Stewart v. Blackwell

Full title: Effie STEWART, et al., Plaintiffs, v. J. Kenneth BLACKWELL, Secretary of…

Court: United States District Court, N.D. Ohio, Eastern Division

Date published: Dec 14, 2004

356 F. Supp. 2d 791 (N.D. Ohio 2004)

Opinion

No. 5:02 CV 2028.

December 14, 2004

MEMORANDUM OPINION

David Dowd Jr., District Judge

I. Introduction

The highly disputed presidential election in 2000, centered on the issue of which presidential candidate, Al Gore or George Bush, was entitled to the critical electoral vote for Florida, gave rise to litigation challenging the use of various forms of balloting procedures. This case involves the plaintiffs' challenge in Ohio to the use of punch card and "central-count" optical scanning technology.

The case was started in the year 2002. It became obvious that the challenge which championed Direct Record Electronic (DRE) technology as the preferred alternative technology could not be effectively resolved in time for the presidential election in 2004, given the certain appellate challenge to any decision by this Court. However, the plaintiffs successfully moved the Court to decide the claims as expeditiously as possible in view of the approaching municipal elections in Cincinnati (Hamilton County) in 2005. In the meantime, congressional legislation known as HAVA (Help America Vote Act) became effective on October 29, 2002. The Ohio Secretary of State, and a defendant in this case, Kenneth Blackwell has embarked on a campaign to replace both the punch card voting machines and optical scan central-count voting machines with DRE voting machines.¹

1 HAVA specifically provided that the purpose of the Act was to provide federal funds to replace punch card voting systems.

Believing that the plaintiffs are entitled to a judicial response, and in view of the on-going issues relating to various forms of voting technology, the Court scheduled and conducted a bench trial over a five-day period on July 26, 27, and 28, 2004; September 30, 2004; and October 1, 2004. At the conclusion of the bench trial on October 1, 2004, the parties agreed to a deadline of November 15, 2004 for the filing of post-trial briefs. The briefs have now been filed. In preparation for the bench trial, the parties entered into a comprehensive fact stipulation which is contained in Section III of this opinion.

The five day bench trial² featured the testimony of Martha Kropf, Dana Walch, Roy Saltman, Richard Engstrom, John Lott, Barabara Tuckerman, and by deposition, Dr. Herbert Asher.³

2 The transcript of the testimony for July 26, 27, and 28, 2004 is located at Doc. Nos. 257, 258, and 259. The concluding testimony on September 30, 2004 and October 1, 2004 can be found at Doc. Nos. 265 and 266. The Court also entertained "interim argument." Interim argument can be found at pages 316-327; pages 525-532; pages 781-785; and finally at pages 954-961 of the transcript.

3 Martha Kropf, whose extensive testimony is found at pages 57-125; 205-239; 844-870; and 876-953 is a professor of political science at the University of Missouri in Kansas City. She has a Ph. D. from American University in Washington D.C. She offered the opinion that the vast majority of "residual votes" cast in the 2000 presidential election were accidental as opposed to intentional. Her opinion rested on the results of post-election polls of voters conducted by National Elections Studies for the presidential elections from 1980-2000 and the Voter News Service for the 1992 presidential election.
Dana Walch, whose testimony is found at pages 131-204 and pages 728-781, is the Director of Election Reform in the Office of the Ohio Secretary of State and previously served as the Director of Elections in the same office. He offered extensive testimony on many issues dealing with various forms of election technology as well as statistical information regarding elections in the State of Ohio.
Roy Saltman, whose testimony is found at pages 246-316, is a consultant in election policy and technology. He enjoys a number of Masters' Degrees including such a degree in Public Administration from American University in Washington, D.C. He was employed at the National Institute of Standards and Technology from 1969 to 1996. He offered testimony about an article he authored in 1988 in which he opined that pre-scored punch cards should no longer be used because of the problems with chads in a recount situation. Mr. Saltman offered testimony on the continuum from voice votes, to paper ballots, to lever machines to computers. Since the contested presidential election of 2000 and because of his 1988 report, Mr. Saltman has been involved as an expert witness in litigation involving voting technology in Florida, California and Maryland. Saltman wrote an article entitled "The use of prescored punch card ballots should be ended" and it was received in evidence as Plaintiff's Exhibit 13. Mr. Saltman was the first witness to introduce the proposition that the prescored punch card ballot is fragile. (See page 268-269).
Mark Salling, whose testimony is found at pages 328-427, has directed a research program at in the College of Urban Affairs at Cleveland State University for 21 years and enjoys a Ph.D. from Kent State University. He manages a team of technicians and researchers who produce demographic analyses and data reports. His work in this case was presented in support of the plaintiffs' action claiming a violation of the Voting Rights Act by use of punch card voting machines in the defendant counties of Summit, Montgomery, and Hamilton.
Richard Engstrom, whose testimony is found at pages 429-564, is a professor at the University of New Orleans with a Ph. D. from the University of Kentucky in political science. He described his specialty as "primarily urban and minority politics and electoral systems." Dr. Engstrom prepared a report to estimate the racial differences in undervoting, overvoting in four counties in Ohio, namely Franklin County and three of the defendant counties, Hamilton, Montgomery and Summit. His report was received as plaintiffs' exhibit # 11. Dr. Engstrom also identified three separate methods of reviewing voting procedures including homogeneous precinct analysis, ecological regression and ecological inference (see page 445) and identified the purpose as determining the percentage of African-Americans that either undervoted or overvoted in the context of vote denial (See page 448)
John Lott, a witness for the defendants whose testimony is found at pages 533-686, holds three degrees from UCLA including a Ph. D. in economics and holds himself out as an economist. Dr. Lott asked for and received information on voting in five election contests in the years of 1992, 1996 and 2000 for the offices of President, United States Senate, United States House, and Ohio Senate and House races to determine how different types of voting mechanisms work. Dr. Lott found that, as one went down the ballot in studying the five offices, the nonvoted ballot rate was higher for other types of voting systems than it was for the punch card ballot. Dr. Lott summarized at page 570 as follows:

. . . but basically what you find here is the simple results that you got for the tables, and that is for the presidential race, the punchcards tend to do relatively poorly, but as you go down the ballot, they improve relative to other types of voting machines that you have there. And again, if you ask the question what type of voting machine produces the lowest number of nonvoted ballots across all five races, punchcards do extremely well. They do much better than the electronic machines that were being used and much better than the lever machines, and they are closely compatible to what you see for the optical scans.

When asked his opinion on whether punch card ballots produce more nonvoted ballots in Ohio than other types of voting devices, he stated, at page 574, that:

if you look across all the races, for the 2000 — for the 1992, the `96 and the 2000 elections, punchcards produce fewer nonvoted ballots than either electronic voting machines or lever machines, and virtually the same as optical scan machines.

Barbara Tuckerman, whose testimony is found at pages 687-727, is the Director of Elections for defendant Sandusky County and she described the use of optical scan voting machines in Sandusky County.
Dr. Herb Asher, called as a witness for the defendants, offered testimony by way of a previously taken deposition. Only part of his deposition was offered by agreement of the parties. The testimony was read into the record and can be found at pages 789 to 844. Dr. Asher is a professor emeritus at Ohio State University. He presently teaches courses on politics, campaign politics and presidential campaigns. He described his interest in evaluating elections in Ohio from a standpoint of fall-off and voter failure to complete a ballot or over-vote in the context of socio-economic factors for the particular precinct or region. He referred to several papers and studies touching on these subjects with specific reference to Ohio.

Based on the testimony and exhibits received into the record, the Court makes additional fact findings as set forth in Section IV of this opinion.

Based on the stipulated facts and the additional fact findings of the Court, the Court finds that judgment should and will be entered on behalf of the defendants. The Court's analysis is set forth in Section V of this opinion.

II. A Summary of the Positions of the Parties

A. The Plaintiffs' Position

Plaintiffs seek declaratory and injunctive relief from Defendants' certification and use of current balloting systems in four Ohio counties. Specifically, plaintiffs allege that punch card voting and "centralcount" optical scanning devices violate their rights under the Due Process Clause, the Equal Protection Clause and (the African-American plaintiffs) their rights under § 2 of the Voting Rights Act of 1965, 42 U.S.C. § 1973.

Plaintiffs assert that their Fourteenth Amendment Equal Protection rights are violated by Ohio's system for voting technology selection which allows counties to choose different types of voting devices. More specifically they argue based on Bush v. Gore, 531 U.S. 98, 105, 121 S.Ct. 525, 148 L.Ed.2d 388 (2000), that their respective counties' use of non-notice voting technologies does not afford their vote with the same weight and dignity of the votes cast in Ohio counties that currently employ notice voting technologies. According to plaintiffs, defendants are illegally favoring some voters over others on the basis of residency by certifying and employing notice voting systems in some counties and non-notice systems in others. The effect of this supposed dual system is to dilute the voting strength of certain Ohioans.

Plaintiffs also claim that defendants are violating their Fourteenth Amendment Due Process Right to Vote. They argue that the use of error prone equipment, such as the punch card ballot, arbitrarily deprives them of their right to vote. It does so by subjecting them to a significantly greater risk that their votes will not be counted. Plaintiffs further contend that this election practice is subject to strict scrutiny because it impacts the right to vote. Since there is no legitimate government interest that justifies this system, plaintiffs argue, it violates their Due Process rights.

To prove their Fourteenth Amendment claims, plaintiffs employ statistical data showing that punch card and central-count optical scan machines had higher residual vote⁴ rates than other systems used in Ohio in the presidential elections between 1992 and 2000. They note that the data of defendants' expert Dr. Lott indicates a residual vote rate of 2.4% for punch cards, 1.0% for DRE, 1.4% for lever machines, and 2.0% for optical scan. Plaintiffs' expert Dr. Kropf produced similar results: 2.29% for punch cards, 0.94% for DRE, 1.04% for lever, and 1.15% for precinctcount optical scan. According to plaintiffs, these statistics show that punch cards and central-count optical scan violate Due Process rights because they are substantially less accurate and violate Equal Protection rights because other Ohio counties are operating the more accurate systems.

4 A "residual vote" is a ballot on which the voter has either overvoted or undervoted. An "overvote" occurs when a voter registers more than the prescribed number of choices in a particular race. An "undervote" occurs when a voter either does not mark a ballot for a particular contest or votes for fewer than the allotted number of choices.

The African — American plaintiffs allege that their respective county's use of punch card ballots violates § 2 of the Voting Rights Act by denying them the right to vote.⁵ Section 2 of the Voting Rights Act extends to protect the proper counting and totaling of votes cast. They argue that the intra-county disparate impact punch card ballots have on the likelihood that an African-American's vote will not be counted amounts to a denial of the right to vote under the Voting Rights Act. According to them, the Court must only compare intra-county vote rates in analyzing their Voting Rights Act claim.

5 Plaintiffs do not claim vote dilution but only vote denial as a violation of the Voting Rights Act.

In support of their Voting Rights Act claim, the African-American plaintiffs claim (1) that the defendants' selection of punch card ballots as a voting device constitutes a state action; (2) that African-American voters suffer far higher rates of ballot rejection than do white voters in Hamilton, Montgomery, and Summit Counties, and (3) that there is a causal relationship between the defendants' selection of punch cards and African-American voters' higher rates of ballot rejection.

To prove the intra-county ballot rejection disparity between African-American voters and white voters, plaintiffs point to the testimony of Dr. Salling and the maps he created. These maps show that in Summit and Franklin counties there is a strong correlation between the precincts with high numbers of residual votes and those with a high percentage of African-Americans. Plaintiffs also present several tables summarizing Dr. Engstrom's analysis to overcome the ecological fallacy.⁶

6 Attributing certain qualities or behavior to individual voters based on aggregate data is known as "the ecological fallacy." It is present here because Dr. Sailing's maps do not show whether it is the African-Americans or the whites in those precincts that are casting the invalid ballots.
Dr. Engstrom employed three different analyses to overcome the ecological fallacy: ecological inference, ecological regression, and homogeneous precinct.

