintiffs to satisfy Rule 23(a). That is,

One or more members of a class may sue or be sued as representative parties on behalf of all only if (1) the class is so numerous that joinder of all members is impracticable, (2) there are questions of law or fact common to the class, (3) the claims or defenses of the representative parties are typical of the claims or defenses of the class, and (4) the representative parties will fairly and adequately protect the interests of the class.

Fed.R.Civ.P. 23(a).

The district court found that "the plaintiffs are not proper representatives for voters in other counties." We disagree. The plaintiffs clearly satisfy the requirements of Rule 23(a) and have satisfied their burden of so proving. See Senter v. Gen. Motors Corp., 532 F.2d 511, 522 (6th Cir.1976). The class is so numerous that joinder is impracticable, there are questions of law and fact common to the class, the claims of the representative parties are typical of the claims of the class, and the representatives will fairly and adequately protect the interests of the class. Further, we find that the plaintiffs have satisfied their burden under the second step of the analysis in Rule 23(b). Under section (b)(2), "the party opposing the class [Secretary of State] has acted or refused to act [by decertifying deficient technology] on grounds generally applicable to the class, thereby making appropriate final injunctive relief or corresponding declaratory relief with respect to the class as a whole." Thus, we reverse the district court's order and conclude that class certification is appropriate for the plaintiffs pursuing state-wide relief under the Equal Protection Clause and sub-class certification is appropriate for the African-American plaintiffs within Hamilton, Summit, and Montgomery Counties.

Violations of the Equal Protection Clause are no less deserving of protection because they are accomplished with a modern machine than with outdated prejudices. The dissent's appeal to caution notwithstanding, "the basic ingredient of [judicial] decision is principle, and it should not be compromised and parceled out a little in one case, a little more in another, until eventually someone receives the full benefit. If the principle is sound and constitutional, it is the birthright of every American, not to be accorded begrudgingly or piecemeal or to special groups only, but to everyone in its entirety whenever it is brought into play." EARL WARREN, THE MEMOIRS OF CHIEF JUSTICE EARL WARREN 6 (Madison Books 2001) (1977). The district court's decision with regard to the plaintiffs' equal protection claim is REVERSED. The case is REMANDED with instructions to enter judgment in favor of the plaintiffs. With regard to the plaintiffs' Voting Rights Act claim, the district court's decision is VACATED, and REMANDED for proceedings consistent with this opinion.

RONALD LEE GILMAN, Circuit Judge, dissenting.

The majority today imposes by judicial decree important changes to the Ohio electoral system, doing so largely in reliance on the Supreme Court's murky decision in Bush v. Gore, 531 U.S. 98, 121 S. Ct. 525, 148 L. Ed. 2d 388 (2000) (per curiam), a vacated Ninth Circuit panel opinion in a California recall-election dispute, and two district court cases that never reached a final judgment on the merits. In my view, these sources cannot bear the weight that the majority places on them, and should not form the basis for subjecting an indeterminate number of state and local election decisions to the strictest level of constitutional scrutiny.

I would take a more cautious approach, one that recognizes the primacy of the executive and legislative branches in the electoral process and the significant costs that the majority's holding will place on state and local governments, the entities that are charged with most aspects of election administration. Although I agree with the majority that the plaintiffs have standing and that the case is not moot, I cannot join the majority's analysis of the equal protection and Voting Rights Act challenges brought by the plaintiffs, or its cursory reversal of the district court's denial of class certification. Accordingly, I respectfully dissent.

My analysis of the equal protection question will proceed in four steps. I will first explain why I believe that the Supreme Court voting-rights precedents cited by the majority are distinguishable from the present case and do not establish, as the majority asserts, that strict scrutiny is the proper constitutional standard of review. The second and third parts of my analysis question the precedential value of Bush v. Gore and the other authorities on which the majority relies, demonstrating why an expansive reading of those authorities threatens to constitutionalize the procedures and mechanisms used to cast a vote under the guise of advancing still unsettled notions of equality. Finally, I will explore the practical effect of the majority's holding in the wake of the Help America Vote Act (HAVA) and Ohio's promise to comply with the terms of HAVA.

A reader unfamiliar with the Supreme Court's voting-rights precedents might conclude after reading the majority opinion that federal courts have long dealt with challenges analogous to the one made in the present case, and that strict scrutiny has been universally recognized as the appropriate standard of review. The majority's lengthy exposition of the fundamental nature of the right to vote cites many cases purportedly to that effect, taking quotations out of their factual and legal context. A closer examination of those cases, however, reveals that the challenges upheld in them were significantly different from the one before us in the present case, and that the appropriate standard of review in voting-rights cases is far from settled.

In Gray v. Sanders, 372 U.S. 368, 83 S. Ct. 801, 9 L. Ed. 2d 821 (1963), for example, the Supreme Court invalidated Georgia's "county unit system" for counting votes in party primaries that were held to nominate candidates for the United States Senate and several state offices. Under the original county unit system, a resident in the least populous county in the state "had an influence in the nomination of candidates equivalent to 99 residents" of the state's most populous county. Id. at 371, 83 S. Ct. 801. The state amended the system to increase the influence of the more populous counties, but the district court still found that "the vote of each citizen counts for less and less as the population of the county of his residence increases." Id. at 372-73, 83 S. Ct. 801 (citation omitted). Articulating its famous "one person, one vote" standard, id. at 381, 83 S. Ct. 801, the Court held that the county-unit system violated the Equal Protection Clause because it gave "one person. . . twice or 10 times the voting power of another person in a statewide election merely because he lives in a rural area or because he lives in the smallest rural county[.]" Id. at 379, 83 S. Ct. 801. The Court, without announcing a standard of review, explained that the constitutional infirmity in the Georgia system inhered in the system's failure to provide "equality of voting power." Id. at 381, 83 S. Ct. 801.

Wesberry v. Sanders, 376 U.S. 1, 84 S. Ct. 526, 11 L. Ed. 2d 481 (1964), also cited by the majority, is not even an equal protection case. The Wesberry Court did not hold, as the majority states, that "Equal Protection requires substantial equality of population amongst the districts established by state legislatures for the election of members to the United States House of Representatives." Maj. Op. at 857 (emphasis added). To the contrary, the Court relied exclusively on two sections found in Article I of the Constitution, stating explicitly that it did not need to "reach the arguments that the Georgia statute violates the Due Process, Equal Protection and Privileges and Immunities Clauses of the Fourteenth Amendment." 376 U.S. at 8, 84 S. Ct. 526 n. 10; see also Daniel Hays Lowenstein & Richard L. Hasen, Election Law: Cases and Materials 113 (2d ed. 2001) ("Wesberry was not decided under the Equal Protection Clause but under Article I, §§ 2 and 4 of the Constitution, and was thus applicable only to the United States House of Representatives."). Article I, §§ 2 and 4 of the Constitution, the Court held, barred states from drawing congressional districts in which "a vote is worth more in one district than in another." Id. at 8, 84 S. Ct. 526. Because it never addressed the plaintiff's equal protection challenge, the Court did not need to — and did not — articulate the appropriate standard of review.