Plaintiffs rely on data from Franklin County to support the third prong of their analysis. Franklin County, whose voters use DRE machines, has a negligible residual vote rate for whites and below 1% for African-Americans. They assert that Franklin County's use of DRE machines and lower racial disparity in residual vote rates indicates a causal relationship between the use of punch card ballots and the higher racial disparity in residual vote rates found in the defendant counties. Because these disparities do not arise from intentional conduct of the voter, plaintiffs assert that they have established a vote denial claim under the § 2 of the Voting Rights Act.

B. The Position of the Secretary of State

The Secretary of State contends that plaintiffs lack standing because they have not suffered an injury in fact and that this matter is moot because H.B. 262 and Ohio's efforts to comply with HAVA adequately address plaintiffs' complaints. Regardless of whether the claim is moot, he also claims that the African-American plaintiffs have not proven a violation of the Voting Rights Act. In support of this claim, the Secretary of State directs the Court's attention to plaintiffs' stipulation that they have not been denied access to the polls and that no evidence was introduced suggesting a denial of the right to vote. He also points to the fact that plaintiffs failed to show that punch card ballots are given exclusively to African-Americans or disproportionately used in African-American precincts. Further, he suggests that plaintiffs' failure to establish any of the Senate factors shows they have not met their evidentiary burden. Additionally, the Secretary of State calls into question plaintiffs' statistical data because it relies on unreliable exit polls.

The Secretary of State, in order to further defeat plaintiffs' Voting Rights Act claim, posits that the statistical data does not prove a great racial disparity in undervoting through the use of punch card ballots. First, he attempts to discredit plaintiffs' statistical data by pointing to the fact that it is based on exit polls that were taken weeks after the election. He then references Dr. Engstrom's testimony that (1) whites in Summit County undervoted at a rate higher than African Americans in Hamilton County, (2) African-Americans in Hamilton County undervoted at a rate lower than, equal to, and higher than African-Americans in Franklin County, and (3) the racial disparity in undervoting in Franklin County (0.81%) was nearly four times higher than in Hamilton County (0.22%). Lastly, he presents the testimony of Dr. Lott and Dr. Asher. Dr. Lott concluded that across years and offices punch card ballots tend to produce fewer undervotes than DRE or lever machines and were nearly identical to optical scan systems. Dr. Asher further concluded that the counties with the highest nonvoted ballot rate in the 2000 Presidential election could not possibly have been caused by African-Americans.

The Secretary of State also argues that plaintiffs have not shown a violation of the Fourteenth Amendment. He first contends that regional differences in voting technology choices within a state do not amount to an Equal Protection violation, citing Bush, 531 U.S. at 134, 121 S.Ct. 525 (Souter, J., dissenting). He then argues that the current voting technology is not rendered unconstitutional merely because another system is more accurate at recording votes.

C. The Positions of Hamilton, Montgomery, and Summit Counties Regarding Plaintiffs' Voting Rights Act Claims

Hamilton, Montgomery, and Summit Counties have Voting Rights Act claims asserted against them. Initially they question the African-American plaintiffs' standing to assert a Voting Rights Act claim because none of them claim that they were denied access to the polls or that they know their vote was not counted. Montgomery County further questions the standing of Professor Vernellia Randall, an African-American who is registered to vote in Montgomery County, because she did not vote in the 2000 presidential election. Montgomery County additionally suggests that this matter is a non-justiciable political question and moot.

Alternatively, these counties argue that they do not violate the Voting Rights Act because an invalid vote does not equate to a denial of the right to vote. They refer the Court to historical instances of vote denial such as poll taxes, grandfather clauses, and literacy tests, and argue that plaintiffs' assertion of a higher rate of invalid votes does not amount to a vote denial under the Voting Rights Act. Hamilton County points to the varying rates of overvotes for African-Americans in recent local elections to suggest that it is not punch cards but something else that causes the overvotes. Specifically, it cites to the fact that the overvote rate for African-Americans was almost non-existent in the 2001 Cincinnati mayoral election but was comparatively higher in the "vote for nine" 1999 race for Cincinnati City Council and the 2000 Presidential election.

D. The Positions of the County Defendants Regarding Plaintiffs' Fourteenth Amendment Claims

Plaintiffs also allege that each of the four County Defendants violates the Fourteenth Amendment Equal Protection Rights and Due Process Rights of its citizens. These counties contend that they are not violating the plaintiffs Equal Protection rights because all of the citizens within each county vote using the same voting technology. They further argue that the use of different voting technologies in other counties does not result in their violating plaintiffs' Equal Protection rights because (1) these counties have no power over the choices made by other counties' boards of elections and (2) there is no requirement that all counties use the same voting technology.

Montgomery County also contends that its use of punch card ballots does not violate Due Process rights because its residual vote rate is not unacceptable. A residual vote rate is not deemed unacceptable until it reaches 3%.⁷ Montgomery County's residual vote rate in the 2000 presidential election was 2.78% and, therefore, not unacceptable. Sandusky County asserts that it too is not violating plaintiffs' Due Process rights by using a centralcount optical scan system, as opposed to precinct-count optical scan, because there are rational bases for employing the central-count system: cost, space, time, and the polling place staff's lack of familiarity with the machines. Sandusky County further argues that it does not violate Due Process rights because as long as a voter understands and follows the directions given to him or her, every intentional vote cast is counted, regardless of where the vote is counted.

7 According to the Cal TechM.I.T. post 2000 election study, a 0-1% residual vote rate is characterized as good, 1-2% is characterized as acceptable, 2-3% is characterized as worrisome, and over 3% is characterized as unacceptable.

III. The Stipulated Facts and the Court's Supplemental Fact Findings.

Counsel for the parties engaged in an extensive stipulation of fact including 98 separate paragraphs of stipulated facts. That stipulation is attached as Appendix I. Following the bench trial the Court made additional fact findings in paragraphs 99 through 146, and those additional facts findings are attached as Appendix II.

IV. Summary of Existing Jurisprudence

In the wake of the 2000 presidential election, several suits, including the present action, were filed challenging the use of punch card ballots as violative of § 2 of the Voting Rights Act, the Equal Protection Clause, and/or the Due Process Clause. Additionally, voters filed suits leading up to the 2004 presidential election challenging the adequacy of the voting systems that replaced the punch card ballots. These other cases provide a background for the climate in which this matter is decided.

In Black v. McGuffage, 209 F.Supp.2d 889, 902 (N.D.Ill. 2002), the Court denied the defendants' motion to dismiss with respect to all counts of the plaintiffs' complaint except the privileges and immunities count. Plaintiffs there sought an injunction prohibiting the use of punch card voting systems and other systems that lack effective error notification. Id. at 894. The plaintiffs alleged that those counties that employed punch cards or optical scan without error notification experienced higher residual vote rates than those counties using optical scan with error notification. Id. Furthermore, while a majority of Illinois counties used punch cards, the plaintiffs contended that those counties which used punch cards had larger populations of minorities than counties using other systems. Id. As a result, the plaintiffs argued that punch cards had a disparate impact on minority voters. Id. The plaintiffs, therefore, alleged that the use of punch cards and other voting systems which lacked error notification violated § 2 of the Voting Rights Act and the Equal Protection and Due Process clauses of the Fourteenth Amendment. Id.

The court in Black concluded that the plaintiffs had sufficiently alleged a claim under § 2 of the Voting Rights Act because they met the statutory requirements. Id. at 896. It stated that two elements are necessary to show a violation of § 2: "(1) the use of an electoral `standard, practice, or procedure,' and (2) a resulting diminution of the opportunity to African American and Latino voters `to participate in the political process and to elect representatives of their choice.'" Id. Accepting the plaintiffs' facts as true, the court concluded that the plaintiffs as voters residing in predominantly Latino and African American precincts that utilized punch cards "b[ore] a greater risk that their votes [would] not be counted than [did] other voters." Id. at 897. Because this could significantly diminish the plaintiffs' participation in the political process, the court found that the plaintiffs had sufficiently alleged a violation of § 2. Id.

Likewise, the court held that if the plaintiffs were able to prove that as a result of different voting technologies "voters in some counties [were] 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," they would establish a violation of the Equal Protection Clause. Id. at 899. In making this determination, the court reasoned that if the different counties' choices of different voting systems resulted in significantly different probabilities of their respective citizens having their votes counted, then one person's vote would be valued over another's vote. Id. "However, once the State has endowed voting rights to its citizens . . . `[it] may not, by later arbitrary and disparate treatment, value one person's vote over that of another.'" Id. at 898 (quoting Bush, 531 U.S. at 104-105, 121 S.Ct. 525). Because this was precisely what the plaintiffs had alleged, they had sufficiently stated a claim for violation of Equal Protection. Id. at 898-899. The court further noted that the plaintiffs allegation of disparate impact on minority groups created "cause for serious concern." Id. at 899.

Lastly, the court denied the motion to dismiss with respect to the plaintiffs' substantive Due Process claim. Id. at 901. The plaintiffs alleged that the vote counting procedure was flawed because it irrationally allowed local election officials to assign greater importance to the votes cast by a portion of the electorate through their choice of vote counting procedures. Id. The court held that "a law that allows significantly inaccurate systems of vote counting to be imposed upon some portions of the electorate and not others without any rational basis runs afoul of the Due Process clause of the U.S. Constitution." Id. Because that is what the plaintiffs had alleged, the defendants' motion to dismiss was denied. Id. at 901-902.⁸

8 Black v. McGuffage, was subsequently settled by the parties.

In Common Cause v. Jones, 213 F.Supp.2d 1106 (C.D.Cal. 2001), the court denied the defendant's motion for judgment on the pleadings against the plaintiffs' Fourteenth Amendment and Voting Rights Act claims. There, the plaintiffs alleged that, because punch card ballots were less reliable than other voting systems certified by the Secretary of State, individuals living in counties where punch cards were used were denied the right to vote protected by the Fourteenth Amendment. Id. at 1108. Because the counties which chose to use punch card ballots had high racial minority populations, the plaintiffs also alleged that this "vote denial" violated the Voting Rights Act. Id.

The court first held that the plaintiff had alleged facts "indicating that the Secretary of State's permission to counties to adopt either punch-card voting procedures or more reliable voting procedures was unreasonable and discriminatory." Id. at 1109. Thus, the defendant was not entitled to a judgment on the pleadings on the Fourteenth Amendment claim. Id. at 1109-1110. The court then stated that the test from Thornburg v. Gingles, 478 U.S. 30, 48-51, 106 S.Ct. 2752, 92 L.Ed.2d 25 (1986)⁹, did not apply to the Voting Rights Act claim because it was a vote denial claim. Common Cause, 213 F.Supp.2d at 1110. The plaintiffs failure to allege facts necessary to establish the Thornburg elements, therefore, did not warrant a grant of judgment on the pleadings, and the court accordingly denied the motion. Id. ¹⁰

9 Thornburg requires a showing of three things: "1) the minority group is `sufficiently large and geographically compact to constitute a majority in a single member district'; 2) the minority group is `politically cohesive'; and 3) `the white majority votes sufficiently as a bloc to enable it usually to defeat the minority's preferred candidate.'" Common Cause, 213 F.Supp.2d at 1110 (quoting Thornburg, 478 U.S. at 48-51, 106 S.Ct. 2752).

10 The issues in this case were subsequently rendered moot because the Secretary decertified punch card ballots. See Common Cause v. Jones, 213 F.Supp.2d 1110, 1113 (C.D.Cal. 2002).

The Ninth Circuit, sitting en banc, reversed a circuit panel's decision and affirmed the district court's denial of a preliminary injunction in Southwest Voter Registration Education Project v. Shelley, 344 F.3d 914, 916-920 (9th Cir. 2003). There, the plaintiffs challenged the use of punch card ballots in some California counties in the gubernatorial recall election of 2003. Id. at 916. They argued that the use of punch card ballots violated the Equal Protection Clause and § 2 of the Voting Rights Act. Id. In deciding the Equal Protection claim, the court noted that it had not previously had occasion to consider this Equal Protection issue and that in Bush, 531 U.S. at 109, 121 S.Ct. 525, the Supreme Court stated that it was not deciding whether local entities may develop different systems for implementing elections. Id. at 918. The court, therefore, could not conclude that the district court had abused its discretion in holding that the plaintiffs had failed to establish a clear probability of success on the merits. Id

Furthermore, while the court concluded that the plaintiffs had made a stronger showing of success on their Voting Rights Act claim, the court denied the injunction because it could not say they had shown a strong likelihood of success. Id. at 918-919. The plaintiffs alleged two things (1) that minority voters disproportionately lived in punch card counties and (2) within those counties, punch card machines discarded minority votes at a higher rate. Id. at 918. Because there was significant dispute as to the degree and significance of any racial disparity, the court could not say that plaintiffs had shown a strong likelihood of success on the merits. Id. at 918-919. The court then determined that the district court did not abuse its discretion in concluding that any hardship the plaintiffs would suffer was not outweighed by the interests of the state and its citizens in continuing the election. Id. at 920. Therefore, the preliminary injunction was denied. Id.

In Weber v. Shelley, 347 F.3d 1101, 1103 (9th Cir. 2003), the court affirmed the district court's grant of summary judgment to state and county officials against a voter who was challenging the lack of a voter verified audit trail. The voter there alleged that the DRE machine's lack of a voter verified audit trail resulted in a violation of her Equal Protection and substantive Due Process rights because it rendered the DRE more prone to fraud. Id.

The court first determined that because the lack of a voter verified audit trail did not severely affect the right to vote, it was not subject to strict scrutiny but only rational basis. Id. at 1106. The court then stated that "it is the job of democraticallyelected representatives to weigh the pros and cons of various balloting systems. So long as their choice is reasonable and neutral it is free from judicial second guessing." Id. at 1107. Because the state and the county had a reasonable, politically neutral, and non-discriminatory basis in choosing the voting system, its choice survived the rational basis test. Id.

In Wexler v. Lepore, 342 F.Supp.2d 1097 (S.D.Fla. 2004), after a bench trial, the court held that Florida's use of different voting systems did not violate the Equal Protection clause. Florida requires a manual recount of ballots if the margin of victory is one-quarter of one percent or less. Id. at 1099-1100. Currently only optical scan and DRE voting systems are certified for use in Florida. Id. The plaintiffs in Wexler argued that because the DRE machine did not produce a paper trail, their Equal Protection rights would be violated if a manual recount ensued. Id. at 1106. While optical scan ballots could be reviewed and interpreted for stray marks, there was no basis for interpretation of a DRE ballot image. Id. at 1106. The court noted that the problem in Bush v. Gore was a lack of uniform standards for determining voter intent with respect to each type of machine, which resulted in disparate standards being applied to identical types of ballots. Id. at 1106-07 (discussing Bush, 531 U.S. 98, 121 S.Ct. 525, 148 L.Ed.2d 388). The court, therefore, held that the state had satisfied the Equal Protection requirements because it had established uniform manual recount standards and procedures to be used throughout the state for each system. Id. at 1108.

The rules promulgated pursuant to the amended statutes comply with the requirements established by Bush v. Gore. Defendants have prescribed uniform, nondifferential standards for what constitutes a legal vote under each certified voting system, and have established procedures for conducting a manual recount of overvotes and undervotes in the entire geographic region.

Id. at 1108.

V. The Court's Ruling

A. The Setting for the Court's Decision

The Court views this case as one of extreme importance as it questions the validity of election machinery and raises the question of the extent to which the judiciary should be involved in issues relating to the integrity of the voting process.¹¹

11 It remains for others to decide whether this Court's experience prior to becoming a United States District Court Judge in 1982 has added or detracted from its management of this case. However, it is difficult for the Court to divorce itself from its past experience. Unlike most federal judges, this branch of the court has been personally involved as a candidate in many elections beginning in 1959 and continuing through 1980. Those elections involved municipal, county, regional, and statewide offices. Moreover, this branch of the court, for over 14 years, was involved in providing legal services to the Stark County Board of Elections. Finally, this branch of the court has cast ballots using punch card voting technology for more than 25 years. In that capacity, this branch of the court votes twice a year; first in the relevant primary and secondly in the general election. Consequently, this court votes eight times in every four-year cycle. However, many voters only vote once every four years and in the general election for the office of president.

Since the 2000 elections, to this Court's understanding, there has been no other case tried on the merits as in this case. It is the Court's view that the presentation of the testimony of the experts, coupled with the parties' stipulated facts as set forth in Appendix No. I and the Court's additional fact findings as set forth in Appendix No. II, provide an appellate court the opportunity to write on the issues raised by the plaintiffs and the defenses provided by the defendants in a comprehensive fashion and certainly the opportunity to disagree with this Court's analysis.

The defendants have argued a lack of standing on the part of the plaintiffs to obtain the ruling sought, that is, that the use of punch card voting technology violates either Due Process or fails to provide equal treatment and additionally that the use of the punch card voting technology in Ohio violates the Voting Rights Act. The defendant Secretary of State has also argued that the controversy is moot because of his intention to have DRE voting technology in place for Ohio by the general election in 2005.

The court 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 court also declines to dismiss the case on the issue of mootness, but rather will decide the case on the merits. It is entirely possible that by the time the case is before the Sixth Circuit, it will be moot or, alternatively, the Sixth Circuit may conclude, as argued by the defendants, that the plaintiffs lack standing to pursue the constitutional claims or the Voting Rights Act claim. By the same token, if the Sixth Circuit should come to a conclusion on the merits of the controversy, it is this Court's view that the extensive fact findings, both by stipulation and by reason of the Court's additional fact findings, provide sufficient material for a thorough appellate review.

B. The Court's Post Trial Preliminary Observations ¹²

12 The deadline for filing post trial briefs, as agreed to by counsel, was set at 45 days after the conclusion of the bench trial. As a consequence of the delay, the Court, while its memory was fresh and without the benefit of a transcript of the testimony or post trial briefs, prepared its own preliminary assessment of the testimony over the five day period and characterized it as "Post Trial Preliminary Observations." In the belief that the observations were and are pertinent, the Court repeats them as an integral part of the decisionmaking process in this case.

1. There is a continuum in voting practices across the ages from voice vote to paper vote to lever voting to computer voting. The struggle today continues with efforts to improve on computer voting.

2. The trend away from paper balloting seems to be motivated by a desire for instant results on election night rather than waiting for the laborious task of counting paper ballots.

3. The common factors that accompany a higher residual vote rate are the extent of the education and income level of the voters. The lower the educational level and the lower the income level, the higher the residual vote is in comparison with other counties demonstrating a higher educational level and income level. Those factors apply without regard to race. See Fact Findings 125-134.

4. Plaintiffs' Exhibit 35 reflects the residual vote rate in the presidential race in the year 2000. The higher percentage of residual votes using punch card voting technology when compared to other technologies, is, in the opinion of plaintiffs' experts, the result of accidental rather than intentional voting.

5. The plaintiffs' case rests on the initial premise that a high percentage of the residual votes in the Ohio 2000 presidential election using punch card voting technology is accidental as opposed to intentional.

6. The plaintiffs contend that the punch card technology is fatally flawed because it is a non-notice system.

7. The plaintiffs contend that because the incidence of residual votes using punch card technology is higher than with other voting technologies in Ohio, that incidence coupled with non-notice constitute constitutional violations, either under a Due Process or Equal Protection analysis and require federal judicial intervention and a remedy ordering that punch card voting machines no longer be used in Ohio.

8. The DRE voting technology, advanced by the plaintiffs as the appropriate constitutional alternative to punch card voting technology, also has its limitations. First, the use of the technology is dependent upon electrical power.¹³ Second, the technology is not available for absentee balloting in all situations. Third, there remains the problem of "premature voting."¹⁴ Fourth, there is at least the claim that the DRE voting technology is subject to computer manipulation.¹⁵

13 One of the plaintiffs' experts, while acknowledging the need for power in the use of DRE voting technology, opined that paper ballots should be available as an alternative in the event of a power outage. However, Dana Walch testified that batteries would be available for DRE equipment in the event of a power loss, and paper ballots would be an unnecessary redundancy. See Walch testimony, pg. 746-747.

14 The DRE voting technology requires the voter to push a button which casts the entire ballot. If the voter mistakenly pushes that button before the voter has finished voting, the voter will not be provided a new ballot as in the case of the voter using punch card technology. See Walch testimony, pg. 756-757.

15 The delay in supplying DRE voting technology for the 2004 election arose from the claim of computer experts that the proposed DRE voting technology approved by the Secretary of State had serious flaws that would have permitted computer manipulation and fraudulent voting. As of September 30, 2004, the Secretary of State had not approved any vendor providing DRE voting technology for use in Ohio. See Dana Walch testimony, pg. 732.

9. The plaintiffs' analysis as to the election results in Ohio is limited to the presidential election in 2000 and only includes four counties. The study of Dr. Lott, defendants' expert, makes a strong case for the proposition that punch card voting technology fares quite well in comparison to other technologies when considering drop-off or residual vote in elections beyond that of the presidency for the years 1992, 1996, and 2000.

10. The belief that the system of choice should be a "notice system" takes the view that the punch card ballot system is "nonnotice" because the experts in the field who promote the use of "notice" systems have chosen to so identify the punch card system as "non-notice," even though the 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.

11. 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.¹⁶

16 See Plaintiffs' Exhibit 13, pg. 1, the Report of Roy G. Saltman, in which he outlined recount problems using punch card voting technology. Specifically, the report states in part as follows:

One major defect of PPC voting systems is the presence of "hanging" chad on the backs of ballot cards following the voters' use of them in voting. . . . Chad retained on the cards result of their incomplete separation in the card-punching process carried out in voting. Voters are instructed to remove chad after voting, but many fail to do so. During the counting process, handing chad may be pressed back into the card and alter the voter's intent. In addition, some chad not intended to be removed may be removed unintentionally in the counting process due to excessive handling or manipulation. Furthermore, some chad may only be indented or pin-pricked in the voting process and not sufficiently separated to indicate a vote to an electro-mechanical card reader. Again, the voter's intent is altered in the counting. Ballots recounted by machine may show different counts because of chad, bringing into question the entire election process in a very close contest. (Emphasis added.)

12. The optical scan central-count system in place in Sandusky County suffers from the problem of uncertain markings on the ballot as indicated in the testimony of the Board of Elections Official from Sandusky County.¹⁷

17 The testimony of Barbara Tuckerman, Director of the Board of Elections for Sandusky County (pages 696-710), indicated that when the four members of the Sandusky County Board of Elections encountered an optical scan ballot with ambiguous markings, the four board members would caucus and determine the intention of the voter. Then the initial ballot would be destroyed and a new ballot containing the consensus opinion of the board of elections would be substituted and then counted. Obviously such a process provides no basis for judicial review in a case challenging the outcome of a recount.

13. The notice system envisioned by House Bill 262 adopted by the Ohio Legislature carries with it no printing of a receipt accurately recording the voter's choices and because of the belief that such a paper trail might result in pressure on the voter to show the voter's choices to the employer, spouse, parent, union, etc.¹⁸

18 The Ohio General Assembly adopted Substitute House Bill 262. A critical component of Substitute House Bill 262 is Ohio R.C. § 3506.01(H), which provides:

(H) "Voter verified paper audit trail" means a physical paper printout on which the voter's ballot choices, as registered by a direct recording electronic voting machine, are recorded. The voter shall be permitted to visually or audibly inspect the contents of the physical paper printout. The physical paper printout shall be securely retained at the polling place until the close of the polls on the date of the election; the secretary of state shall adopt rules under Chapter 119.[sic] of the Revised Code specifying the manner of storing the physical paper printout at the polling place. After the physical paper printout is produced, but before the voter's ballot is recorded, the voter shall have an opportunity to accept or reject the contents of the printout as matching the voter's ballot choices. If a voter rejects the contents of the physical paper printout, the system that produces the voter verified paper audit trail shall invalidate the printout and permit the voter to recast the voter's ballot. On and after the first federal election that occurs after January 1, 2006, unless required sooner by the Help America Vote Act of 2002, any system that produces a voter verified paper audit trail shall be accessible to disabled voters, including visually impaired voters, in the same manner as the direct recording electronic voting machine that produces it.