The pathbreaking decision in Reynolds v. Sims, 377 U.S. 533, 84 S. Ct. 1362, 12 L. Ed. 2d 506 (1964), marked the first time that the Court announced a standard of review similar to the strict-scrutiny standard currently applied in reviewing laws that discriminate against suspect classes or infringe on certain constitutionally protected rights. See id. at 562, 84 S. Ct. 1362 ("[A]ny alleged infringement of the right of citizens to vote must be carefully and meticulously scrutinized."). Again, however, the Court announced this standard in a case that dealt with the "weighting of votes;" specifically, unequally populated state legislative districts. See Lowenstein & Hasen, supra, at 105, 114; Reynolds, 377 U.S. at 563, 84 S. Ct. 1362 ("Over-weighting and overvaluation of the votes of those living here has the certain effect of dilution and undervaluation of the votes of those living there."). Although it invalidated Alabama's legislative apportionment plan and held that both houses in a bicameral state legislature must be apportioned on a "substantially equal population basis," the Court did not do so on the ground that the law was not narrowly tailored to further a compelling governmental interest, see, e.g., Adarand Constructors, Inc. v. Pena, 515 U.S. 200, 235, 115 S. Ct. 2097, 132 L. Ed. 2d 158 (1995), but instead because the existing apportionment plan violated the "one person, one vote" principle — that "the weight of a citizen's vote cannot be made to depend on where he lives." Reynolds, 377 U.S. at 567, 84 S. Ct. 1362 (emphasis added).

The two other cases subjecting state voting practices to strict scrutiny cited by the majority are likewise inapposite. In Harper v. Virginia Board of Elections, 383 U.S. 663, 668, 86 S. Ct. 1079, 16 L. Ed. 2d 169 (1966), the Court invalidated the applicable laws of the four remaining states that charged a poll tax to vote in state elections. The Court applied strict scrutiny in Harper, citing Reynolds for the proposition that, "where fundamental rights and liberties are asserted under the Equal Protection Clause, classifications [that] might invade or restrain them must be closely scrutinized and carefully confined." Id. at 670, 86 S. Ct. 1079. Virginia's poll tax, the Court reasoned, determined the qualifications of voters on the basis of wealth, a ground that is "traditionally disfavored," id. at 668, 86 S. Ct. 1079, and was therefore unconstitutional.

The second of the two cases is Dunn v. Blumstein, 405 U.S. 330, 334, 92 S. Ct. 995, 31 L. Ed. 2d 274 (1972), where the Court struck down Tennessee's requirement that, in order to qualify as a registered voter, a person must have resided in the state for at least one year and in the county for at least three months. Strict scrutiny was the appropriate standard of review, the Court held, because the durational-residency requirement affected two fundamental rights — the right to vote and the right to travel. See id. at 338, 92 S. Ct. 995 (explaining that strict scrutiny applied, despite an earlier durational-residency-requirement case using a less stringent standard of review, because the Tennessee law "penaliz[ed]" only people who had recently exercised their right to travel). Dunn is therefore unique because the Court faced a state law that impinged on two fundamental rights protected by the Fourteenth Amendment. Indeed, courts since Dunn have emphasized the importance of both the Tennessee law's impact on two constitutionally protected activities and Dunn's status as a voter-qualification precedent. See, e.g., Donatelli v. Mitchell, 2 F.3d 508, 515 (3d Cir.1993) (explaining that the Court in Dunn reviewed a law that both denied access to the ballot and "directly burdened citizens' fundamental constitutional right to travel"); Greidinger v. Davis, 988 F.2d 1344, 1349 (4th Cir.1993) (classifying Dunn as a case "involving voter qualifications and ballot access").

Stated simply, none of these Supreme Court precedents resembles the claim made by the plaintiffs in the present case. These plaintiffs do not allege that Ohio has imposed an impermissible voter-qualification requirement, as in Harper and Dunn. Nor do they argue that the votes in Ohio are "weighted" differently — that is, that a properly marked ballot "is worth more in one district than in another." Wesberry, 376 U.S. at 8, 84 S. Ct. 526; see also Reynolds, 377 U.S. at 563, 84 S. Ct. 1362. Instead, they maintain that voters in counties that employ inferior voting equipment "are subjected to a significantly greater risk that their votes will not be counted." In other words, properly marked ballots are counted and accorded the same value everywhere in the state, but voters in some counties have a reduced chance of turning in a properly marked ballot that will actually be counted because certain voting methods do not catch the voters' own inadvertent mistakes.

This challenge to the nuts-and-bolts of election administration, regardless of its merit, cannot be equated with either discriminatory voter-qualification requirements or generally applicable state laws that deny "equality of voting power," Gray, 372 U.S. at 381, 83 S. Ct. 801, which are the principal types of state actions that the Supreme Court has subjected to strict scrutiny. See generally Richard L. Hasen, Bush v. Gore and the Future of Equal Protection Law in Elections, 29 Fla. St. U.L.Rev. 377, 393 (2001) (explaining that the Warren Court precedents govern only the equal opportunity to vote and the equal weighting of votes, not "equality in the procedures and mechanisms used for voting").

B. Rational basis is the proper standard of review

The Supreme Court has never adhered to the view that all legislation or practices that affect the right to vote must be subjected to strict scrutiny. See Greidinger, 988 F.2d at 1350 (collecting cases in which the Supreme Court has applied rational-basis scrutiny in upholding the denial of absentee ballots to pretrial detainees and in sustaining a state requirement that persons seeking to vote in a primary must register with a political party a specified number of days before a general election). As the First Circuit succinctly explained, "[t]he Supreme Court has eschewed a hard-and-fast rule, and instead has adopted a flexible framework for testing the validity of election regulations." Werme v. Merrill, 84 F.3d 479, 483 (1st Cir.1996). That "flexible framework" was most recently articulated by the Court in Burdick v. Takushi, 504 U.S. 428, 112 S. Ct. 2059, 119 L. Ed. 2d 245 (1992), a case that the majority attempts to limit to the specific setting of political candidates' access to the ballot. Maj. Op. at 859-63. I believe, however, that Burdick sets forth a general test for evaluating challenges to state voting laws and practices under the Fourteenth Amendment, and therefore cannot be so easily "dismiss[ed]." Maj. Op. at 861.

The Supreme Court in Burdick began by observing that "[e]lection laws will invariably impose some burden upon individual voters." Burdick, 504 U.S. at 433, 112 S. Ct. 2059; see also Harper, 383 U.S. at 673, 86 S. Ct. 1079 (Black, J., dissenting) ("All voting laws treat some persons differently from others in some respects."). Subjecting "every voting regulation to strict scrutiny," the Court warned, "would tie the hands of States seeking to assure that elections are operated equitably and efficiently." Burdick, 504 U.S. at 433, 112 S. Ct. 2059. Courts facing challenges to state election laws or practices should therefore start by "weigh[ing] the character and magnitude of the asserted injury" to voting rights against the state's articulated interests, "taking into consideration the extent to which those interests make it necessary to burden the [voter's] rights." Id. at 434, 112 S. Ct. 2059 (citations and quotation marks omitted). The Court continued:

Under this standard, the rigorousness of [the judicial] inquiry into the propriety of a state election law depends upon the extent to which a challenged regulation burdens First and Fourteenth Amendment rights. Thus, . . . when those rights are subjected to severe restrictions, the regulation must be narrowly drawn to advance a state interest of compelling importance. But when a state election law provision imposes only reasonable, nondiscriminatory restrictions upon the First and Fourteenth Amendment rights of voters, the State's important regulatory interests are generally sufficient to justify the restrictions.