C. The Court's Ruling

The right to vote is central to our freedoms. Wesberry v. Sanders, 376 U.S. 1, 84 S.Ct. 526, 11 L.Ed.2d 481 (1964). The public's confidence in the integrity of the vote tabulation is absolutely essential to the respect we, as Americans, accord our elected officials.¹⁹ The governmental control of the voting process is bestowed primarily on state government. In turn, in Ohio, the voting process is supervised by the elected Secretary of State, in this case the defendant Kenneth Blackwell. But the Ohio Legislature has granted primary control and management of elections, local, state and federal, to the 88 County Boards of Elections.²⁰

19 The Court is of the view that rational basis is the level of scrutiny applicable to Ohio's use of punch card voting technology. See, e.g., Mixon v. State of Ohio, 193 F.3d 389 (6th Cir. 1999). While the right to vote is of great importance, states still enjoy the opportunity to put limits on the right to vote such as age, citizenship, requirement to register, restriction on votes by felons, and the hours that the polls are open. In this Court's view a rational basis level of scrutiny should be applied as opposed to plaintiffs' contention that the defendants' use of punch card voting machines should be subjected to strict scrutiny. However, if the Court were to apply strict scrutiny, the Court's ruling would be the same.

20 Title 35 of the Ohio Revised Code is entitled "Elections" and contains 15 separate chapters. The Ohio system for elections delineates the Secretary of State as the chief elections officer (§ 3501.04). However, the conduct of elections is controlled primarily by the Board of Elections in each of the 88 counties. The duties of the Board of Elections are set forth in § 3501.11.

The 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. Voting has been on a continuum in this country over two centuries. First, only white males enjoyed the privilege. Then, suffrage was extended to non-whites, but only after a bitter Civil War. Women did not obtain the right to vote until after the conclusion of World War I. It was not until the aftermath of the assassination of President John F. Kennedy in 1963 that African-Americans had a realistic opportunity to vote in certain areas in the country.

Voting has gone from oral votes, to paper ballots, to lever machines and now, in the era of computers, to the use of computer technology. The determination of the applicable voting process has always been focused in the legislative branch of the government. See Weber v. Shelley, 347 F.3d at 1105-07. In the Court's view, subject to constitutional amendment, that is where the determination should remain.

Turning now to the issues at hand, the Court finds that the plaintiffs have failed to make a case for judicial intervention with respect to the challenged voting technologies in this case, i.e., the punch card voting technology in use in Hamilton, Montgomery and Summit Counties and the central-count optical scan technique in use in Sandusky County.

While it is true that the percentage of residual or nonvoted ballots in the 2000 presidential election ran slightly higher in counties using punch card technology, that fact standing alone is insufficient to declare the use of the system unconstitutional.²¹ Moreover, the highest frequency in Ohio of residual voting bears a direct relationship to economic and educational factors, negating the Voting Rights Act claim against the three counties Hamilton, Montgomery, and Summit.

21 See Appendix III, which is Plaintiffs' Exhibit 35, setting forth the residual vote on a county-by-county basis for each of the voting technologies employed in the year 2000 presidential race.

In arriving at these conclusions, the Court accepts for the purpose of these conclusions, standing on the part of the plaintiffs, and makes the following pivotal findings.

1. The use of the punch card voting technology is neither confusing nor difficult to operate.

2. The testimony of Dr. Kropf, accepting as accurate the data from the NES and VNS studies ( see Fact Finding No. 100), leads to the conclusion that use of the punch card voting technology results in some small fraction of voters mistakenly failing to cast a ballot for the office of president in the election in the year 2000. The analysis follows.

(a) Dr. Kropf's report, Exhibit # 1, concludes with the following opinion:

17. Thus, the survey evidence we analyze indicates that differences across racial groups in intentional undervoting are insignificant, controlling for other factors and differences associated with income, while statistically significant, are relatively small. This evidence suggests that accidental undervoting and overvoting account for most of the invalidated presidential ballots in poor and minority precincts. As indicated before, one of the most likely sources of accidental undervotes is voting equipment. (Emphasis added)

(b) Earlier in her report, Dr. Kropf discussed Incidence of Intentional Undervoting in the following paragraphs:

6. Invalidated votes occur as the result of undervotes (where voters intentionally or unintentionally record no selection) or overvotes (where voters select too many candidates, thus spoiling the ballot). Invalidated votes in the presidential contest may occur for several reasons. Extensive empirical evidence has found that accidental undervotes and overvotes may occur because of faulty voting equipment, as alleged in this case. However, some of the invalidated votes occur when voters intentionally do not cast a vote in a particular race. Voters may cast intentional undervotes in presidential contests for several reasons including alienation from the political process or because they came to the polls to vote for another office on the ballot (senate or a local office such as dogcatcher) but did not have the time or energy to learn who to vote for in the presidential contest.

7. When just examining ballots, talking to election officials, or analyzing precinct- or county-level data, it is not possible to distinguish intentional from accidental undervotes. Voter self-reports represent the only systematic way to estimate the incidence of intentional undervoting. Survey questions from the National Election Studies over a period of 20 years (NES) and from the Voter Research and Surveys exit polls (VRS, more familiarly known as Voter News Services — VNS — as it was known until 2002) are used here to estimate the number of intentional undervoters. Based on responses to these surveys, we found that a minimum of one-ninth, but no more than two-fifths of invalidated presidential votes are accounted for by intentional undervoting. [Footnotes omitted].

(c) Against that background, Plaintiffs' Exhibit 35, attached as Appendix III, and recording the residual vote in Ohio on a county-by-county basis for the office of the President in 2000, bears scrutiny. The percentage of residual votes using Punch Card Votomatic machines was 2.3% across the state, meaning that out of 1000 voters, 23 did not cast a vote for any candidate in the 2000 presidential election. By the same token, the residual vote using electronic machines was 0.7%, meaning that out of 1000 voters, 7 did not vote for any candidate for the president.

(d) Upon acceptance of Dr. Kropf's determination in paragraph 7 above and using the minimum of 1/9 factor and multiplying the 23 nonvotes by 1/9 (.11) the result is a determination that approximately 2.5 persons out of 1000 intentionally did not vote for the office of the presidency and using the two-fifths percentage, then slightly over nine persons out of 1000 intentionally did not vote for the presidency. Then applying the minimum/maximum numbers used by Dr. Kropf in paragraph # 7, the accidental nonvote using punch card technology in the presidential election in the year 2000 ranged from a high of over 20 votes out of 1000 to a low of 14 votes out of 1000. Under either analysis, the residual vote based on accidental nonvoting with punch card technology is higher than the residual vote using the electronic voting machine; i.e., seven out of 1000 votes cast. Thus, the number of accidental nonvotes, under Dr. Kropf's analysis, would range from a minimum of seven to a maximum of thirteen in a group of 1000 voters.

(e) Viewing the plaintiffs' case in a light most favorable to them, leads to the conclusion that seven to thirteen voters out of 1000 using punch card technology accidentally failed to record a vote in the year 2000 in the presidential election.²² 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.²³

22 The adoption of HAVA, the passage of Substituted House Bill 262 and the determination of the defendant Secretary of State to replace punch card voting technology certainly indicate that confidence in the punch card voting technology has waned. The punch card technology has apparently served its time. However, the current delay in perfecting the DRE voting technology and the recitation of difficulties with the DRE technology in the recent 2004 election suggest that Ohio will need to make use of punch card voting technology for the immediate future.

23 A troubling feature of the punch card voting technology is the "fragility" of the ballot and the difficulty encountered in re-counts because of the fragility of the ballot. However, that feature does not, in the court's view, justify a conclusion that the use of the punch card voting technology constitutes either a violation of the Constitution or a violation of the Voting Rights Act.

3. To label the punch card voting technology a non-notice system is to ignore the reality that the careful voter has every opportunity to scrutinize his or her ballot after removal from the voting tray to determine if a mistake has been made in the context of an undervote or an overvote or a mistaken vote and to request a new ballot in the event of a recognized mistake. While a DRE voting technology system, when finally adopted for the state of Ohio, will provide the voter a better recognition of a nonvote and prevent a mistaken overvote, that forthcoming apparent improvement does not justify a declaration that the punch card technology is unconstitutional.

4. The highest level of residual voting is located in counties in Ohio with a very small African-American population which indicates that residual voting is not raceoriented.

D. Additional Legal Determinations

In addition to these findings, the Court also reaches the following legal determinations regarding plaintiffs' Voting Rights Act and Equal Protection claims, which further support the Court's ruling that the plaintiffs have failed to make a case for judicial intervention.

1. Plaintiffs' Voting Rights Act Claims

Plaintiffs' Voting Rights Act claims fail because their alleged injury does not amount to a vote denial under § 2 of the Voting Rights Act. Section 2(a) of the Voting Rights 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." 42 U.S.C. § 1973(a). Under Section 2(b), an election practice or procedure violates Section 2(a) whenever, based on the totality of the circumstances, members of a protected class have less opportunity to participate in the political process than other members of the electorate. 42 U.S.C. § 1973(b).

A violation of subsection (a) of this section is established 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.

Id. There are two separate and distinct theories under which a plaintiff can assert a claim under the Act, vote denial and vote dilution. See Holder v. Hall, 512 U.S. 874, 114 S.Ct. 2581, 129 L.Ed.2d 687 (1994); Farrakhan v. Washington, 338 F.3d 1009 (9th Cir. 2003). Plaintiffs here only allege a claim of vote denial. See Plaintiffs' Mem. Opposing SJ, pg. 6, Doc. No. 187 ("In their summary judgment brief, the Plaintiffs made clear that they were asserting only a vote denial claim, not a claim for vote dilution."). A vote denial arises when a state or a municipality employs a "practice or procedure" that results in the "actual" denial of the right to vote on account of race. 42 U.S.C. § 1973(a); see also Muntaqim v. Coombe, 366 F.3d 102, 105 (2d Cir. 2004).

Here, none of the plaintiffs, including the African-American plaintiffs, claim that they have been denied access to the polls. Rather, the African-American plaintiffs contend that punch card ballots subject them to a greater probability that their votes will not be counted than whites. They do not argue that punch card ballots are employed disproportionately in African-American areas of the state, indeed punch card ballots are widely implemented throughout Ohio, including in counties where less, than 1% of the population is African-American. Furthermore, in Hamilton, Montgomery, and Summit Counties, all voters, be they African-American, white, or otherwise, use punch card ballots.

When coupled with the previously referenced de minimis effects of the punch card ballot, these facts do not allow this Court to conclude that an "actual" denial of the right to vote on account of race occurs. All voters in a county, regardless of race, use the same voting system to cast a ballot, and no one is denied the opportunity to cast a valid vote because of their race. Thus, African-American voters have the same opportunity to participate in the political process as other members of the electorate. The Court, therefore, holds that the plaintiffs have not established their vote denial claim.

2. Plaintiffs' Equal Protection Claims

The Court further notes 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. The Equal Protection Clause provides that once a state has granted its citizens the right to vote, it "may not . . . value one person's vote over that of another." Bush, 531 U.S. at 104-105, 121 S.Ct. 525 (per curiam). Local variety in voting technology, however, does not violate the Equal Protection Clause, even if the different technologies have different levels of effectiveness in recording voters' intentions, so long as there is some rational basis for the technology choice. 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; local variety can be justified by concerns about cost, the potential value of innovation, and so on.

Id. "Rational basis review does not require us to identify the legislature's actual rationale for the distinction; rather, we will uphold the statute if `there are plausible reasons for [the government's] action.'" Hamama v. Immigration Naturalization Serv., 78 F.3d 233, 237 (6th Cir. 1996) (quoting United States R.R. Retirement Bd. v. Fritz, 449 U.S. 166, 179, 101 S.Ct. 453, 66 L.Ed.2d 368 (1980)). Thus, so long as there are plausible reasons for the maintenance of different voting mechanisms, that practice will not be considered unconstitutional.

Here, defendants have offered reasons for their continued use of punch card ballots. Defendants suggest that the punch card ballot has been a cost effective vote tabulating device. They have further indicated some concern regarding the security of DRE machines that the machine vendors are currently resolving. While DRE machines are certified for current use, they have not been HAVA certified by the Secretary of State due to security concerns. The Court finds defendants' cost and security reasons for the use of punch card ballots plausible.

Furthermore, HAVA requires that polling places have a DRE machine for use by disabled voters. 42 U.S.C. § 15481(a)(3). The Secretary, therefore, did not require the replacement of punch card ballots with precinct-count optical scan systems because a county may determine that it desires all of its voters to be able to cast their votes on the same machines. Because no DRE machines have been HAVA certified, a county could wish to await certification before purchasing DRE machines to avoid the security concerns and future expense in updating the machines. The Court finds these reasons for delaying the replacement of punch card ballots, plausible and reasonable. The Court, therefore, holds that there is a rational basis for defendants' continued use of punch card ballots.

Moreover, Sandusky County's use of central-count optical scan also has a rational basis. The Court heard the testimony of Barbara Tuckerman, who stated that Sandusky County uses central-count optical scan because of the cost of buying, storing, and moving the larger number of machines necessary for implementing the precinct-count optical scan technique. She further indicated that the optical scan system was chosen because of its relative simplicity for the voter, as compared to DRE machines. Additionally, she noted that the poll workers would not be familiar with the optical scan machines and, therefore, would be unable to fix a machine if it were to breakdown.²⁴ Thus, the Court finds that there are plausible reasons for Sandusky County's use of the centralcount optical scan technique, and, therefore, it has a rational basis.

24 Sandusky County's concerns with the precinct-count optical technique serve as further reasons why the other defendant counties have not opted for the use. of such a system.

As the Court has previously noted, voting technology has been a continuously developing aspect of civic culture. Unfortunately, we have yet to produce the perfect voting system, one that allows for immediate error-free results. If states were not permitted to employ different types of voting technologies within their borders, this development could very well come to a halt.

JUDGMENT FOR THE DEFENDANTS.

APPENDIX I PRETRIAL STIPULATION OF FACT SUBMITTED BY PARTIES

1) Plaintiffs, Erin Otis and Vernellia Randall, are citizens of Ohio and registered voters who reside in Montgomery County.

2) Plaintiffs, Howard Tolley and Art Slater, are citizens of Ohio and registered voters who reside in Hamilton County.

3) Plaintiffs, Effie Stewart and Marco Sommerville, are citizens of Ohio and registered voters who reside in Summit County.

4) Plaintiff, Linda See, is a citizen of Ohio and registered voter who resides in Sandusky County.

5) Plaintiffs, Randall, Slater, Stewart and Sommerville are African-Americans.

6) Defendant, J. Kenneth Blackwell, is the Secretary of State of Ohio and chief elections officer of the State of Ohio.

7) Defendants, Raymond Butler, Geraldine Lewis and Larry Loutszenhiser are members of the State Board of Voting Machine Examiners for the Approval of Electoral Marking Devices.

8) Defendants, Alex Arshinkoff, Joseph Hutchinson, Wayne Jones and Rus Pry are members of the Summit County Board of Elections.

9) Defendant, Christopher Heizer is Director of the Montgomery County Board of Elections, Defendant Stephanie Harsman is Deputy Director of the Montgomery County Board of Elections, and Defendants, Sue Finley, Thomas Ritchie, Sr., James Nathenson, and Dennis Lieberman are members of the Montgomery County Board of Elections.

10) Defendants, Timothy Burke, Michael Barrett, V. Daniel Radford and Todd Ward are and at all relevant times were members of the Hamilton County Board of Elections.

11) Defendants, Harry Heyman, Thomas Yonker and John Rettig are and at all relevant times were members of the Sandusky County Board of Elections.

12) Pursuant to Ohio Rev. Code Ann. § 3506.06(B) (Anderson 2003), Defendant, J. Kenneth Blackwell has certified and Defendant, Ohio Board of Voting Machine Examiners ("State Defendants") have approved a variety of voting systems for use in Ohio elections, including equipment that contains a feature which gives voters notice of overvotes and/or undervotes and also equipment that does not contain such a feature.

13) The election systems that State Defendants have certified and that county officials currently utilize are: 1) punch card voting; 2) optical scan voting; 3) direct record electronic (or "DRE") voting; and 4) lever machines (also known as "automatic voting machines").

14) Ohio is pervasively a punch card voting state. Sixty-nine (69) of Ohio's 88 counties use punch card voting. Those 69 counties include 72.5% of the states' registered voters, and 74% of the 11,756 precincts in the State of Ohio.

15) Among the 19 non-punch card counties in the State of Ohio, 11 use optical scan equipment, six use electronic voting equipment, and two use lever machines.

16) Some voting systems have a feature which provides some form of active feedback from a ballot reader that instantly informs a voter of potential errors on her ballot and allows the voter to make any needed corrections. This feature is commonly know as a "second chance" feature.

17) Most counties in Ohio do not utilize voting machines which have the described "second chance" feature.

18) Punch card voting machines do not provide voters with actual notice of overvotes and undervotes.

19) The punch card system relies on a ballot card with pre-scored, square perforations or "chad." The names of candidates are not on the ballot card, but each chad is assigned a number which appears in small type on the ballot card. The ballot card has a stub at one end, with two holes that allow the card to be held in place over pins, when the ballot is inserted into the "Votomatic" device. This device is placed on a table in front of the voter and holds the ballot card during the voting process. The device has attached to it a loose-leaf booklet, with the names of candidates or other items to be voted on listed in the booklet. In the center of the booklet is a slot which should line up with columns in the ballot card. Along the slot are holes. If the ballot card is correctly positioned within the device and the device has been correctly assembled, each chad lies beneath a different hole. A voter casts his or her vote by placing the ballot card in the device, and exerting pressure on a stylus through the hole to "punch" chad in the ballot card corresponding to the desired choice. After the voter has completed the voting process, he or she removes the card from the device. The stub is removed from the ballot. The cards are thereafter run through a mechanical tabulator and recorded by a computer.

20) Hamilton, Summit, and Montgomery Counties are among the 69 counties that use punch card voting equipment.

21) There are two types of optical scan systems used in Ohio: "centralcount" optical scans, and "precinctcount" optical scans. Precinct-count optical scan systems have an errorcorrection feature which gives voters a chance to discover and correct possible mistakes at the poll, while central-count optical scan systems does not have such a feature.

22) In the 2000 election, Sandusky County used a punch card voting system. Beginning with the 2001 election and continuing for all subsequent elections, Sandusky County has utilized a central-count optical scan system.

23) Electronic or "DRE" voting machines can be programmed to alert voters that they have not made a choice in a particular race. DRE technology advises voters of the choices they have made, showing whether they have undervoted, and gives them a message allowing them to review and verify their choices before casting their votes. Additionally, DRE technology can be programmed to prevent overvoting.

24) An "overvote" occurs when a voter registers more than one choice for a candidate in a particular race and thereby disqualifies his or her vote for that particular race. An "undervote" occurs when a voter does not mark a ballot in a particular race or votes for fewer than the allowed number of candidates. "Residuals votes" are overvotes and undervotes combined.

25) The counties in Ohio experiencing the highest percentages of residual votes in the 2000 presidential election were those in which voters used punch card ballots, while the counties experiencing the lowest percentage of residual votes in this election used other technology. The 29 counties in Ohio with the highest residual vote percentages in the 2000 presidential election were all counties that use the punch card method of voting. The seven counties with the lowest residual vote percentages in the 2000 election were all counties that did not use punch cards as their primary voting system.

26) The information contained in Tables 1-6 of John Lott's Report is accurate.

27) The information contained in Tables 1 and 2 of Martha Kropf's Affidavit is accurate.

28) In the 2000 presidential election, the statewide residual vote rate was 1.88%.

29) The information contained in the Summary Report for the 2000 election in Ohio is accurate. See Doc. No. 187, Appendix E.

30) Some residual votes occur when voters intentionally choose not to cast votes in a particular race. Examining ballots does not allow for intentional and unintentional undervotes to be distinguished.

31) The information contained in Table 1 of Kropf's Report is accurate.

32) The three counties that are the subject of Plaintiffs' Voting Rights Act claim, Hamilton, Summit and Montgomery Counties, all use the Votomatic punch card voting system.

33) The demographic information for Hamilton, Summit, Montgomery and Franklin Counties contained in Richard Engstrom's Report is accurate.

34) The Census 2000 Summary for Hamilton, Summit and Montgomery Counties is accurate.

35) Dr. Richard Engstrom used three analytic procedures to assess the extent to which the African-American voters overvoted or undervoted at different rates than non-African-American voters in each of these counties. Homogeneous precinct ("HP") analyses simply report the percentage of the voters that overvoted or undervoted in the precincts in which over 90 percent of the voting age population was not African American and in those in which over 90 percent was African American. Ecological Regression ("ER") analyses provide estimates of these rates for African American and non-African American voters based on the votes cast in all of the precincts in an election. The third methodology is called Ecological Inference ("El"). This is an estimation procedure that also takes into account all of the precincts in which votes are cast.

36) The information concerning overvoting in Hamilton and Summit Counties contained in Engstrom's Report is accurate.

37) The information concerning undervoting in Hamilton and Summit Counties contained in Engstrom's Report is accurate.

38) The information concerning overvoting and undervoting in Montgomery County contained in Engstrom's Report is accurate.

39) The 2000 Ohio Census Data giving rise to State Defendants' Table of residual vote rates in fourteen Ohio Counties is accurate.

40) Adams County, which has a total of 48 African-Americans living in the entire county, had 492 residual votes for President in the 2000 Presidential Election. Holmes County, which has 127 African-American residents, had 792 residual ballots in the 2000 Presidential election. Jackson, Meigs, Monroe, Pike, and Vinton Counties also had more residual votes for President in the 2000 election than African-Americans living in the jurisdiction.

41) African-Americans make up 27.4% of the population of Cuyahoga County, and there are more African-Americans living in Cuyahoga County than there are people in 83 of Ohio's 88 Counties.

42) In the 2000 presidential election, Cuyahoga County used punch cards.

43) Delaware County, Ohio uses punch cards while Franklin County uses electronic voting machines and, in the 2000 elections, Delaware County had 55,959 total votes cast and 55,403 votes cast for President for a residual vote total of 556 votes or 0.99% while Franklin County had 417,800 total votes cast and 414,074 votes cast for President for a residual vote total of 3,726 votes or 0.89%.

44) Shortly after the effective date of the Help America Vote Act ("HAVA"), Ohio Secretary of State J. Kenneth Blackwell reconfigured his office and required many of his senior staff to focus on implementing HAVA in the State of Ohio.

45) In May of 2003, Secretary Blackwell's Office finalized Ohio's HAVA Plan and published the document for the public's review.

46) In September 2003, the Secretary of State's Office qualified the following four vendors to offer voting devices in Ohio: 1) Sequoia Voting Systems; 2) Diebold Elections Systems; 3) Election Systems Software; and 4) Maximus/Hart Intercivic/DFM Associates, pending a security review for all prospective voting devices.

47) The security reviews revealed that there were 57 potential security risks within the software and hardware for prospective voting devices and, in December 2003, the Secretary of State ordered the qualified voting device vendors to resolve the identified security concerns.

48) The Ohio Secretary of State's Office submitted a request to the State's Controlling Board for release of monies to begin the process of replacing the State's existing voting technologies, and the request was temporarily delayed. The Ohio General Assembly created a House-Senate Ballot Security committee to ensure that all replacement machines will provide accurate, reliable and tamper-proof results.

49) On May 7, 2004, Governor Bob Taft signed H.B. 262 into law.

50) Punch card voting equipment in Hamilton County is maintained on an ongoing basis and it is cleaned at least annually by having the displaced chad removed from the interior tray. The cleaning includes sharpening the styluses in advance of every election.