Id. (citations and quotation marks omitted). In other words, only state laws and regulations that impose "severe restrictions" on the right to vote should be subjected to strict scrutiny.

Federal courts applying the Burdick framework have evaluated challenges to various state voting regulations under a rational-basis standard of review. See, e.g., Werme, 84 F.3d at 485-86 (upholding under rational-basis review a New Hampshire law that prevented a member of the Libertarian Party from serving as a ballot clerk on Election Day); Donatelli, 2 F.3d at 514-15 (upholding under rational-basis review a state reapportionment plan that temporarily assigned a state senator to a district that had not elected him); see also id. at 515-16 (collecting court-of-appeals and district-court cases that reviewed claims of "temporary disenfranchisement" due to reapportionment under a rational-basis standard). These cases from our sister circuits serve as reminder that, the tone of the majority opinion notwithstanding, many state and local voting regulations are evaluated under the rational-basis standard of review.

I believe that this court's decision in Mixon v. Ohio, 193 F.3d 389 (6th Cir. 1999), establishes a framework consistent with the one articulated by the Supreme Court in Burdick. In Mixon, this court held that state legislation that "grants the right to vote to some residents while denying the vote to others" is subject to strict scrutiny, whereas legislation that "does not infringe on the right to vote" is examined "under the rational basis standard." Id. at 402. Two aspects of this statement stand out. The first is that Mixon might not readily apply to the present case because the plaintiffs have not challenged any legislation that "grants the right to vote to some while denying the vote to others." See id. Instead, they are challenging "electoral practices," none of which purport to grant the right to vote to anyone. This court has not yet decided whether the Mixon analysis applies to state practices other than legislation.

Assuming that Mixon does apply, however, there is a second noteworthy aspect of the standard enunciated in that case — namely, that this court drew a distinction between state laws and regulations that "infringe on the right to vote" and those that do not. Id. This standard approximates the line drawn by the Supreme Court in Burdick, where the Court emphasized the difference between laws that severely restrict voting rights and those that "impose[] only reasonable, nondiscriminatory restrictions" on such rights. Burdick, 504 U.S. at 434, 112 S. Ct. 2059. If we are to harmonize Mixon and Burdick, as I think we must, then we should review the challenged voting practices under the rational-basis standard because those practices do not "severe[ly] restrict[]" the plaintiffs' right to vote. See id.

Also consistent with Burdick is the Ninth Circuit's decision in Weber v. Shelley, 347 F.3d 1101 (9th Cir.2003), a case relied on by the district court below and one that I believe is directly on point. See Stewart v. Blackwell, 356 F. Supp. 2d 791, 800 (S.D.Ohio 2004). In Weber, a voter brought an Equal Protection challenge to a California county's decision to replace paper ballots with an electronic-voting system. 347 F.3d at 1102. The computerized system, the voter maintained, was susceptible to manipulation and potential fraud because it lacked a paper audit trail. Id. at 1104-1105. Applying the Burdick framework, the Ninth Circuit held that the "use of paperless, touchscreen voting systems [did not] severely restrict[] the right to vote," and that the voting procedure was therefore subject only to rational-basis review. Id. at 1106. Because the electronic-voting machines constituted a reasonable method of ameliorating the problems inherent in paper ballots, the change in voting procedures was rationally related to the state's interest in "ensuring that elections are fair and orderly." Id.; see also id. at 1107 (opining that the state and the county had "made a reasonable, politically neutral and non-discriminatory choice to certify [and to use] touchscreen systems as an alternative to paper ballots"); accord Am. Ass'n of People with Disabilities v. Shelley, 324 F. Supp. 2d 1120, 1127 (C.D.Cal.2004) (reviewing the Calfiornia Secretary of State's decision to decertify Direct Recording Electronic (DRE) voting machines under the rational-basis standard of review).

Weber is significant for a number of reasons. First, when faced with an Equal Protection challenge to the type of voting technology certified by the California Secretary of State and utilized by the county, the Ninth Circuit found the appropriate standard of review in Burdick — not in Reynolds, Harper, Dunn, or any other pre-Burdick precedents. See Weber, 347 F.3d at 1106. The second reason is that, in applying Burdick, the Ninth Circuit held that the increased potential for voter fraud with the new voting technology did not constitute a "severe restriction" on the right to vote, so that the decision to implement such electoral changes was subject only to rational-basis review. Id.

Finally, and perhaps most importantly, Weber is the closest on its facts to the challenge levied by the plaintiffs in the present case. Indeed, Weber is the exact converse of the present case. Whereas the plaintiffs here decry the refusal of the counties to utilize improved voting technologies, the voter in Weber complained precisely because her county had abandoned traditional paper ballots in favor of the newly available technology. The contrast between the two cases epitomizes the predicament in which election officials find themselves — damned if they upgrade voting machines, damned if they do not — and further supports adopting a constitutional standard of review that affords state and local officials the flexibility necessary to regulate an election system constantly in flux.

I cannot help but note, however, the majority's insistence that the result here would be the same even if it applied rational-basis scrutiny. Maj. Op. at 872. The majority makes this unsupported assertion despite having chided the district court for doing the same thing — that is, summarily concluding that the result would be the same regardless of the appropriate standard of review. Maj. Op. at 852-53. If the state's proffered justifications for employing certain voting methods were truly as flimsy as the majority portrays them, then the majority would have no need to discuss at length why the state's practices must be judged under the strict-scrutiny standard.

The proper standard of review, in my opinion, is critical to the judicial inquiry in this case. Because I believe that the Supreme Court's decision in Burdick has supplanted the factually and legally distinguishable precedents on which the majority relies, and that the use of different types of voting technologies does not constitute a severe restriction on the right to vote, I would evaluate the challenged voting practices under the rational-basis standard of review. I reach this conclusion notwithstanding the Court's post-Burdick opinion in Bush v. Gore, 531 U.S. 98, 121 S. Ct. 525, 148 L. Ed. 2d 388 (2000) (per curiam), a case whose precedential value, as I will explain, is at best questionable.