51) Hamilton County poll workers are trained on how to set up the precinct polling location, how to conduct the election, and what to do when the polls close.

52) Hamilton County poll workers are instructed to ask every voter if they require assistance or would like a demonstration on how to cast a valid vote.

53) All Hamilton County polling locations have a demonstrator voting machine so that voters may practice prior to voting.

54) Hamilton County poll workers are trained on provisional balloting.

55) Hamilton County poll workers are trained on what to do with a soiled or defaced ballot.

56) There are voting instructions permanently affixed to all individual voting booths in Hamilton County. Further, instructions are included on the ballots themselves and handouts of the instructions are available to the voter.

57) The instructions in the booths and on the ballots in Hamilton County are mandated by the Secretary of State.

58) The voting instructions placed in each booth in Hamilton County provide a written description of the six step process involved in casting a valid ballot and a visual diagram on how it is to be accomplished.

59) The voting instructions placed in each booth in Hamilton County also advise the voter: 1) what to do if they require assistance due to blindness, disability, or inability to read or write; and 2) that they may return a torn, soiled, defaced, or erroneously marked ballot to the precinct election official and have a new ballot issued to them.

60) The voting instructions included on each ballot in Hamilton County advise the voter to: 1) insert your ballot card with both hands; 2) be sure the red pins are through the two holes; 3) use only the punching tool attached to the Votomatic; 4) if you make a mistake, return your ballot card to the precinct official and receive a new one; 5) to writein a candidate name, write the candidate name and office on the top of the ballot card.

61) The Hamilton County Board of Elections has procedures for remaking ballots that are defaced, ripped, or put in backwards.

62) In Hamilton County, ballots that are defaced, ripped or may have some problems with it, or where a ballot was placed in backwards in the machine, are identified and, where it is possible to determine what the voter intended, the Ballot is "remade" and then counted.

63) The ballots are remade by a bipartisan team comprised of staff members of the Hamilton County Board of Elections. If there is a dispute regarding remade ballots, it is brought to the members of the board.

64) The Chairman of the Hamilton County Board of Elections is unaware of any specific instance in which the Board has been unable to agree on if, or how, a ballot should be remade.

65) In Hamilton County, in the event a ballot is remade, both the original ballot and the remade ballot are retained by the Board of Elections.

66) When a Sandusky County voter receives her optical scan ballot, a poll worker gives personal instructions to that voter, including to use only the pencil provided in the voting booth and to look at the instructions printed on the top of the ballot. Additionally, the voter is shown the candidates and issues and is also shown the corresponding ovals, and is also told to color in the oval to the left of their choice. Finally, the same instructions are written and posted in the voting booth right in front of the voter's eyes.

67) In Sandusky County, in order to cast a vote on the optical scan ballot, the voter simply blackens in the oval to the left of the voter's choice with the pencil provided. Voters are informed that they should return their ballot to the presiding poll judge if they make a mistake on their optical scan ballot to receive a new ballot. When the voter has completed voting on the optical scan ballot, she takes it and places it into the ballot box. After the polls are closed, the ballot boxes are locked and brought back to the office of the Sandusky County Board of Elections. These boxes are then opened and several employees examine the ballots to make sure that they are all facing the correct way. The ballots are then passed on to a board member who brings them up to the tabulating machine (a.k.a.tabulator), puts the ballots into the tabulating machine, and then runs the ballots through the tabulating machine. The tabulating machine is operated by either a clerk or by the Director of the Board of Elections. If a ballot is rejected by the tabulating machine, the machine stops and a digital read-out on the machine states what the problem is. The problem is that an overvote, an undervote, or a blank ballot was detected. When this occurs, two board members then look at the ballot, and, according to the guidelines given by the Ohio Secretary of State, they determine what to do with the ballot.

68) In Sandusky County, when a ballot is rejected, for example, because a voter did not fill in the oval enough, and it is clear that the voter did not intend to vote for someone else, then that oval is filled in completely.

69) A voter might use an unauthorized device, such as an ink pen, to fill out their ballots. However, in Sandusky County, this only will occur if the voter ignores one explicit oral instruction from a poll worker not do that, as well as two written instructions.

70) With the optical scan ballot system currently used by Sandusky County, like the punch card ballot system before it, the Director of the Sandusky County Board of Elections and the Deputy Director hand count at least two precincts after every election in order to double check that the machines are accurately counting the votes cast. To date, none of the hand counts has demonstrated a discrepancy. The Director and Deputy Director will continue to do a hand count. This will continue to occur even when touch screen ballots are utilized.

71) Sandusky County switched from punch card voting machines because it was difficult to get the machines repaired. Sandusky County decided to use an optical scan system with central location tabulation, as opposed to an optical scan system with in-precinct tabulation because the in-precinct tabulation machines are more expensive than the central tabulation machines and because the inprecinct machines are more cumbersome.

72) Sandusky County has and continues to uniformly administer the optical scan ballot with central location tabulation to all of its voters.

73) While it is not possible to cast an overvote in Franklin County, which uses electronic voting equipment, it is still possible to cast a mistakenly valid vote (cast a valid vote, but only for someone other than who the voter intended to vote for).

74) In Franklin County, a voter error occurs when a voter fails to press the "vote" button, or a voter-parent takes a child into the booth with them and the child presses the "vote" button before the voter-parent has completed the voting. Nothing can be done to remedy these errors.

75) Voters do report that they intentionally undervoted, or did not cast a vote, in presidential elections.

76) Short of violating voters' privacy rights, asking voters if they voted in a particular race or not (through an exit poll) is the only way to figure out if an undervote is intentional or accidental.

77) Sandusky County did not use an optical scan ballot in 2000, but instead used a punch card ballot.

78) In a special election in Sandusky in May 2003, which only had one issue on the ballot, with the exception of one district that had two issues on the ballot, there were only three overvotes cast that were rejected by the tabulation machine.

79) In the 2000 General Election in Ohio, a 0.93% residual vote rate occurred in Allen County (which used optical scan with in-precinct tabulation) and 2.4% residual vote rate occurred in Mahoning County (which used optical scan with central tabulation). The only other county that used an optical scan ballot with in-precinct tabulation in the 2000 Presidential election was Hancock County, which had a residual vote rate of 1.2%. Ashland, Geauga, Hancock^1a and Ottawa Counties all used optical scan ballots with central tabulation in that election and had the same or a lower residual vote rate than did Hancock County. Additionally, Coshocton County only had a residual vote rate of 1.5% in that race. Coshocton County used an optical scan ballot with central tabulation.

80) The parties stipulate to the authenticity and accuracy of the following remarks of the Chairman of the Hamilton County Board of Elections, Mr. Timothy Burke, as made in a letter to Congresswoman Stephanie Tubbs Jones, dated July 18, 2001 (Doc. 171-40-7c):

a) "2.25% of the people who voted in the 1999 Cincinnati City Council Election had their votes disqualified because they overvoted by voting for more than nine candidates."

b) "Twenty-three of the twenty-five precincts in Hamilton County that experienced the highest rates of overvoting in the 1999 Cincinnati City Council Election were minority, largely poor precincts."

c) "Of the 25 precincts that had the most overvotes cast in the 2000 Presidential Election in Hamilton County, all 25 appear to be minority, mostly poor precincts."

81) In the presidential election of 2000, the highest rates of overvoting in Hamilton and Summit Counties occurred in Cincinnati Precinct 22-F (6% overvote, 95% black) and Akron Precinct 3-K (6% overvote, 80% black) where substantial percentages of black voters reside. (Docs. 206-3-c, 206-4-d).

82) The voting databases from the Ohio Secretary of State's Office and the Summit, Hamilton, and Montgomery County Boards of Election, which all counsel and experts on both sides of the case have used to determine the level of residual ballots, overvotes, and undervotes in the precincts, wards, and counties of Ohio in the 1992, 1996, and 2000 presidential and U.S. Senate elections, as well as the City of Cincinnati mayoral and city council elections of 2001 and the Hamilton County Primary Election of 2004, are authentic and accurate.

83) The State Help America Vote Act (HAVA) Report (Doc. 115-1) is authentic and accurate.

84) All exhibits in trial binders are authentic.

85) All named Plaintiffs voted in the 2000 presidential election, with the exception of Vemellia Randall. All named Plaintiffs, including Vernellia Randall, plan to vote in subsequent presidential elections.

86) There is no evidence in the record of these proceedings that the named Plaintiffs from Hamilton, Montgomery, Sandusky, and/or Summit County were denied in any way equal access to the polls.

87) There is no evidence in the record of these proceedings that the named Plaintiffs from Hamilton, Montgomery, Sandusky, and/or Summit County were denied in any way equal access to the voter instructions placed in each voting booth, on each ballot, and available at the polling location.

88) There is no evidence in the record of these proceedings that the named Plaintiffs from Hamilton, Montgomery, Sandusky, and/or Summit County were denied in any way equal access to assistance from poll workers in casting a ballot, if they required it.

89) There is no evidence in the record of these proceedings that the named Plaintiffs from Hamilton, Montgomery, Sandusky, and/or Summit County were prevented from attempting to cast their vote in the 2000 general election.

90) All deposition testimony arising out of these proceedings is indeed the witness' testimony, and it may be admitted into evidence.

91) The information contained in Tables 2, 3, and 4 of Martha Kropf's Report is accurate.

92) In the 2001 Cincinnati mayoral election, a Hamilton County election, 89,000 votes were cast, of which 113 votes were overvotes — a citywide rate of slightly more than one-tenth of one percent. (Percentages were obtained by dividing the overvote by the ballots cast.) (Doc. 192; Williams affidavit at ¶ 2, exhibit 1).

93) In the 2001 Cincinnati mayoral election, no ward had a percentage rate of overvotes greater than fourtenths of one percent. Twenty-five precincts had more than one overvote while no precinct had more than three overvotes. The range of overvotes of all precincts where more than one overvote occurred was 0.34 to 1.6 percent with six of these precincts falling within the range of equal to or greater than one percent. (Doc. 192; Williams affidavit at ¶ 2, with attached exhibits).

94) In the Cincinnati City Council election of 2001, out of 376 precincts, 101 precincts had an overvote rate of 2%, 79 precincts had an overvote rate of 3%, 50 precincts had an overvote rate of 4%, 18 precincts had an overvote rate of 5%, 10 precincts had an overvote rate of 6%, 9 precincts had an overvote rate of 7%, 2 precincts had an overvote rate of 8%, and 2 precincts had an overvote rate of 10%.

95) In the Republican Party Primary for Hamilton County Commissioner which was held on March 2, 2004, overvoting occurred within the townships with largely white populations at a higher rate than the rate of overvoting which occurred in the 2001 Cincinnati mayoral election. All but one of these townships had rates over one percent, with a range of 0.7 to 2.0 percent.

96) The percentage of white residents within Hamilton County Townships together with the percentage of overvoting occurring in the 2004 Republican Party Primary for Hamilton County Commissioner are: Anderson Township, 96.57% white population, 1.5% overvote; Colerain Township, 87.77% white population, 1.1% overvote; Columbia Township, 61.74% white population, 1.3% overvote; Crosby Township, 97.89% white population, 0.7% overvote; Delhi Township, 97.65% white population, 1.8% overvote; Green Township, 97.53% white population, 2.0% overvote; Harrison Township, 98.23% white population, 1.5% overvote; Miami Township, 98.71% white population, 1.7% overvote; Springfield Township, 67.06% white population, 1.5% overvote.

97) The parties stipulate to the admissibility of the optical scan ballot used by the Sandusky County Board of Elections in the November, 2002 General Election, which is entitled "General Election Ballot — November 5, 2002, Sandusky County Ohio."

98) The parties stipulate to the admissibility of the written directions that the Sandusky County Board of Elections has posted in the voting booth right in front of the voter's eyes in each election in which an optical scan ballot has been utilized.

1a Hancock County used both optical scan ballots with both in-precinct and central tabulation.