II. EQUAL PROTECTION ANALYSIS AND BUSH V. GORE

As I have explained above, neither the venerable Warren Court precedents nor current Supreme Court voting-rights cases support either the majority's decision to subject the challenged voting practices to strict scrutiny or the majority's conclusion that those practices are unconstitutional. What actually provides the analytical basis for the majority opinion, then, is the Supreme Court's decision in Bush v. Gore and a series of lower-court cases that have purported to adopt the reasoning of that decision. For the reasons ably articulated by a leading election-law expert — reasons to which I will add a few thoughts of my own — I believe that we should heed the Supreme Court's own warning and limit the reach of Bush v. Gore to the peculiar and extraordinary facts of that case. See Hasen, Bush v. Gore and the Future of Equal Protection Law in Elections, 29 Fla. St. U.L.Rev. at 379 ("Language in the per curiam opinion limits [the holding] to the facts of the case, or, at most, to cases where jurisdiction-wide recounts are ordered."). The majority has chosen a different path, one that unjustifiably expands Bush v. Gore into a landmark precedent designed to fundamentally transform federal election law.

A. Bush v. Gore should not be given an expansive reading

Professor Hasen summarized his three "reasons for doubting Bush v. Gore's precedential value" as follows:

[T]he limiting language in the opinion, the lack of seriousness with which the Court undertook its own analysis, and the inconsistency with other jurisprudence by this majority of Justices all point in the direction of assuming that Bush v. Gore is not good precedent for an expansive reading of equal protection law in elections.

Id. at 391. I will focus on the first two of these reasons. In its per curiam opinion, the Court majority cautioned that its analysis was "limited to the present circumstances, for the problem of equal protection in election processes generally presents many complexities." Bush v. Gore, 531 U.S. at 109, 121 S. Ct. 525. The Court elaborated as follows:

The question before [us] is not whether local entities, in the exercise of their expertise, may develop different systems for implementing elections. Instead, we are presented with a situation where a state court with the power to assure uniformity has ordered a statewide recount with minimal procedural safeguards. When a court orders a statewide remedy, there must be at least some assurance that the rudimentary requirements of equal treatment and fundamental fairness are satisfied.

Id.; see also id. at 134, 121 S. Ct. 525 (Souter, J., dissenting) ("It is true that the Equal Protection Clause does not forbid the use of a variety of voting mechanisms within a jurisdiction, even though different mechanisms will have different levels of effectiveness in recording voters' intentions[.]"); Vikram David Amar, Adventures in Direct Democracy: The Top Ten Constitutional Lessons from the California Recall Experience, 92 Cal. L.Rev. 927, 955 (2004) (explaining that the Supreme Court in Bush v. Gore "explicitly disavowed" the notion that "the Equal Protection Clause . . . invalidates any statewide election where different kinds of voting machinery throughout the state may lead to nontrivial differential error rates across counties").

In the present case, of course, no state-court order is at issue, and no governmental entity has ordered a "statewide remedy." The allegations of these plaintiffs are a far cry from the lack of uniform rules for discerning the meaning of votes already cast, which is what the Court in Bush v. Gore found to be a constitutional violation. See Hasen, 29 Fla. St. U.L.Rev. at 384 (explaining that the Bush v. Gore majority identified the Florida Supreme Court's failure to formulate uniform rules for determining voter intent as the reason the state court's recount procedures violated the Equal Protection Clause). These significant differences alone constitute a "legitimate basis [and] principled manner of distinguishing Bush v. Gore." See Maj. Op. at 875.

Since Professor Hasen's article, the Supreme Court has had ample opportunity to prove him wrong by explaining, or even citing to, its decision in Bush v. Gore. See, e.g., Clingman v. Beaver, 544 U.S. 581, 125 S. Ct. 2029, 161 L. Ed. 2d 920 (2005); Vieth v. Jubelirer, 541 U.S. 267, 124 S. Ct. 1769, 158 L. Ed. 2d 546 (2004); McConnell v. Fed. Election Comm'n, 540 U.S. 93, 124 S. Ct. 619, 157 L. Ed. 2d 491 (2003); Georgia v. Ashcroft, 539 U.S. 461, 123 S. Ct. 2498, 156 L. Ed. 2d 428 (2003). But despite taking a steady load of election-related cases, the Court has not cited Bush v. Gore even once — not in a majority opinion, in a concurrence, or in a dissent — in the more than five years since that case was decided. The Court, in other words, has adhered to the limiting language that it conspicuously included in the Bush v. Gore opinion. I believe that we should follow the same path.

Bush v. Gore also provides little support for the majority's conclusion that a decision by a state or local government to employ certain voting technology must be subjected to strict scrutiny. As Professor Hasen has explained, the Bush v. Gore Court described voting as a fundamental right, but did not "even bother[] to undertake" the "hornbook" analysis of asking whether the state's interest in meeting an impending federal election deadline was "compelling" enough to overcome the fundamental right of voters to have their vote counted. See Hasen, 29 Fla. St. U.L.Rev. at 389. I am not convinced that the Court's unexplained citation to Reynolds and Harper can be read as displacing the Burdick framework and instructing lower courts to employ strict scrutiny, particularly when the Court itself neither applied that standard nor articulated any standard of review.

An expansive reading of Bush v. Gore portends problems far graver than simply a heightened standard of review. Professor Hasen has convincingly shown that Bush v. Gore applied the Equal Protection Clause to an area of election law entirely different from prior caselaw, but did so without explaining "which kinds of procedures and mechanisms used for voting constitute arbitrary and disparate treatment that value one person's vote over another" and therefore violate the Constitution. Hasen, 29 Fla. St. U.L.Rev. at 393 (quotation marks omitted). The absence of guidance on these crucial matters leaves lower courts to evaluate for themselves not just the jurisprudential complexities that arise in election law cases of first impression, but also the potential policy implications of extending a Supreme Court precedent whose rationale is far from clear.

Those policy implications, in my view, counsel strongly against extending the rationale of Bush v. Gore to the distinct area of voting technology. Professor Hasen has focused on three significant concerns, beyond the obvious monetary ones, raised by constitutionalizing local governments' choice of voting mechanisms. The first is that doing so "eases the way for federal court intervention in state and local elections over nuts-and-bolts disputes better left to local authorities." Id. at 400. Like Judge Posner, I do not think that the Court meant to invite more legal challenges to election methods and results. See Richard A. Posner, Florida 2000: A Legal and Statistical Analysis of the Election Deadlock and the Ensuing Litigation, 2000 Sup.Ct. Rev. 1, 41 (2001) ("The last thing we need is more election litigation.").

This first concern gives rise to the second one — namely, that a federal constitutional rule on voting technology "undermines federalism" by imposing a uniform national standard on localities that might otherwise allocate their resources in a different yet equally just manner. See Hasen, 29 Fla. St. U.L.Rev. at 401 ("Bush v. Gore is tantamount to a holding that the purchase of ambulances by a very poor county is less important than a move from punch cards to optical scanners.").

Professor Hasen's third concern, however, is what I see as the most important downside to extending Bush v. Gore to voting technology. See id. at 401 (reading Bush v. Gore as creating "disincentive[s]. . . for jurisdictions to experiment with new methods of voting"). If the Equal Protection Clause bars one county in a state from adopting a new voting mechanism simply because the new mechanism may create a disparity between that county and its neighbors, local officials will have little incentive to innovate. The reason is simple: each time a county implements a change in equipment, it risks becoming party to a lawsuit alleging disparities in voters' chances of properly completing a ballot and having that ballot counted. Because of this powerful disincentive to innovate, Bush v. Gore, absent other legislative solutions, might "have the unintended effect of freezing our voting mechanics at the current level of technology." Id. at 402. I will return to this point in Part III below.

In the final analysis, I believe that the best course is to understand the Supreme Court's decision in Bush v. Gore as a shield to preserve the status quo in an electoral process beset by extraordinary temporal and political pressures. I therefore decline to join the majority in permitting litigants to use that decision as a sword to strike down state election policies that, while ripe for improvement, were previously on solid constitutional ground. See Posner, 2000 Sup.Ct. Rev. at 41 (explaining that differences in the technology and vote-counting methods used by counties "had not previously been thought to deny equal protection of the laws").

In reaching this conclusion, I am not, as the majority charges, making the nonsensical claim that Professor Hasen's law review article has overruled Bush v. Gore. See Maj. Op. at 874. I am instead faithfully following the Supreme Court's explicit admonition in its decision that the analysis employed was "limited to the present circumstances." Bush, 531 U.S. at 109, 121 S. Ct. 525. Lost in the majority's rebuttal to my dissent is the recognition of a key aspect of Bush v. Gore that we all agree is clear: the Court's unequivocal statement that it was not announcing a general rule for future cases. See Maj. Op. at 859 n. 9 (acknowledging the Court's limiting language). Unlike my colleagues, I cannot look past the Court's own words and attribute to the Court the intention — in a per curiam opinion issued one day after oral argument in the midst of a national crisis — to revolutionize the election laws and practices of this country.

B. Other lower courts have improperly expanded the reach of Bush v. Gore

This is not the first case in which plaintiffs have utilized Bush v. Gore as their principal tool for effecting change in electoral policy. See Steven J. Mulroy, Lemonade from Lemons: Can Advocates Convert Bush v. Gore Into a Vehicle for Reform?, 9 Geo. J. on Poverty L. & Pol'y 357, 358-59 (2002) (citing various cases, all filed in the wake of Bush v. Gore, that challenge the use of certain voting methods under the Equal Protection Clause and the Voting Rights Act). The majority bases much of its analysis on the strength of these test cases, including Black v. McGuffage, 209 F. Supp. 2d 889 (N.D.Ill.2002), and Common Cause v. Jones, 213 F. Supp. 2d 1106 (C.D.Cal.2001). Maj. Op. at 862-67; see Mulroy, 9 Geo. J. on Poverty L. & Pol'y at 358-60 (citing Black and Common Cause as cases that "voting rights advocates" filed in an attempt "to use [Bush v. Gore] to push for long overdue electoral reform, by invoking the very equal protection theory relied on by the conservative Bush majority"). Like the district court below, however, I find these cases unpersuasive.

The chief weakness of these cases is their reliance on the per curiam opinion in Bush v. Gore as the primary basis for their decisions. See Southwest Voter Registration Educ. Project v. Shelley, 344 F.3d 882, 894 (9th Cir.) (per curiam) (Shelley I), rev'd en banc, 344 F.3d 914 (9th Cir.2003) (per curiam) (Shelley II); Black, 209 F.Supp.2d at 898; Common Cause, 213 F.Supp.2d at 1109. For the reasons set forth in Part II.A. above, I do not view Bush v. Gore as a landmark precedent designed to vastly extend the reach of the Equal Protection Clause.

The case that most openly relies on Bush v. Gore is the Ninth Circuit's now-vacated panel decision in Shelley I. In an opinion that looks and sounds strikingly like that of the majority in the present case, a panel of the Ninth Circuit enjoined the use of punch-card ballots in the 2003 gubernatorial recall election in California. 344 F.3d at 888. The theory of the California plaintiffs, which is identical to the one advanced by the Ohio plaintiffs in this case, was "that using error-prone voting equipment in some counties, but not in others[,] will result in votes being counted differently among the counties." Id. at 895. Describing the plaintiffs' allegations as stating "a classic voting rights equal protection claim," id. at 895, the panel concluded that the plaintiffs had "demonstrated a sufficient likelihood of success on the merits regardless" of whether strict scrutiny or rational basis was the appropriate standard of review. Id. at 900. Success on the merits was likely, in the panel's view, because the plaintiffs' claim "present[ed] almost precisely the same issue as the [Supreme] Court considered in Bush," where the Court found an Equal Protection violation. Id. at 895. Within a week of the panel's decision, however, the Ninth Circuit granted en banc review, vacated the panel opinion, and unanimously reached a conclusion opposite to that of the panel. Shelley II, 344 F.3d at 917, 920.

Despite the rather firm rebuke that the en banc court dealt the panel decision in Shelley I, the majority insists that the Ninth Circuit "did not reject outright the Equal Protection argument pursuant to the Bush v. Gore rationale . . . ." Maj. Op. at 866. I concede that the rejection was not "outright," but the en banc court's description of the panel's rationale as "one over which reasonable jurists may differ" can hardly be construed as a ringing endorsement of Bush v. Gore's precedential value or its application to the issue of variations in voting technology. Shelley II, 344 F.3d at 918. At the very least, the Ninth Circuit's unanimous decision in Shelley II divested the panel decision of whatever precedential value it otherwise would have had.

Undeterred, the majority struggles to accord such value to Shelley I because the rationale adopted by the majority is virtually identical to the one provided by the Shelley I panel — namely, that the plaintiffs' novel equal protection claims are controlled by Supreme Court precedents that mandate the equal weighting of votes. Compare Maj. Op. at 870-71 (holding that the counties' use of outdated technology "results in the weighting of votes differently"), with Shelley I, 344 F.3d at 894-95 (describing the plaintiffs' contention "that a vote cast in Los Angeles or San Diego is entitled to the same weight as a vote cast in San Francisco"). As I explained in Part I.A. above, however, the plaintiffs are not asserting that one properly marked and tabulated vote has less value than another, but instead that certain technologies decrease a voter's chance of handing in a properly marked ballot that will then be tabulated. This claim, when properly characterized, does not fall within the ambit of Gray, Wesberry, Reynolds, or any of the Supreme Court's caselaw establishing the right to an equally-weighted vote.

Stripped of their connection to longstanding Supreme Court precedent, the majority opinion and Shelley I are persuasive only to the extent that Bush v. Gore is controlling. Neither opinion, in my view, successfully refutes the compelling reasons supplied by Professor Hasen for refusing to "take Bush v. Gore's equal protection holding seriously." Hasen, 29 Fla. St. U.L.Rev. at 380. What I believe makes the majority's opinion even more troubling than that of the Ninth Circuit panel is its failure to articulate precisely which aspects of Ohio's voting system violate the Equal Protection Clause, and what state and local officials should do in the future to ameliorate the perceived constitutional problems. I turn now to these concerns.