APPENDIX II SUPPLEMENTAL FACT FINDINGS BY THE COURT AFTER THE TRIAL

Dr. Martha Kropf Trial Testimony:

99) Dr. Kropf opined that undervoting on purpose is rare, but relied on the data for this opinion from the University of Michigan's National Election Study (NES) and Voters News Service (VNS) studies, which were based on post-election polling.^1b

100) NES studies consistently over-report voter turnout.^2b

101) The VNS database, which is based on a nationwide survey, has 54,806 voters for the 1992 election, and is not just specific to Ohio.^3b

102) The 2000 NES survey involved questions submitted to 7,699 voters but there was no proof that any of the voters were from Ohio.^4b

103) Dr. Kropf's report was based on national data and not data specifically from Ohio.^5b

104) Dr. Kropf's report, which included a paper on voided ballots in the 1996 election, did not include education as a factor when evaluating intentional undervoting.^6b

105) The 1992 NES survey began on November 4, 1992 and ended on January 13, 1993.^7b

106) The 1996 NES survey began on November 6, 1996 and ended on December 24, 1996.^8b

107) The NES overestimated voter turnout.^9b

108) The 2000 NES survey began on November 8, 2000 and ended on December 18, 2000.^10b

109) The 1988 NES survey began on November 8, 1988 through January 30, 1989.^11b

110) The only way to tell if someone who has been the subject of a postelection questionnaire to determine whether the person so questioned actually voted is to look at voting and registration records.^12b

111) The 1988 NES survey checked on those surveyed to determine if there was a record of them voting but determined that there was no record of an actual vote by 3-5% of those surveyed.^13b

112) Dr. Kropf conducted a nationwide regression analysis with controlling factors, but not in each county in Ohio.^14b

1b See Trial Transcript, pp. 85-86.

2b The following testimony was elicited from the cross-examination of Dr. Kropf as follows:

Q. Did you hear Mr. Coglianese's opening statement?

A. I heard a little bit of it.

Q. Did you hear him say that there had been reports that NES data consistently overreports voter turnout?

A. I've heard that separately from his opening statement, and I did hear a little bit about that, yes.

See Transcript, pg. 109, ln. 15-19.

3b See Transcript, pg. 110.

4b See Transcript, pg. 112.

5b See Transcript, pg. 121.

6b Dr. Kropf offered the following testimony regarding her paper on voided ballots in the 1996 election as follows:

Q. You mentioned that you did a paper, Voided Ballots in the 1996 Election?

A. Yes, sir.

Q. And that that was a county level analysis?

A. Yes, sir.

Q. What county was that we are talking about?

A. Well, it's looking at about two-thirds of the counties in the 1996 election for which data were available.

Q. All right. On a national basis?

A. On a national basis.

Q. Okay.

A. That's right.

Q. All right. In the report that you submitted, looking at Tables 1, 2 and 3, and 3 in particular, did you control for education?

A. No. When we had originally ran the analysis, education hadn't mattered, so we just didn't include it because it reduced our sample size unacceptably. It's not asked in every state on the exit polls and, I don't know, but it's not.

Q. Do you think that would be something that might be important with regard to a question of intentional undervoting?

A. Well, I'll say yes, it is important, but it is not included here.

See Transcript, pg. 121, ln. 9 through pg. 122, ln. 8.

7b See Transcript, pg. 208, Exhibit TT, pg. 28.

8b See Transcript, pg. 209, Exhibit UU, p. 9, Table 3.

9b See Transcript, pg. 210, and Exhibit UU, p. 11.

10b See Transcript, pg. 213 and Exhibit W, p. 10.

11b See Transcript, pg. 214 and Exhibit SS, p. 9.

12b See Transcript, pg. 215, and Exhibit SS, p. 18.

13b See Transcript, pg. 216.

14b See Transcript, pg. 220.

Dana Walch Trial Testimony:

113) Ohio counties used four types of voting equipment in 2000 — punch cards, lever machines, optical scan devices and direct recording electronic devices (DRE's).^15b

114) House Bill 262 requires that, in 2006, all DRE voting devices in Ohio must use Voter Verified Paper Audit Trails (VVPAT). House Bill 262 also mandates that any receipt or paper verification be behind glass.^16b

115) House Bill 262 requires that a paper record be the official vote if there is a recount, so vendors are working on different systems to keep official voter records with VVPAT.^17b

116) The Secretary of State intends to use HAVA money for DRE or precinct-count optical scan systems. Such non-notice systems would provide more options to voters in reducing error rates.^18b

117) The Secretary of State has progressed with his procurement process, and wants to implement new voting technology by 2005.^19b

118) No vendor has passed a final security check in Ohio for new DREs.^20b

119) The Secretary of State did not approve and deploy optical scan inprecinct count because he wanted the counties to retain the ability to choose what system they wanted. Also, simply choosing optical scan does not meet the requirements of HAVA because such a system does little in helping people with disabilities vote.^21b

15b See Transcript, pg. 134.

16b See Transcript, pg. 142-143.

17b See Transcript, pg. 147-148.

18b See Transcript, pg. 159-160.

19b See Transcript, pg. 732.

20b See Transcript, pg. 732.

21b See Transcript, pg. 752, 753.

Dr. Herb Asher's Testimony:

120) Dr. Asher's study of punch card ballots found that precincts with a higher concentration of poverty had a nonvoted ballot rate higher than the average nonvoted ballot rate in Ohio.^22b

121) Factors such as race, education and poverty are predictor variables.^23b

122) Dr. Asher's 1982 report contends that punch card voting has a higher ballot completion at the bottom of the ballot.^24b

123) In his report for the year 2000, Dr. Asher studied the Ohio Appalachian counties of Pike, Adams, Vinton, Meigs, Noble, Monroe, Jackson, and Gallia, which all have a substantially higher percentage of Caucasians in their population when compared with the percentage of Caucasians in other Ohio counties.^25b

124) The population of African Americans in Adams County is 0.2% and it has a lower education attainment standard than Ohio's average, and its median income level is lower than the rest of the state as a whole. The residual or nonvoted ballot rate for President in the year 2000 was 4.6%.^26b

125) The population of African Americans in Jackson County is 0.6%, and it has a lower educational level, a lower number of high school graduates, and a lower median in come level than the rest of the state as a whole. The residual or nonvoted ballot rate for President in the year 2000 was 3.3%.^27b

126) The population of African Americans in Gallia County is 2.7% and it has a lower educational level, and a lower median income level than the state as a whole. The residual or nonvoted ballot rate for President in the year 2000 was 3.2%.^28b

127) The population of African Americans in Monroe County is 0.3% and it has a lower educational level, a lower number of high school and college graduates, and a lower median income level than the state as a whole. The residual or nonvoted ballot rate for President in the year 2000 was 3.6%.^29b

128) The population of African Americans in Pike County is 0.9% and it has a lower educational level, a lower education attainment standard, and a lower median income levels than the state as a whole. The residual or nonvoted ballot rate for President in the year 2000 was 4.7%.^30b

129) The population of African Americans in Vinton County is 0.4% and it has a lower population of African Americans (0.4%), a lower educational level, a lower number of high school and college graduates, and a lower median income level than the state as a whole. The residual or nonvoted ballot rate for President in the year 2000 was 4.6%.^31b

130) The population of African Americans in Noble County is 6.7% and it has a lower population of African Americans (6.7%), a lower educational level, a lower number of high school and college graduates, and lower median income levels than the state as a whole. The residual or nonvoted ballot rate for President in the year 2000 was 3.6%.^32b

131) The population of African Americans in Meigs County is 0.7% and it has a lower population of African Americans (0.7%), a lower educational level, a lower number of high school and college graduates, and lower median income levels than the state as a whole. The residual or nonvoted ballot rate for President in the year 2000 was 4.2%.^33b

132) In Pike, Adams, Vinton, Meigs, Noble, Monroe, Jackson, and Gallia Counties, where African Americans make up less than 1% of the population, something besides race caused nonvoted ballots, such as educational and income levels.^34b

133) One factor could never be the sole factor in error rate and the explanation for everything.^35b

22b See Transcript, pg. p. 797.

23b See Transcript, pg. 800.

24b See Transcript, pg. 813.

25b See Transcript, pg. 819.

26b See Transcript pg. 820 and Plaintiffs' Exhibit 35.

27b See Transcript, pg. 821 and Plaintiffs' Exhibit 35.

28b See Transcript, pg. 821-822 and Plaintiffs' Exhibit 35.

29b See Transcript, pg. 822 and Plaintiffs' Exhibit 35.

30b See Transcript, pg. 822-823 and Plaintiffs' Exhibit 35.

31b See Transcript, pg. 823. and Plaintiffs' Exhibit 35

32b See transcript, pg. 823-824 and Plaintiffs' Exhibit 35. The Court has accurately recorded the figure of 6.7% as the percentage of African-Americans in Noble County reported in Dr. Asher's deposition testimony that was read into the record. The Court, however, believes that figure to be inaccurate. Dr. Asher's later statement, which comprises fact finding # 132, that African-Americans are less than 1% of the population of Noble County leads the Court to believe that the actual figure is more likely 0.67% or 0.7%. The Court, therefore, will recognize the accuracy of fact finding # 132 and treat Noble County as having an African-American population of less than 1%.

33b See Transcript, pg. 824 and Plaintiffs' Exhibit 35.

34b See Transcript, pg. 825.

35b See Transcript, pg. 826

Dr. Richard Engstrom Trial Testimony:

134) Dr. Engstrom's study of the Presidential election in the year 2000 in Ohio was limited to Hamilton, Franklin, Summit and Montgomery counties.^36b

135) Dr. Engstrom's study in the year 2000 did not include any other race except for the office of the President. He did not look at any other county in Ohio, or any other elected office except the 2000 presidential race.^37b

136) Dr. Engstrom's study of the Presidential race in the year 2000 as to residual or nonvoted ballots did not account for income levels, education levels, age, disabilities, or foreign languages.^38b

137) The residual vote for non-African-Americans in Summit County in the 2000 Presidential election was higher than residual vote of African-Americans in Hamilton County.^39b

138) Using ecological inference analysis, racial disparity is four times higher in Franklin County than in Hamilton County.^40b

139) Dr. Engstrom found that using the ecological regression analysis, Hamilton County's residual vote rate was slightly higher and, in comparing homogeneous precincts in Hamilton County and Franklin County, the residual vote rate was nearly identical.^41b

140) Dr. Engstrom conceded, that in terms of undervoting, Franklin County voting technology actually increases, decreases or has the same racial disparity as Hamilton County.^42b

36b See Transcript, pg. 484.

37b See Transcript, pg. 485.

38b See Transcripts, pg. 488-489.

39b See Transcripts, pg. 490-491.

40b See Transcripts, pg. 493.

41b See Transcripts, pg. 493.

42b See Transcript, pg. 493, line 1 through pg. 494, line 6, where the following testimony of Dr. Engstrom appears:

Q. Okay, Let's take a look at the difference [sic] categories as well.

It appears as though, at least according to ecological inference, when we compare Hamilton County to Franklin County, the racial disparity in Hamilton County is .22, while the racial disparity in Franklin County is .81.

So, in fact, it is a true statement that according to ecological inference, the racial disparity in Franklin County is approximately four times higher than Hamilton County, is that correct?

A. That would be correct.

Q. For purposes of the ecological regression, the racial disparity in Hamilton Count is .75 compared to .65 in Franklin County, so Franklin County fairs slightly better in that analysis, correct?

A. Yes. Hamilton — excuse me — Franklin —

Q. Franklin County fares slightly better?

A. Okay. Lower, the disparity is lower, correct.

Q. Right. But for purposes of homogeneous precinct, in Hamilton County it's .85 and in Franklin County it's .84, I think you previously told His Honor that's identical for all practical purposes, is that right?

A. Yes. I wouldn't make a distinction between them.

Q. So depending on which numbers we want to look at, actually we can claim the voting technology in Franklin County actually increases the racial disparity, decreases the racial disparity, or is the same racial disparity as it relates to Hamilton County, is that correct?