The majority holds that the certification and use of non-notice voting technology in one county but not in another violates the Equal Protection Clause because "punch card and central-count optical scan technologies result in a greater likelihood that one's vote will not be counted on the same terms as the vote of someone in a notice county." Maj. Op. at 871. Notwithstanding its cautionary language to the contrary, see Maj. Op. at 876-77, the majority is calling for virtually absolute equality in voting methods and procedures across all electoral districts in Ohio.

A closer look at some of the relevant numbers tells much of the story. The key statistic, as the plaintiffs present it, is that the residual-vote rate across the state was 2.29% for punch-card machines, 0.94% for electronic-voting machines, 1.04% for lever machines, and 1.15% for precinct-count optical scanners. In other words, the error rate in punch-card districts is twice that of counties with precinct-count optical scan equipment, and more than twice that of counties using lever or electronic-voting machines. The plaintiffs also rely on direct inter-county comparisons, pointing out that Cuyahoga County, which used punch-cards in the presidential elections at issue, had an error rate four times higher than Franklin County, where electronic-voting machines were used.

What the majority does not indicate, however, is which of these numbers is constitutionally significant. Is the constitutional problem the fact that voters in punch-card districts face an error rate twice as high as voters in precinct-count optical-scan districts? If so, then improvements in technology will not necessarily cure the equal protection problem. This is so because a voter in a county with a much-improved residual vote rate of 0.2% could point to a neighboring county where the rate is 0.1% and say the same thing — namely, that the error rate in her county is twice as high as in the neighboring county. Moreover, as voting technology improves and error rates decline, whatever disparities remain among the different types of voting equipment will be magnified. That is to say, even a technology that yields a negligible error rate of 0.1% might violate the Equal Protection Clause if other jurisdictions within the state employ technology with a rate of 0.01%, which is one-tenth as high. The majority refers to a "greater likelihood" that a vote will not be counted, but fails to articulate a coherent constitutional threshold — a point at which such a likelihood renders state voting practices unconstitutional.

Inter-county comparisons likewise fail to clarify the majority's constitutional ruling. According to the plaintiffs, a disparity such as the one between Franklin and Cuyahoga Counties — where the residual-vote rate in the latter is four times that of the former — triggers strict scrutiny and, in all likelihood, dooms the challenged county practice. Under that logic, however, punch cards and central-count optical scan equipment are not the only voting mechanisms that might violate the Equal Protection Clause. Lucas County, which used a lever-style AVM (Automatic Voting Machine), had a residual vote rate of 0.4% in the 2000 presidential election. See Stewart v. Blackwell, 356 F. Supp. 2d 791, 826 (S.D.Ohio 2004). In that same election, Ross County used electronic voting and had a residual rate of 1.2% — three times that of Lucas County. See id. Do these numbers dictate that, in a suit by a voter from Ross County, the county's decision to use electronic-voting machines should be reviewed under a strict-scrutiny standard? After all, the continued use of the voting machines subjects that voter to a "greater likelihood" than her counterpart in Lucas County that her vote will be excluded as improperly cast.

The strict-scrutiny standard exacerbates the hazards posed by the majority's inexact approach. Any incremental change that leads to a potential disparity in voter-error rates becomes susceptible to legal challenges in which the state must show that its chosen practice or regulation is "necessary to promote a compelling state interest." Mixon v. Ohio, 193 F.3d 389, 402 (6th Cir.1999); see also Craigmiles v. Giles, 312 F.3d 220, 223 (6th Cir.2002) (explaining that a challenged state regulation survives strict scrutiny only if it "serve[s] a compelling state purpose and [is] narrowly tailored to achieving that purpose"). But there is no guarantee that local governments would be able to satisfy this stringent standard when making the most mundane of decisions, such as replacing one brand of electronic-voting equipment (with, say, a 0.2% residual-vote rate) with another that happens to have a 0.4% residual-vote rate. What kind of justification would suffice, and how could the governmental body prove that the chosen course constituted the least restrictive means of meeting its goal?

The majority reads the above paragraphs as my faulting it for not articulating "a precise mathematical formula for determining when voting technology is constitutional and when it is not." Maj. Op. at 876. But my reference to a "coherent constitutional threshold" does not require a concrete numerical error rate of x% or y%. See id. Rather, I have criticized the majority for failing to provide a framework for determining when the numerical differences that are unavoidable in the election setting become constitutionally problematic. In responding to this aspect of my dissent, however, the majority does no more than reaffirm its conclusion that the certification and use of non-notice technology in this case is unconstitutional. The majority does not explain why these particular differences in voting technology are unconstitutional — a failing that exposes its approach as something other than the "reasoned analytical" one that it urges courts to employ "when confronted with difficult constitutional questions." Id. The next court facing the "difficult constitutional questions" inherent in a statistics-based equal protection challenge to state voting practices will therefore know only that, on these facts and these numbers, Ohio's use of varied voting methods violated the Equal Protection Clause, but will not know why.

Moreover, I am not the one who has made number-crunching a dominant factor in the equal protection analysis. The plaintiffs themselves did so when they brought a case whose central premise is that statistical differences in voter-error rates suffice to show a constitutional violation. Because these numbers are squarely at issue, I have provided the above hypotheticals in order to highlight the majority's failure to articulate a limiting principle beyond its assurances that it is not announcing a general rule and that "[f]uture claims of this sort will be evaluated with [legal, financial, and practical] factors in mind." Maj. Op. at 876.

The majority has responded by comparing apples to oranges, implying that the narrow "margin of decision" across all voters in recent presidential elections infuses the residual-vote rates at issue in the present case with added constitutional significance. Maj. Op. at 871 n. 18. But again the majority does not explain how these state-by-state margin-of-victory numbers affect, if at all, the equal protection analysis. I assume, for example, that the majority does not mean to suggest that the comparative error rates of voting machines would have to be less than the 0.009% margin-of-victory in Florida's 2000 presidential election in order to pass constitutional muster.

In the end, my colleagues attempt to disclaim their reliance on quantitative differences, describing "[a] judicially imposed mathematical formula for evaluating voting rights cases" as "purely arbitrary." Maj. Op. at 876. Such a formula, however, strikes me as no more arbitrary than the majority's declaration that the specific differential in error rates across voting districts in the present case violates the Equal Protection Clause.

I also believe that the majority fails to grasp that "equality" is not a one-way street. If what the Equal Protection Clause requires is that counties across Ohio utilize voting mechanisms that yield a substantially similar residual-vote rate, then the use of punch-card ballots is presumably permissible so long as every county uses punch cards. That this voting method leads to a relatively high percentage of discarded votes statewide should make no difference to the equal protection analysis, since every voter would face the same chance of casting a ballot that is later deemed not to have been properly marked. In other words, while the majority presumes that requiring substantial equivalency in residual-vote rates across districts will prompt local governments to "ratchet up" to notice equipment, those governments would also be free to turn the ratchet down to cheaper non-notice technology.

This all-or-nothing approach carries with it substantial practical concerns. As Professor Hasen observed, when one district adopts a new voting method designed to decrease the percentage of voter error, voters in districts that have not yet implemented the new method would have a colorable claim, under the majority's theory, that their constitutional rights have been violated. See Hasen, 29 Fla. St. U.L.Rev. at 401-02. The risk of a lawsuit each time improved technology emerges accordingly creates a perverse incentive for "freezing. . . voting mechanics at the current level of technology." Id. at 402.

Beyond the concern that the threat of lawsuits will halt innovation, the all-or-nothing approach increases the risk that technological setbacks could invalidate the electoral results on a statewide basis. Assume, for example, that one or more counties in Ohio want to experiment with Internet voting because such a mechanism might serve to decrease the error rate. If the trail-blazing counties are permitted to implement the new system on their own, security or logistical problems that call into question the validity of election results would affect only those counties, leaving intact the results from the rest of the state. On the other hand, if every county is required to simultaneously adopt a still imperfect system, a widespread technical failure could wipe out the entire state's election results. Instant equality, while fine in theory, does not account for these realities of the electoral system.

In light of the uncertain doctrinal foundations of and the practical problems posed by the majority's approach, I believe that the proper course is to review the challenged state and county election practices under the rational-basis standard. The Ninth Circuit followed this course in Weber v. Shelley, 347 F.3d 1101, 1106 (9th Cir.2003), a case in which the voter contested not the use of punch-card ballots, but rather the decision to abandon punch cards in favor of computerized voting. In rejecting the voter's claim, the Ninth Circuit applied the Burdick framework, concluded that the increased risk of fraud inherent in the new technology did not severely restrict the right to vote, and upheld the officials' decision as "a reasonable, politically neutral and non-discriminatory choice." Id. at 1107.

The same result should follow in the present case. Voters in counties that employ non-notice voting technology — whether punch-card ballots or central-count optical-scan equipment — do face a slightly elevated chance that a ballot they believe was properly marked will later be found invalid. In my view, however, this slight increase in the risk of improperly marking a ballot simply does not constitute a "severe restriction" on the right to vote. See Burdick, 504 U.S. at 434, 112 S. Ct. 2059. We should not, after all, lose sight of the fact that nearly 98% of the votes cast even in the punch-card counties were properly counted in the 2000 presidential election. The Secretary of State's decision to certify different types of voting equipment, and the counties' decision to use those mechanisms, are therefore subject to rational-basis review, where "the State's important regulatory interests are generally sufficient to justify" the government action. Id. at 434, 112 S. Ct. 2059 (citation and quotation marks omitted); accord Weber, 347 F.3d at 1106.

Like the district court, I believe that the government defendants have proffered sufficient reasons for their actions to survive the highly deferential rational-basis standard of review. See FCC v. Beach Communications, Inc., 508 U.S. 307, 313, 113 S. Ct. 2096, 124 L. Ed. 2d 211 (1993) (explaining that legislation is presumably valid and "must be upheld against equal protection challenge if there is any reasonably conceivable state of facts that could provide a rational basis for the" provision at issue). The majority dismisses the state's concerns about the cost and administrability of the new voting systems, Maj. Op. at 869-70, but does not even mention a primary justification given by the state and accepted by the district court — that of security. Concerns about potential "computer manipulation and fraudulent voting," according to one of the experts who testified, were substantial enough to delay the certification of DRE voting machines in the months leading up to the November 2004 elections. See Stewart, 356 F.Supp.2d at 803 n. 15. Recent events have revealed that these concerns were far from illusory. See Ann E. Marinow, Touch-Screen Voting Fallible, Ehrlich Says, Wash. Post, March 6, 2006, at B01 (discussing Maryland's decision to abandon its $90 million electronic-voting system because of concerns that the system was "vulnerable to tampering"). The state might also have worried about the all-or-nothing aspect of some DRE voting systems, which, unlike punch-card ballots, do not allow a voter who mistakenly casts an incomplete ballot to fill out a new ballot. See Stewart, 356 F.Supp.2d at 803 n. 14. These justifications suffice to show that the government defendants had a rational basis for choosing the challenged voting methods.

Furthermore, for the reasons that I mentioned briefly in Part I above, I cannot accept the majority's attempt to classify the plaintiffs' claims as covered by the older Supreme Court precedents on voter-qualification requirements and the weighting of votes. The majority cites to the Court's decision in Harper, 383 U.S. at 668, 86 S. Ct. 1079, insisting that "[t]he technology provided to a voter by the State, like that voter's wealth, has no relation to voting qualifications or the value of that vote." Maj. Op. at 870 (emphasis added). This statement conflates two different inquiries. Harper and other cases recognize that a state can determine voter qualifications absent invidious discrimination. See, e.g., Dunn v. Blumstein, 405 U.S. 330, 336, 92 S. Ct. 995, 31 L. Ed. 2d 274 (1972) (citing Harper and other cases that hold that "States have the power to impose voter qualifications, and to regulate access to the franchise in other ways"). Other decisions recognize that states generally have wide latitude in controlling the administration of elections. See, e.g., Storer v. Brown, 415 U.S. 724, 730, 94 S. Ct. 1274 39 L. Ed. 2d 714 (1974) ("[A]s a practical matter, there must be a substantial regulation of elections if they are to be fair and honest and if some sort of order, rather than chaos, is to accompany the democratic processes."). The plaintiffs' challenge in the present case goes to this latter ground — election administration — not voter qualifications. Harper, and its emphasis on the irrelevance of wealth as a factor for determining voter eligibility, is thus inapposite.

Reynolds also does not control the outcome of the present case. The use of non-notice voting technology — even assuming that such technology leads to appreciably higher rates of residual votes — is not the same as "[w]eighting the votes of citizens differently." Maj. Op. at 870 (quoting Reynolds, 377 U.S. at 563, 84 S. Ct. 1362). Again confusing the refusal to count an improperly marked ballot with a state's ascribing greater value to the votes of people living in one area rather than another, the majority asserts that voters in non-notice districts "have an unequal chance of their vote being counted, not as a result of any action on the part of the voter, but because of the different technology utilized." Maj. Op. at 870. I agree with the district court, however, that the improper marking of a ballot is primarily due to the voter's own actions, however unintentional those actions may have been. See Stewart, 356 F.Supp.2d at 803 ("[T]he voter has every opportunity to check the punch card ballot before submitting it to the election official at the polls and to be given a new ballot if a mistake is discovered."). So although the technology utilized has a statistical correlation to residual-vote rates, a person's failed attempted to cast a vote is primarily the result of human error. Some technologies admittedly account for such human error better than others. Until today, however, the Equal Protection Clause has not served as a check on human error in federal elections.

Technology that dilutes voting strength, the majority tells us, "is no less a [constitutional] violation than any other invidious method." Maj. Op. at 871. But the use of different voting mechanisms has little to do with the type of "invidious discrimination" that the Equal Protection Clause was designed to prevent. That is why I believe that many of the broad statements in the majority opinion, although eloquent, are ultimately hollow. E.g., Maj. Op. at 879-80 ("Violations of the Equal Protection Clause are no less deserving of protection because they are accomplished with a modern machine than with outdated prejudices."). However unpleasant the prospect, human error remains a part of the democratic process, and no constitutional rule is going to change that fact. See Powell v. Power, 436 F.2d 84, 88 (2d Cir. 1970) (holding that neither the Equal Protection and Due Process Clauses of the Fourteenth Amendment nor Article I, § 2 of the Constitution "guarantee against errors in the administration of an election"). What does have more potential to reduce the effect of human error on election results is incremental change of the kind that Congress enacted in the Help America Vote Act of 2002.