A. In terms of undervoting?

Q. Yes.

A. Yes.

Dr. John Lott Trial Testimony:

141) Dr. Lott examined 1992, 1996 and 2000 presidential election data, and looked at other political races besides the presidential race.^43b

142) Dr. Lott examined the entire state of Ohio and not just three counties.^44b

143) Based on Dr. Lott's study, punch card voting technology in the U.S. Senate races for the years 1992 and 2000 had a lower nonvoted ballot rate that electronic or lever machines. The residual vote or nonvoted ballot rate increases with races other than for the presidency. However, the residual vote does not increase as quickly as the residual vote from most other forms of voting technology.^45b

144) Punch cards outperformed electronic machines in congressional races regarding nonvoted ballot rates. By looking to other races, the data demonstrates that focusing only on presidential races does not produce a complete picture because other races have much higher rates of nonvoted ballots.^46b

145) For presidential races, punch cards tend to do relatively poorly, but as the voter goes down the ballot, the punch cards improve relative to other voting machines.^47b (Punch cards indeed do extremely well across all five races with a low number of nonvoted ballots.)

146) For the races of 1992, 1996 and 2000, punch cards produced fewer nonvoted ballots than either electronic voting machines or lever machines and produced virtually the same results as optical scan machines.^48b

43b See Transcript pg. 544.

44b See Transcript, pg. 545.

45b See Transcript, pg. 561.

46b See Transcript, pg. 562.

47b See Transcript, pg. 570.

48b See Transcript, pg. 574.

APPENDIX III

Summary Report — 1

Sums by Voting Type and Race

Ballots Cast For For Voting System 2000 Election President Senator AVM (Lever) 176,467 175,642 162,022 Electronic 537,474 533,910 504,512 Punchcard 3,593,958 3,512,191 3,319,157 Datavote 38,246 37,118 35,618 Votomatic 3,555,712 3,475,073 3,283,539 Optical Scan 492,182 483,838 466,475 Procinct Count 58,188 57,612 53,963 Central Count 433,914 426,226 412,512 Summary Report — 2

Residual Vote by Voting Type and Race

Ballots Cast For For Voting System 2000 Election President Senator AVM (Lever) 176,467 825 14,445 Electronic 537,474 3,564 32,962 Punchcard 3,593,958 81,767 274,801 Datavote 38,246 1,128 2,628 Votomatic 3,555,712 80,839 272,173 Optical Scan 492,102 8,254 25,627 Precinct Count 58,188 578 4,226 Central Count 433,914 7,688 21,402 Summary Report — 3

Residual Percentage by Voting Type and Race

Ballots Cast For For Voting System 2000 Election President Senator AVM (Lever) 176,467 0.5% 8.2% Electronic 537,474 0.7% 6.1% Punchcard 3,593,958 2.3% 7.6% Datavote 38,246 2.9% 6.9% Votomatic 3,555,712 2.3% 7.7% Optical Scan 492,102 1.7% 5.2% Precinct Count 58,188 1.0% 7.3% Central Count 433,914 1.8% 4.9%

Sorted DataSet — Gathering of County Sums and Percentages 2000 Election

TOTAL AVM 176,467.0 175,642.0 162,022.0 0.5% 8.2% TOTAL Electronic 537,474.0 533,910.0 504,512.0 0.7% 6.1% TOTAL Votomatic 3,555,712.0 3,475,073.0 3,283,539.0 2.3% 7.7% TOTAL Datavote 38,246.0 37,118.0 35,616.0 2.9% 8.9% TOTAL Central Scan 433,914.0 426,226.0 412,512 1.6% 4.9% Total Ballots Ballots For Ballots For Residual Residual County Voting System Percentage Percentage Cast President Senator President Senator Hardin AVM 11,211.0 11,126.0 10,408.0 0.8% 7.2% Lucas AVM 185,256.0 164,516.0 151,614.0 0.4% 8.3% Franklin Electronic 378,963.0 376,636.0 356,566.0 0.6% 5.9% Knox Electronic 21,488.0 21,263.0 20,393.0 1.0% 5.1% Laka Electronic 92,763.0 92,244.0 85,461.0 0.6% 7.9% Pickaway Electronic 17,912.0 17,743.0 17,041.0 0.9% 4.9% Ross Electronic 26,348.0 26,024.0 25,051.0 1.2% 4.9% Adams Votomatic 10,727.0 10,235.0 9,882.0 4.8% 7.9% Ashlabuta Votomatic 40,378.0 39,472.0 38,164.0 2.2% 5.5% Athens Votomatic 25,888.0 25,447.0 24,167.0 1.7% 6.6% Auglaize Votomatic 20,212.0 19,892.0 19,485.0 1.6% 3.6% Belmont Votomatic 31,039.0 30,141.0 28,798.0 2.9% 7.2% Brown Votomatic 18,882.0 16,429.0 15,864.0 2.6% 5.9% Buller Votomatic 138,992.0 136,737.0 131,961.0 1.8% 5.1% Carroll Votomatic 12,576.0 12,261.0 11,917.0 2.5% 5.2% Champaign Votomatic 16,035.0 15,680.0 15,205.0 2.2% 5.2% Clark Votomatic 58,676.0 57,559.0 56,589.0 2.2% 3.9% Clinton Votomatic 15,366.0 15,070.0 14,666.0 1.9% 4.6% Columbiana Votomatic 45,294.0 44,427.0 43,360.0 1.9% 4.3% Crawford Votomatic 19,622.0 19,178.0 18,945.0 2.3% 3.5% Cuyahoga Votomatic 590,473.0 574,777.0 495,535.0 2.7% 18.1% Darke Votomatic 23,784.0 23,267.0 22,759.0 2.2% 4.3% Defiance Votomatic 16,610.0 16,242.0 15,797.0 2.2% 4.9% Delaware Votomatic 55,959.0 55,403.0 53,602.0 1.0% 4.2% Fairfield Votomatic 54,913.0 54,094.0 52,604.0 1.5% 4.2% Fayette Votomatic 9,484.0 9,278.0 9,099.0 2.2% 4.1% Franklin Votomatic 38,837.0 37,436.0 38,160.0 3.6% 6.9% Fulton Votomatic 19,161.0 18,896.0 18,408.0 1.4% 3.9% Gallia Votomatic 13,203.0 12,776.0 12,183.0 3.2% 7.7% Greans Votomatic 86,524.0 65,204.0 64,134.0 2.0% 3.8% Guemsey Votomatic 15,854.0 15,429.0 14,934.0 2.7% 5.8% Hamilton Votomatic 384,336.0 377,898.0 369,277.0 1.7% 3.9% Herdin Votomatic 948.0 942.0 899.0 0.6% 5.2% Harrison Votomatic 7,380.0 7,161.0 6,913.0 3.0% 6.3% Henry Votomatic 13,484.0 13,252.0 12,951.0 1.7% 4.0% Highland Votomatic 15,854.0 15,447.0 15,081.0 2.6% 4.9% Hocking Votomatic 11,034.0 10,756.0 10,484.0 2.5% 5.0% Holmes Votomatic 9,937.0 9,145.0 8,822.0 8.0% 11.2% Huron Votomatic 21,788.0 21,360.0 20,727.0 2.0% 4.9% Jackson Votomatic 12,918.0 12,490.0 11,863.0 3.3% 8.2% Jefferson Votomatic 35,449.0 34,636.0 33,253.0 2.3% 6.2% Lake Votomatic 10,584.0 10,320.0 9,787.0 2.5% 7.5% Lawrence Votomatic 25,179.0 24,451.0 23,271.0 2.9% 7.5% Licking Votomatic 63,490.0 62,466.0 61,401.0 1.6% 3.3% Logan Votomatic 18,823.0 18,454.0 18,148.0 2.0% 3.6% Lorain Votomatic 114,480.0 112,172.0 108,594.0 2.0% 5.1% Madison Votomatic 14,960.0 14,566.0 13,823.0 2.0% 7.6% Marion Votomatic 25,371.0 24,815.0 24,439.0 2.2% 3.7% Medina Votomatic 67,850.0 66,883.0 64,350.0 1.4% 5.1% Meigs Votomatic 10,228.0 9,795.0 9,253.0 4.2% 9.5% Mercar Votomatic 18,848.0 18,294.0 17,998.0 2.9% 4.5% Monroe Votomatic 7,377.0 7,115.0 6,721.0 3.6% 8.9% Montgomary Votomatic 237,680.0 230,987.0 204,136.0 2.8% 14.1% Morgan Votomatic 6,158.0 5,993.0 5,787.0 2.7% 6.0% Morrow Votomatic 13,145.0 12,839.0 12,510.0 2.3% 4.8% Muskingum Votomatic 33,520.0 32,624.0 31,738.0 2.7% 5.3% Noble Votomatic 6,210.0 5,988.0 5,704.0 3.6% 8.1% Paulding Votomatic 9,214.0 8,946.0 8,632.0 2.9% 8.3% Perry Votomatic 13,145.0 12,828.0 12,439.0 2.4% 5.4% Pike Votomatic 11,084.0 10,580.0 10,238.0 4.7% 7.6% Portage Votomatic 64,026.0 62,899.0 60,602.0 1.8% 5.3% Preble Votomatic 18,506.0 18,166.0 17,882.0 1.6% 4.5% Putnam Votomatic 17,743.0 17,344.0 16,870.0 2.2% 4.9% Richland Votomatic 54,088.0 52,779.0 51,467.0 2.4% 4.8% Sandusky Votomatic 28,441.0 25,744.0 25,121.0 2.8% 5.0% Scioto Votomatic 30,786.0 29,945.0 28,854.0 2.7% 5.3% Saneca Votomatic 24,931.0 24,351.0 23,808.0 2.3% 4.5% Shelby Votomatic 20,131.0 19,670.0 19,108.0 2.3% 5.1% Stark Votomatic 153,061.0 159,844.0 156,084.0 2.0% 4.3% Summit Votomatic 232,262.0 224,838.0 218,487.0 3.2% 5.9% Trumbull Votomatic 98,440.0 96,239.0 84,107.0 2.2% 14.6% Union Votomatic 17,288.0 17,024.0 16,427.0 1.5% 5.0% Van Wert Votomatic 13,471.0 13,219.0 12,889.0 1.9% 4.3% Vinton Votomatic 5,184.0 4,946.0 4,713.0 4.6% 9.1% Warren Votomatic 70,109.0 69,078.0 67,278.0 1.5% 4.0% Wayne Votomatic 43,151.0 42,436.0 41,126.0 1.7% 4.7% Williams Votomatic 16,170.0 15,919.0 15,439.0 1.6% 4.5% Wood Votomatic 52,832.0 52,194.0 50,427.0 1.2% 4.6% Wyandot Votomatic 10,059.0 9,827.0 9,607.0 2.3% 4.5% Tuscarawas Datavote 38,246.0 37,118.0 35,618.0 2.9% 8.9% Allen Precinct Scan 44,207.0 43,795.0 40,550 0.9% 8.3% Hancock Precinct Scan 13,981.0 13,817.0 13,407 1.2% 4.1% TOTAL Precinct Scan 58,188.0 57,612.0 53,963 1.0% 7.3% Ashland Central Scan 21,633.0 21,269.0 20,583 1.2% 4.4% Clarmont Central Scan 71,240.0 69,949.0 68,534 1.8% 3.8% Coshocton Central Scan 14,493.0 14,279.0 13,600 1.5% 4.8% Erie Central Scan 36,836.0 35,015.0 33,991 2.3% 5.1% Geauga Central Scan 42,963.0 42,600.0 40,947 0.8% 4.7% Hancock Central Scan 16,977.0 16,800.0 16,262 1.0% 4.2% Lucas Central Scan 23,163.0 22,834.0 22,063 1.4% 4.7% Mahoning Central Scan 116,889.0 114,119.0 109,529 2.4% 6.3% Miami Central Scan 43,555.0 42,878.0 41,818 1.6% 4.0% Ottawa Central Scan 20,185.0 19,968.0 19,577 1.1% 3.0% Washington Central Scan 27,080.0 26,515.0 25,408 2.1% 6.2%