If the end result of the majority's holding is a requirement that all Ohio counties adopt some form of notice technology, then Congress has beaten my colleagues to the punch. Key provisions of HAVA impose detailed notice-technology requirements on states receiving federal funds for election administration, whether those states employ lever, optical-scan, or DRE voting systems. See 42 U.S.C. § 15481(a)(1)(A) (2005). HAVA stops short, however, of banning the use of paper and punch-card ballots, explicitly permitting such ballots to be used so long as the state establishes a voter education program to explain the possibility of voter error, and provides instructions on how to correct those errors prior to the casting of a ballot. Id. § 15481(a)(1)(B); see also id. § 15481(c)(2) (preserving a state's power to use paper ballots in federal elections). Punch-card voting systems thus survive HAVA even though, as the district court recognized, a primary "purpose of the Act was to provide federal funds to replace punch card voting systems." Stewart, 356 F.Supp.2d at 792 n. 1.

As one would expect, HAVA has become a focal point for election litigation, spawning a series of lawsuits in this court and the district courts within the circuit. See, e.g., Sandusky County Democratic Party v. Blackwell, 387 F.3d 565 (6th Cir.2004) (per curiam); White v. Blackwell, 409 F. Supp. 2d 919 (N.D.Ohio 2006); Bay County Democratic Party v. Land, 347 F. Supp. 2d 404 (E.D.Mich.2004). These suits generally seek enforcement of obligations recently imposed by Congress on those states that choose to accept federal funds for election administration. The present suit, in contrast, eschews the significant but balanced changes mandated by HAVA, and instead seeks a broad constitutional rule. But because HAVA is coextensive with the majority's constitutional rule in so far as both require that participating states like Ohio implement some form of a notice system, that constitutional rule is unnecessary. Ohio's compliance with HAVA will give the plaintiffs substantially all the relief that they claim to seek.

What concerns me about the majority's decision, therefore, is not the effect that the equal protection holding will have on election practices in Ohio in the upcoming November elections. Rather, I am troubled by two other aspects of the decision. The first is that the majority has inscribed its preferred principles of federal election law into the Constitution, thereby taking critical decisions out of the hands of government officials more familiar with the realities of the problems and the possible solutions. This course of action should be especially disfavored at a time when the democratic process — through HAVA and the discussions and debates engendered by that legislation — has responded with a national plan that still allows for variations among states and localities.

My second concern stems from the uncertain scope of the majority's equal protection holding, which I explored in Part III above. HAVA requires that states employ some type of a notice system, regardless of the voting equipment used. But, as I have demonstrated, the majority's holding might render even some notice systems unconstitutional if other notice systems are found to have a substantially lower residual vote-rate. That is to say, the majority's novel interpretation of the Equal Protection Clause sweeps far more broadly than does HAVA, and threatens to eliminate the ability of state and local governments to experiment with procedures and technology that maximize security, minimize cost, and make the voting process as inclusive as it should be. Cf. New State Ice Co. v. Liebmann, 285 U.S. 262, 311, 52 S. Ct. 371, 76 L. Ed. 747 (1932) (Brandeis, J., dissenting) ("There must be power in the states and the nation to remould, through experimentation, our . . . practices and institutions to meet changing social and economic needs."). I therefore cannot join the majority's decision to declare unconstitutional the use of different voting technologies across the state of Ohio.

Although the district court's analysis of the claim brought by the plaintiffs under the Voting Rights Act could have been more thorough, I am not convinced that the court reached an erroneous conclusion. Indeed, the majority recognizes the district court's finding that the residual-vote rate in rural counties with minimal percentages of African-American residents was higher than in the counties whose practices are challenged here. See Stewart, 356 F.Supp.2d at 820. These findings are fully consistent with the district court's view that the political process in the defendant counties was "equally open to participation by" African-American voters. See 42 U.S.C. § 1973(b). The majority nevertheless decides to remand the plaintiffs' claim for further findings and additional legal determinations, and holds that "the plaintiffs properly stated a claim under the Act . . . ." Maj. Op. at 879. This latter conclusion strikes me as both premature and inconsistent with the majority's decision to remand for further findings of fact.

I also disagree with the majority's instructions to the district court on remand. The majority purports "[t]o aid" the district court's analysis by citing to Johnson v. DeGrandy, 512 U.S. 997, 1019, 114 S. Ct. 2647, 129 L. Ed. 2d 775 (1994), a case that the majority says "reject[s] the argument that vote dilution in one part of the state can be remedied by another part of the state." Maj. Op. at 879 (emphasis added). DeGrandy, however, provides little guidance in the present case for two reasons. The first is that the plaintiffs here have alleged a vote denial claim, not a vote dilution claim like the one addressed in DeGrandy. See Stewart, 356 F.Supp.2d at 808 (S.D.Ohio 2004); see also Johnson v. Governor of the State of Fl., 405 F.3d 1214, 1227 n. 26 (11th Cir.) (en banc) (explaining the requirements for proving a claim of vote denial), cert. denied, ___ U.S. ___, 126 S. Ct. 650, 163 L. Ed. 2d 526 (2005). Second, DeGrandy is a reapportionment case that does not speak in terms so general as to cover differences in voting technology across counties.

I therefore dissent from the majority's decision to vacate and remand and from the imprecise instructions it has provided to the district court. But because the district court will have an opportunity on remand to supplement the factual record and to address the Voting Rights Act claim anew, I see no point in fully analyzing this claim on what is essentially an interlocutory basis. Indeed, as the en banc Ninth Circuit found under similar circumstances in Shelley II, the application of the Voting Rights Act to allegations like those made by the plaintiffs presents close, fact-sensitive questions. See 344 F.3d at 918-919. I will therefore explore those questions if and when the plaintiffs' claim under the Voting Rights Act returns to this court.

The expansive reading of Bush v. Gore by the Ninth Circuit panel lasted less then ten days before being vacated en banc, leaving today's decision as the only one from a federal court of appeals to invalidate state-election practices on the basis of Bush v. Gore's equal protection holding. Because I believe that the Supreme Court's precedents outside of Bush v. Gore permit the states significant leeway in the administration of elections, that the Constitution does not require local governments to employ voting technology that assures every voter a statistically identical chance of turning in a properly marked ballot, and that HAVA already mandates many of the changes that the majority has elevated to constitutional requirements, I would affirm the grant of summary judgment in favor of the defendants with regard to the claim under the Equal Protection Clause. With respect to the claim under the Voting Rights Act, I am not convinced that the district court erred either in its ultimate conclusion or in denying the plaintiffs' request to certify a class. I therefore respectfully dissent from the majority's contrary conclusions.